Data Modelling & AI : A Deep Dive into the Future of Tableau with Kirk Munroe

0:02Boom.

0:03We're back.

0:04Um yeah, so I took a break uh to welcome the newest member of my family, Estelle, into the family and just enjoy time with the kids.

0:12It's super, super important when you

0:14uh have this opportunity to take it and make sure you make the most of that now that said it did come at a really bad time for Tableau content 24.

0:232 had dropped

0:24And actually leading up to me taking time off, I was even more busy because of a bunch of other things going on in life.

0:30And so I took a break and I was a little bit behind the schedule.

0:33And so here we are with the video again

0:36But just to say, look, we've got a ton of content to get through.

0:3924.

0:392.

0:39I haven't even covered some of the main features, multi-fact analysis, so many things to cover there.

0:45And then we have 24.

0:473 just round the corner.

0:48So I'm gonna have to start straddling two releases at the same time over the next month and a half.

0:52And then we're going to be at DataFam Europe.

0:55That's a new Tableau conference here in Europe.

0:57So that's going to be super exciting because it's in my hometown, London.

1:00So we've got a long run up to Christmas and hopefully we'll have lots of content every single week

1:06In today's video, I'm talking to Kirk Munro.

1:09Now, Kirk and I exchange conversations all the time on LinkedIn.

1:13He's the kind of person where, you know, he just has so many useful insights over the years with Tableau.

1:18He's a very esteemed member of the Tableau community and he's also written a book about data modeling in Tableau.

1:24So what I said to Kirk is hey Kirk, instead of us having these sort of small exchanges on LinkedIn and in various places in LinkedIn DMs

1:30Let's just have the conversation here on the channel and we talked for two hours.

1:35So full disclosure, the whole conversation unedited is right here in this video.

1:39It's two hours.

1:40Yes, it's long.

1:41It's more like a podcast.

1:43We don't do any screen sharing.

1:44We're just focused on talking about these topics.

1:46I think it was really valuable to get his time and his experience of what he's seen in the industry in this conversation.

1:52So

1:53As ever, everything is timestamped, it's super clearly laid out, and we covered everything from data modeling, uh some of the sort of conceptions around data modeling, but also some of the missed opportunities

2:03And then secondly, we talked about Tableau's future, some of the AI capabilities, but also the way Tableau is positioning quite a few things.

2:09Now, the context of this is we recorded this a month and a half ago and I've only just got round to releasing this video now

2:16So if it looks like we haven't watched the most recent Dream Force event, that's probably what the context is there.

2:21But nonetheless, as ever, let's get stuck in.

2:24It's good to say that again.

2:25Kirk, how are you?

2:27I'm doing great.

2:29Good, good.

2:29I'm I'm I'm glad we set this up.

2:31I I did my best at ignoring you for about a month and a half and then uh

2:35I finally picked up the the bevo and uh messaged you about three weeks ago to say, Hey, when are we doing this?

2:41And realised it was me that was uh not replying to you.

2:44Well you you're more important stuff going on in the

2:47No, yeah, no, it's a pretty pretty hectic time of ti time in my life.

2:51But uh I appreciate your patience and sort of uh agreeing to come on at very short notice.

2:55So I think we we pinned this down last no last week, yeah, very beginning of last week.

2:59So

3:00Good to be finally talking to you.

3:01Um so many topics are discussed today.

3:04Um I think we we keep we keep crossing paths and lots of different threads on LinkedIn, on YouTube.

3:09I always see your comments there.

3:10Always very insightful and there's always they're always fantastic little nuggets.

3:14So I'm keen to have you know an hour or maybe more with you just uh to get it all out of you and sort of go back and forth on a on a couple of topics.

3:21So yeah, I really appreciate it.

3:23Yeah, and I'm really looking forward to it.

3:25Just just for the benefit of the audience, um maybe it's good to start with an introduction of you know who you are, how long you've been using Tableau, and yeah, just just just how you got sort of passionate about Tableau.

3:35Um Yeah, so

3:37Well first I guess I've been in the BI space for a really long time, which is why I see a lot of things come back.

3:43Um I had uh

3:45I'd move from I grew up in a relatively small place and moved to um Ottawa, Ontario and Canada to run a startup in '99.

3:52Um and that startup didn't work.

3:55But it was interesting.

3:56It was when the internet was really, you know, burgeoning and, you know, the dot-com bubble and everything.

4:01And um somehow Cognos, who would have been along with Business Objects, one of the two big

4:08early BI vendors, um, somehow I convinced them, I got hold of their senior vice president of products and convinced without any background in it, convinced them that they should hire me for um

4:19What do they call the job at the time?

4:21It was a product manager for performance, sizing, and scalability because they were going to the web, right?

4:27Like so they were doing servers and everything in the client.

4:30So

4:31Right.

4:31With Tableau moving everything to the web client, I'm like, well, we did that at Cocnos twenty-two years ago or something.

4:37It's not new.

4:38Do you know what I mean?

4:39So

4:40It was really new then because parallels were pretty crap compared to today and stuff.

4:45Um anyway, so I ended up working at Cognos for

4:4910 years.

4:50Um and then I went to a oh and a funny story.

4:54So the first time I saw Tableau, I was a product manager in the first five or six

4:58And then at the end I ran global sales enablement for us after we got bought by IBM.

5:03Right, yeah.

5:04Including having the competitive team.

5:06And I remember someone who worked on PowerPlay, which so I was on a product called Metric Studio, which I think still exists, believe it or not.

5:14And then one of the product managers from PowerPlay ended up running the competitive team on my team.

5:20And she said to me, this is about 2000.

5:22Hate maybe.

5:24She goes, I've something um that you have to see that you're not gonna want to see.

5:29And she shows me Tableau Desktop.

5:31And at the time it was probably Tableau two or something.

5:34I don't even know.

5:37And it's a little thing that I think some people still don't get today, but the people who get it who get it is

5:43I think she dragged over some measure in a time dimension and it automatically made a line chart.

5:48And I go, why can't we do that?

5:49Like I know it seems little, but like a lot of people

5:52uh especially if you're going after true business users can't yeah like don't get that like they don't know that that should be a liar.

6:00I mean every other product to this day it's like bring your data in how do you want to visualize it?

6:04Why doesn't the product help me out of visualize

6:06Anyway, um so then uh I went to a supply chain analytics company called Canacis program product comp

6:14I ran out of marketing and product management for a couple of years, did another startup.

6:17Anyway, that startup didn't work.

6:19I went, that's enough to rise the startup.

6:21And uh I thought

6:23I'm gonna go back and be an individual contributor.

6:25I've had enough of this climbing the corporate ladder trying to run company stuff.

6:28And um and Tableau had an SE role.

6:32And I knew a lot of the people from Tableau Canada

6:35uh from the Cognos phase and I went um uh and I went and they're like the only thing is do you really want to go back and be a sales engineer?

6:43I'm like I actually really want to go back and be a sales engineer

6:45sales though like I just want something monstrous, right?

6:49Um yeah.

6:50And then so I remember going through the interview process.

6:52So this will get us onto the data part too right away.

6:54And I looked at Super Storm.

6:55I'm like

6:56Everyone's data looks like that's great.

6:58Yeah.

6:58Like data doesn't look like that.

7:00Right?

7:00So I remember and they made me do a demo.

7:02I go, I'll only do a demo if I can not use super snoring.

7:06And they went sure.

7:07And I went uh inside Airbnb, which I still use today for demos all the time.

7:12Yeah.

7:13And went, I went, well, what if I was a city, what if I work for the city of Toronto?

7:17How would I use this?

7:18data.

7:18If I was a host, how would I use this data?

7:21How would I and that was my demo and I went right and that made me think, oh man, I love this product even more than I thought, but it's even harder to model data in.

7:29Yes, definitely.

7:31So anyway, I worked that I worked at Tableau from 99 to sorry 2017 to 2021

7:38And then I left because my wife had actually started, she's a three-time Cavalier ambassador at this point.

7:43She'd started a consulting company and I went

7:46Oh, that's the next stage, let's do that.

7:48So we have a very small Tableau consultancy called Paint With Data, and we just help clients get better Tableau, basically.

7:56Good.

7:57And yeah, so then just super quick, I guess, to get in the book.

8:00How the book came about was what I think what we have people with I think more than anything is they'd already been using Tableau.

8:08The perf there it doesn't look that good.

8:10It's not communicating well, so there's that side.

8:13But the other side it performs like garbage usually.

8:15And like so we help them

8:17Work quicker.

8:18And we have a philosophy that everyone should be able to get from the highest level of aggregation down to offending record in ten seconds and three clicks or

8:26So we can't always do it, but that's an obsession.

8:29So I'm like, so you can't do that if your data is not modeled really well.

8:33And then

8:34The publisher hit me up and I'm like, I don't really want to write a book and but I did a search and there's no books on the topic.

8:41So like Carl's got a book on Tableau prep uh prep and there's a million on Tableau, but there's none really on data modeling.

8:47Data model

8:49And I'm like, sure, I'll I'll take this on.

8:51And then that made me realize I thought I was good at data modeling and Tableau.

8:55And obviously I got way better because it forces you to get way better at it.

8:59And then Yeah it does, yeah.

9:01Yeah, since then I seem to talk about it a little bit.

9:05Amazing, amazing.

9:06Yeah.

9:06That that's such a rich and colourful like uh heritage with Tableau and it it's it's also

9:12You're quite fortunate because yeah, you you've you've come from uh a different era in analytics, very much so.

9:18You've probably seen

9:20Uh the end of one with cognos, the start of another with Tableau, and I think we're now full circle again.

9:26We it's kind of the third

9:28next era just forming.

9:29We'll get onto this later.

9:30And yeah.

9:38Right.

9:38And and and trends come back into into into fruition.

9:41And so I think

9:42One of the interesting things I've always asked other, you know, people with Tableau specifically, you know, in that sort of time frame, I've already I've I've been using Tableau for Tableau.

9:51just just under a decade as it were.

9:53So I've only really seen one phase of this.

9:55Um how have you kept your your passion in analytics?

9:58How have you kept your passion in

10:00in data going, is it actually the fact that you found new things to discover and do?

10:05And that sort of constant pursuit to make things better is sort of pushed you that way?

10:09Or is there there's some secret source that maybe

10:12You can synthesize for the audience.

10:14That's a great question.

10:18I think my Venn diagram is uh is an obsession around two things.

10:23Um the first is um

10:26People making decisions completely based on their opinion and not fact.

10:29It's always driven me crazy from the time you were young.

10:33Like I I uh Yeah

10:35I could tell so many stories on that, but everything I want.

10:38But you know, I want it's not that I won't make a decision without data, but if I make it without data, I'm like, I'm gonna collect data as I do

10:44do this.

10:45Because um I absolutely love the whole field of behavioral economics, behavioral psychology, and I just know how garbage we are as humans at doing that.

10:54Do you know what I mean?

10:54So I have to check myself and use data.

10:57So I'm very passionate about using data to um to bat to make the right

11:02Keep you on course, right?

11:03Um I could go on about strokes game and golf for hours, right?

11:07Uh and then um and then on the other side, I'm obsessed about um

11:13uh uh that using technology for productivity to make things quicker and easier.

11:19Right.

11:19So as an example, one time I was

11:23Trying to espouse why Tableau is better than Power BI, which isn't the easiest thing to explain, but yes, I'm explaining it to this customer who's a Power BI customer.

11:32And uh he said, well, other than it being fast and easier, how you know, how's it better?

11:37And I said, why would you adopt any technology?

11:40For any purpose other than for it to be faster and easy.

11:43Like that's technically slower and harder.

11:46Why would you do that?

11:47Like every literal thing.

11:49Like you could drive a bicycle instead of a car if it didn't get you there fast.

11:53You know what I mean?

11:54Like we wouldn't have

11:55Tons of problems, right?

11:57Like faster and easier is the reason.

11:59So in Tableau, it just is the intersection of those two things.

12:03things right so um i know other people don't like tab or sorry people struggle with tableau i think because you know this and you and i got into this on linkedin but it starts with the aggregate and you go down

12:13And they don't cost the number.

12:15But I'm like, but if I have to start with all the details and aggregate up, you've already slowed me down.

12:19I'd rather start at the top and drill into

12:23Yeah.

12:24So and and Tapov's done a great job.

12:28Yeah, I guess we could talk about whether we think this is ending or not, but they've done a great job of innovating for at least the past seven years.

12:34I've been

12:35using it pretty deeply.

12:37Like we all have our audience about laying out a dash like that I can't snap to a grid and align.

12:42Like all the things were dashboarded, but they've still done

12:46So many big innovative things that they've kept up, like in and other products don't typically.

12:53So

12:53Correct.

12:54And you know, um we I'm sure you know Jock McKinley's um, you know, uh innovation chart, right?

13:01That should shows and he I think he still updates it on Tableau Public Hall.

13:05I'll put it up on screen.

13:06You know, it shows all the innovation happening in Tableau um over the years.

13:11And what is what is pretty incredible about that diagram is if you can you can literally plot in a linear s on a linear level, like just with each release.

13:19How frequently big momentous changes come about to the platform.

13:24You know, um, there's some features which are sort of incremental to the platform.

13:27They add a bit of quality of life, they add a bit of um, you know, finesse.

13:31But then there's some changes which are, oh, okay, actually, we've noticed a trend that's hard for our customers.

13:38We're going to invest a lot of time into enabling this whole swathe of capabilities.

13:44And to me, data modeling was that sort of, you know, thing, relationships as a specific thing.

13:49Tableau prep was kind of the precipice of that, right?

13:52but it was solving it in a different space.

13:54It was solving it in a in a sort of, let's say, um pre-pre-analysis sort of those.

14:00Do this, then do that, right?

14:02And then data modeling came about and it actually opened up the possibility to do this while doing that.

14:07So it allowed you to kind of do that stuff at the same time as you're asking questions.

14:11And we'll come onto this in a bit more detail.

14:13But I think that's

14:14That's just been such a huge shift.

14:16And I think part of the reason I wanted to talk to you today was because I think that shift was kind of missed by a lot of people, right?

14:22I think I think a lot of people haven't taken the opportunity to really evaluate what that means for their work

14:28Right.

14:30And I think a big reason it's missed is because so many of these community projects

14:37Right.

14:37And the things people talking about can you use very small data.

14:40So when they join it and explode it, it's still not that big.

14:44And I must not send anything down.

14:46Yeah.

14:46So as a consultant, the biggest thing I feel like I talk to

14:51I have to unlearn, like if people unlearn stuff, they learn from the community that I go, what they taught you was completely fine when you've got 10,000 rows of data, but it's not going to work on your data set.

15:03And therefore some of those people don't, I think, ever learn it, like the leaders in the community, because they're not working with these massive data sets.

15:13I'll give you a really good example is I take I show people this to show simple relationship kind of in the impact of answer all the questions without and not expose your data, right?

15:23I go joins, expose your data, and not your questions.

15:26Right.

15:27So I take three simple ones from inside Airbnb where I can go.

15:31I'm going to take listings, I'm going to take bookings, and I'm going to take sorry.

15:38Listings, reviews, because you don't have bookings because they're scraping, they don't know bookings, right?

15:42Listings, reviews, and reviews are a proxy for bookings, and comments.

15:47And I can take, picture how what that would explode out.

15:50Do you know what I mean for a city say this like Toronto or London or whatever?

15:53I go, but because it's a relationship

15:55I can go, all the listings have you drilled down on my whole 10-second three-click thing, right?

16:01I can show you the calendar of the next 365 days for the price you pick.

16:06So you have to drill into it first.

16:08And I can show you not only the reviews, but I can show you the comments on the reviews.

16:12And the thing performs like

16:14Snap Chris.

16:15Right.

16:15But people are trying, but people try to join.

16:18We get this all the time.

16:19We have a few customers with call center stuff and they try to join the call center.

16:23Like join it in.

16:24I'm like

16:25Yeah, yeah.

16:25They're like, why does it take forty-five seconds?

16:27I'm like, yeah, it's impressive you can get in forty-five seconds.

16:30You know what I mean?

16:31If you were a relationship, it would come instantaneous.

16:34Yeah, exactly, exactly.

16:36And you know, the most the simplest example I always give about that is like support support ticket data.

16:41You know, someone wants to know how many tickets are active on any given day.

16:45historically, right?

16:46So you've got a trending number of open tickets at a specific moment in time.

16:50And the hard thing about that question, you don't know it's hard until you try it.

16:54is that you have to have a starting point, right, if you're gonna do it the old way.

16:58And then you have to preserve that starting point all the way through time to get the running total at any given time.

17:04Like absolute nightmare

17:06And then I did it for the first time with relationships.

17:09And it, you know, it it wasn't something I even thought to do.

17:12I just thought, what if you did this instead?

17:15And I tried it and it worked.

17:16And I was like,

17:17Wow, and this is this is incredible.

17:20And and that was the first time I realized okay, I've been doing this wrong with the data model entirely.

17:25Like this is what it really meant for.

17:27Yeah, that yeah, what I ran into recently, and there's a lot of these is because people come up with these questions, right?

17:34So they wanted to know um

17:36Yeah, it's effectively a call.

17:38Like when a call opened, they wanted to know in any given week how many calls opened this week, how many calls closed this week, they were open this week, how many call how many were open this week

17:50Um close this week but started in another week.

17:53And then how many pass through?

17:55You know what I mean?

17:56Like we're already open and pass through.

17:58And they're like, how do I even do this?

17:59I'm like, hmm.

18:00We'll scaffold this to a new date.

18:02We'll go to that date.

18:03The thing responds instantly and it took me, I don't know, an hour to come up with that logic for them.

18:07It's like I'm like, I wouldn't even know how to create a join to do like that logic's crazy.

18:13Because you have to evaluate it every week to get and I go the great thing about this one is if you want to drill that up to a month, all the logic's gonna work.

18:20Yeah.

18:21Right?

18:22Go down to a day.

18:23It wouldn't make sense if a day didn't have enough volume, but it would work, right?

18:27Like

18:28But it's a painful thing.

18:29And you know, even if you sort of live up to the what I would call the traditional data stack and you have a data engineering team and you ask them to instantiate something that solves this

18:41The thing they'd come back with wouldn't give you the flexibility that Tableau gives you.

18:45And this is sort of what I call last mile analytics, which is it's fine to push this logic sometimes back into the data warehouse.

18:51Absolutely sort of live with that philosophy.

18:54But

18:54There does come a point where you lose agility in the speed and the ability to answer the question.

18:59And I think there is there is sometimes

19:03a space to keep that stuff flexible in Tableau.

19:06Yes, push back the metadata, absolutely.

19:09But I s I st I s I still

19:11I know I I work in a consultancy with lots of data engineers and I think their philosophy is move everything back into the warehouse so everything can see it.

19:18But when you're doing analysis, and this is hard because you know

19:23We'll come onto this as well.

19:25Sometimes the person building it isn't the person using it.

19:28And I often feel like, you know, you as a consultant, me as a consultant,

19:33We y we dog feed our own products.

19:36Right.

19:37Right.

19:37And so we have that perspective.

19:40Whereas sometimes in an organization with reporting teams and business users, the reporting builders are a bit disconnected.

19:48from the frustration of trying to answer that question.

19:51And the person can articulate what it should do, but can't articulate how to do it.

19:55And then the reporting person knows how to do it.

19:58But they don't understand why they want it done, right?

20:00And those two never meet in a in a sort of a logical way.

20:03It's so hard to bring those two words together.

20:05Right.

20:05Yeah, no, I I think if

20:08It sets people back a lot of the time, but when we start a new consulting engagement, the almost always the first question I ask is, can you walk me through the application which is your people are inputting stuff to generate this data?

20:18And they're like, why would I do that?

20:19I go, because I don't know what I'm looking at

20:22Like it's just numbers.

20:23Like I want to know, like is it validated first?

20:26Is it do you want like and that's what's like no one's ever thought to ask that question before?

20:30I'm like, I don't know how they can help you with your data then.

20:33Correct.

20:33Do they know the answer to that?

20:37Sometimes they do, but it often makes them run that person down, which then gets them into a conversation that they should have been

20:42having me right exactly yeah like you'll see sometimes Bethany Lyons and I'll get into this on LinkedIn right I'm like my thing always is like oh the hardest thing for a data engineer to

20:53do to me is that if you think about applic like software in the traditional sense, they have three tiers, right?

21:00They have a UI and they have

21:02Almost always they have a layer that's got logic and then they have the data, right?

21:06And and then people extract the data from those applications and they hand it to a data engineer and then some kind of analyst and go

21:13give me answers out of that data and I'm like, but all the rules are gone.

21:16They were in the mid-tier.

21:17They were never in the data in the first place.

21:18Like I'm like, if you don't know how to recreate those roles in the data, it's it's not a data cataloging problem.

21:24It's uh how did that data get generated?

21:27problem.

21:27And I think in consultancy sometimes I think it frustrates people because I'm like, even with great data modeling tools, I go, I'm gonna annoy you how long we're gonna spend in getting your data model right.

21:39But then

21:40You'll get answers so quick it you won't even be able to wrap your head around how quick it is.

21:45Cause like once like once it's clear what we're pulling like you know in Tableau speed

21:50When those pills were dragging on when we know what they are and the relationship between them, it's just oh that's where table is so fast.

21:57But if you don't know what you're pulling on, which I see people do all the time.

22:00And then they they pull it on until they go, that number looks like I think what the number is on.

22:04like that we can't do that.

22:05Yeah, you like you do need to learn to like what's the word?

22:11Not test it, but like

22:13You kind of need to have a lived experience of how that metric comes to life, right?

22:17And you have to actually ask someone

22:20Sometimes not even the person who's asking for the report.

22:23Sometimes it actually you have to go back sort of a few layers down to the person who actually lives that world and they'll tell you the nuance of that metric and then

22:31You can kind of pull that through.

22:33100%.

22:34And we and we have a logistics fight now to your point about last mile, how you can solve lots of things in the last mile.

22:40Yeah.

22:41They said we can't just use

22:43um transit number or whatever it was, right, to do it because some of them get billed to the customer and some don't.

22:49And then they quickly go, well we can go back to our data engineering team and we can work them out.

22:53I'm like

22:54But how would you know looking at the application?

22:56And they're like, yeah, well the ones that we because they tr they use the transit number to move from one of their own warehouses to another, right?

23:02Yeah.

23:02I'm like, it wouldn't have a PO on it.

23:04I'm like, great, we'll just filter PO and blah blah blah.

23:07Like you don't have to go back to your DM edge.

23:10We'll just filter the ones out with an Alpia.

23:12Like in their hands, right?

23:14Yeah, there you go.

23:14Yeah.

23:16I just say if you like

23:17Five thousand dollars engineer data engineering time or whatever in a month.

23:21Do you know what I mean?

23:22Yeah.

23:23Right?

23:23Like some of these you don't need to engineer.

23:26Or even build in the ability to set that context depending on the question that's being answered.

23:32Yeah.

23:33Yeah, like the great thing about this then is if you want to know what percentage of

23:37Times you're just sending things between warehouses and not to customers.

23:40Now that's not gone.

23:41Do you know what I mean?

23:42You can answer this.

23:42Exactly.

23:43Exactly.

23:44Exactly.

23:45So yeah, like drilling into this data modeling skill, how did you um obviously you talked about consulting, you've seen you've seen a lot of your customers have have challenges in this space.

23:55Um, maybe let's dive deeper into sort of what what drove you to write the book and like what problem were you trying to solve the book?

24:02If if if someone had read through your book and come out the end of your book, you know, having really sort of engaged it

24:09What would what would be sort of your desirable effect to that personality?

24:13So if I if I read your book end to end, what would what what would you hope for me if that makes sense?

24:18Yeah, that's a great one that we're asked it that way.

24:20Um I think

24:22I think two things depending on the role, right?

24:26But at a minimum, the first thing I would get them to expect is, oh, if I put more time into this.

24:31Yeah.

24:32Like the end experience in Tableau for both authors and um viewers is going to be so much better.

24:40Right.

24:40Um so if they're a dashboard developer, they'd still get, I can do that.

24:44But what I hope organizations would get out of the book

24:48Is, you know what, we shouldn't have everyone with a creator license creating data sources.

24:53We should have a relatively small number of people creating data sources that are really good at it, that are creating high performance ones, as published data sources

25:01And then have default go to that, right?

25:03Or if they have data management even virtual connections, which I didn't get, but they were pretty new, right?

25:07So I didn't get that deep in.

25:09But I'm like, again, for that fast and easy.

25:11The quickest way to be fast and easy is a lot of few people build really good data models and share them.

25:19Because then Tableau just becomes so much easier.

25:23Do you know what I mean?

25:24If you do that.

25:24So that would be Yeah.

25:27And uh

25:28That's that's a really powerful um point and I want to put a pin on data management.

25:33I'll come to that in a second.

25:35There's a bit of that that drives me nuts uh to do with pricing and how it splits like that.

25:41It's like

25:42God.

25:42Like you yeah, absolutely.

25:44We'll come back to that.

25:45But like I I don't want to sort of distract from the the conversation about your book.

25:48So I kind of I started at the end with your book by asking like

25:52What what would someone get out of your book?

25:53And you you've given a really sort of clear, cohesive answer.

25:56And to anyone who's watching or listening, uh, you know, I I couldn't recommend your book more.

26:01I think it's such a

26:03it will take you out of your uh day-to-day sort of um you know when you're working in analytics sometimes you find yourself just repeating the same habits right

26:12Even even though you've done something, you've done a use case before, you know how to do it.

26:15You just you just do what you've been doing all along.

26:17And I think the great thing about your book and the way it structs is it forces you to

26:20to to to at a high level think how you approach things in a new way.

26:25And I think your content more generally, even your Tableau conference um session most recently.

26:30I think was a really powerful demonstration of how that can have a huge impact on on the work you do.

26:37And, you know, I I couldn't recommend it more.

26:39So

26:39What I wanted to do is now go into the book a little bit more and just just give you an opportunity to talk about the book and uh you know what's your favorite chapter?

26:47Like uh, you know, if you were to tell someone just to read one part of the book, what what would it be?

26:52Um the problem with this is is it I should have the book in front of me.

26:56I at the last chapter I think is 15.

26:58Um We can grab it.

27:00Yeah.

27:01My favorite chapter um was the last chapter because I always think

27:06It's best to walk people through use cases of you know when I would embed a data source, when I wouldn't um embed a data source, uh, you know, all that kind of thing based on um the specifics of whoops.

27:20of uh uh of the use cases because I think that's always the most important thing which was kind of what the hopefully the presentation was the TC that kind of feel that yeah yeah yeah

27:30The problem with it is unless you know taboo you can't jump to the bedroom after.

27:34Because you need to know all the things before.

27:36Like uh like um like uh like on the

27:41Similar it won't go back to the book, but I think it's similar.

27:44Like today I just had um a blog on uh Kevin and Kevin's um blog on on Included.

27:53Yeah, and mostly it's on

27:57It's on include and exclude, but mostly it's on how not to use them.

28:01Except for one include.

28:02But

28:04But they're a perfect way to understand Tableau's order of operations.

28:09So what it's really about is what it means, the order of operations, and what being

28:14in the view means.

28:15Correct.

28:16But um but those are kind of boring conceptual topics.

28:21But this if you work back from it, you'll get the you know what I mean?

28:24If you work back from why in include and exclude don't work the way that almost anyone thinks they

28:28work right yes and you work back you're like oh but now I understand the order of operations otherwise it's just I can follow this chart but I don't really get it.

28:35Yes.

28:35Right.

28:36So um so the the the last chapter of the book is definitely my favorite chapter of the book but it's kind of things like

28:43People have to take the time to understand how relationships work, I think, is a good example, right?

28:48And then before you do publish data sources, you have to know how they work, how they're secure.

28:53Right.

28:54And if if I'd done the book today, I would definitely have a chapter on pulse because pulse is easy, except

29:03You have to understand published data versus how they're secured, uh, how robotical security works, or pulses for good.

29:09And if you have that set up.

29:11You can I don't have logos, you can create a pulse metric in five minutes.

29:14I'm like you can create one in 30 seconds if your model's good.

29:16If your model's not good, forget it, right?

29:19Correct.

29:21So within the model actually for Tableau Pulse.

29:23Yeah, I would almost never have a model without row level security, which wouldn't, which is not about row level security per se, but it's because

29:30Then I only have to create one metric, and depending on who looks at it, gets the metric they do without me and create a bunch of metrics for people, right?

29:37It's not even about the security, it's about

29:39Yeah.

29:40So I guess back to the book in the last chapter.

29:43But but I think there's a lot and I tried to write the shortest book I could write.

29:47Now there's a lot of screenshots in it, but and it still came up to three hundred and twenty five pages.

29:51My prediction for the publisher was oh we get

29:56But but it's um yeah, there's a lot of moving pirates is the problem.

30:01Like there's uh Yeah, yeah.

30:03But like I think I think I think um I've I've had a really difficult experience with publishing.

30:08We're going off track here, but I've been approached by publishers before and obviously I'm a video person, right?

30:13So yeah,

30:14Th there's there's a there's there's just something in my head around I think I could make a three hour video better than I can make a chapter.

30:25You know what I mean?

30:26They're they're roughly the same length in terms of time to consume and assimilate.

30:31But I think I can structure a video better.

30:34I don't know what it is about that.

30:35And you know, I've had publishers, oh why don't you just

30:38take that video and you know synthesize it into a book and I'm like because I think I'd struggle with that.

30:44It's it's just something that I really, really struggled.

30:45So when you first, you know, wrote wrote a book,

30:49Did you have any sort of reservations like that where, you know, maybe maybe you're more familiar with blogging and someone's asking you to write a book or if the format was intimidating anyway and uh

30:58Did did you overcome that or actually was it a much easier process than you you maybe thought?

31:02For for anyone else who sort of an author.

31:06Selfish question as well.

31:08It's a great question.

31:09It was um

31:11No, it was definitely harder than I thought, which most things of that magnitude are.

31:17Um uh I think

31:21And um I mean it doesn't sound like a John Grism or anything.

31:26So um uh or Pickett, whoever, right?

31:29Um uh I think with the book the oh

31:34The almost the only really good thing about a book, I think, is it's such a great reference when people have to go back later to reference things.

31:41But it's um that's what's hard to get at a video, I think.

31:45And um

31:47But it's hard to get someone to read your book and it's hard to read a book anymore.

31:50Because I think we're not only you say are you a video guy, I think we're way more video consumer.

31:55Right?

31:55Generally, yeah.

31:56Yeah.

31:57I'm very auditory when it comes to things like if I were to listen to something

32:03I don't know, like the state of politics in this country or something.

32:05That'd be a podcast for me.

32:07Like I wouldn't have to see them doing it.

32:09I'd be doing yard work or something at the same time and because it's easy to process those two things at once.

32:14Um so yeah, I'm not sure I would it'd be hard for me to recommend people to write a book other than it's a really cool thing to have done.

32:24Uh yes, especially.

32:26Yeah, my favorite Tenta

32:28And there's been a lot of good ones.

32:29But my favorite one is Tim Urban's on procrastination.

32:33I don't know if you've ever seen it.

32:35And he it's great because he talks about I'm paraphrasing so I'll be slightly wrong, but he goes, I knew six months ahead I was giving it.

32:42And it's on procrastination, remember.

32:44And he goes, I'm gonna get it done now.

32:45And then three months OB hasn't started, and then a month I always still hasn't started.

32:49But then his line, which I love, and writing the books exactly like this.

32:53is he goes, it was at that point I realized I didn't want to give a TED talk.

32:57I wanted to have been someone who gave a TED talk in the past.

33:01And that's kind of what the experience is like

33:05Right, right.

33:06I think a lot of life is like that.

33:08I wanted to have done it.

33:09Like I wanted to have done it, but I don't want to

33:12Yes.

33:13Not exactly put the work in.

33:14Do you know what I mean?

33:15Because I I don't mind working for it.

33:17It's like it is a labor because what the great advantage of video over of

33:23book is that you can add all the color to it that you can.

33:26And right the book's driving me crazy because I'm trying to be concise as I possibly can.

33:30And I want to say more.

33:32And I'm like, people aren't going to read it.

33:33It's already 325 pages.

33:37But I'm like, but that's why there's such good references though.

33:39It's so funny to so I'm glad I've written it for a reference.

33:43Do you know what I mean?

33:44That'll sit there for a reference for a while.

33:47Yeah.

33:47And I think I think there is I always I'm actually quite envious of people who have written books because as so many make videos, one of the things um

33:56you you you have to do when you're making videos.

33:58You have to get into rhythm um with necessarily YouTube or whatever.

34:01The algorithm likes it it's a bit like a TV show essentially.

34:04If a game show happens every week at a set time, guess what?

34:07You tend to stick with it and you watch it for a couple of weeks.

34:10Whereas if it happens at irregular times, you kind of fall out of love with it because it never it never sinks with your schedule.

34:15Same is true with video

34:16And the thing I think authors and people who've written books have is that is that is that roadmap, right?

34:23You can take your content and seeing as you own the book, you've got the freedom to be able to say, right.

34:28I'm actually going to make this more accessible in a specific structure.

34:31And so I think I mentioned this to you, like, you know, uh like um in in messaging.

34:35I was thinking,

34:36The pairing of writing a book and pairing that with video, I think is a it's a super powerful mechanism.

34:44Because for the people who aren't into reading books, it brings them into the topic, I think, just enough.

34:50To pique their interest in the reference.

34:52And for people who are interested in the reference, it allows them to have something to share to be able to demonstrate the thing that they have.

35:00So there's a social element to it, which books

35:03Books I think struggle with that social element because once you read something, you like it.

35:06How do you how do you how do you share it?

35:08Like you can't always take a screenshot of the book.

35:10That you know, that infringes copyright in some respects.

35:12But and and

35:14There's no digital thing if unless you're reading on an e-reader or you're gonna paraphrase the book or quote the book.

35:19You know, people do lots of different things, but it's hard to capture like a chapter, the sentiment of a chapter.

35:24and share it with someone.

35:25Uh and so that's that's where I think, you know, books and videos or content and videos can can really help because the videos can get that reach.

35:33They can get that broad brush appeal, bring people in.

35:36And if they're passionate

35:37You kind of you you kind of bring them in.

35:39So that that's sort of um I it's not something I'm gonna explore.

35:43I still do not have time to write a book but

35:48I'm surprised, not surprised, because no one else like in the history of the world, almost no company has gone from one model to another, but I'm surprised a good publisher hasn't

35:58Like set up a pretty good platform for that.

36:01Do you know what I mean?

36:02Because it's correct.

36:03It's worth giving up.

36:04Um I would say it's worth giving up a lot of whatever you're gonna make on something for someone

36:09Yeah.

36:15So you don't have to write the formula.

36:18Do you know what I mean?

36:19Like it's like I don't want to

36:22I mean, since the company doesn't exist anymore, I'll give you an example of a cognos that of some exec team that drove me crazy, is that they brought in uh a consultant to help come up with the development process, right?

36:33Like a better kind of

36:35Yeah, you know, before Agile was big or not like that hasn't been ruined either.

36:39But uh and I'm like, but is there a core competency like developing um

36:46Like, is our core company developing a development process?

36:48Because Microsoft had something called Microsoft Solution Framework, which told other software companies how to develop software.

36:54I'm like, why don't we just adopt something like that?

36:55that.

36:56Yeah.

36:56Like, no, I'm like, it's not our core composition.

36:58It doesn't make sense.

36:59So the parallel to this is I'm like, I don't want I don't want to figure out like what you just said made total sense to me.

37:05I love this idea of reference video reference.

37:08Like why doesn't someone have that platform?

37:09Or if they do, and someone's put it in their class.

37:12Do you know what I mean?

37:14Exactly

37:15They can have do you know what I mean?

37:17Just uh Yeah, exactly.

37:18I don't want I don't want to learn how to build that thing.

37:21Do you know what I mean?

37:21I just want to deliver the content.

37:23Yeah, yeah.

37:24And you know, the the closest I got to a book was with one publisher and like m my my line m the line I wouldn't cross is I wanted to make a but about something I was passionate about.

37:37And I wanted to make it in a way that felt authentic to me because I felt like if I just put out a book and I already have a a presence and people know my style, they know my communication.

37:49Like it it would be a bit off, right?

37:50Oh, Tim just put out a book because he can, kind of thing, rather than, oh, this is Tim taking, you know, what what he knows and putting it in in into another format.

38:00Um and the the route I I sort of ended up sort of thinking maybe this is something to do for the future is like going down the route of self-publishing.

38:08Not not to discredit the um

38:10the the the the force that and you know ri rigor that publishers have they have they they have it down to a T right.

38:17They have editors, they have

38:19um people who can print the book, they have distribution, they have m like they have it all down.

38:24And to to take that on in self-publishing is is really, really challenging.

38:29But

38:30If you're not so precious about the reach of the book and you know, you know, where it goes, I think you can you can sacrifice a little bit of that to just put that idea into a f into that format.

38:41And let the world do with that format what it can.

38:43So I thought about self-publishing something only to essentially produce a PDF.

38:53What if I just go down the street, but all we do is we make a PDF and we just put it out into the community and let people do what we want with it.

38:59And it was around the sketchnotes idea, you know, explaining Tableau and data to people.

39:02It's a format in the video that I think works well and I think it could easily translate to a book in lots of different interesting ways.

39:08And so I've parked that idea for now, but it's that that's sort of if I come back to that, that's exactly sort of where I'd go back to.

39:14And it would be just a

39:19You know what's brilliant is Amazon does this really well.

39:23True they do

39:24Publishing, but I don't know.

39:29Correct.

39:29Correct.

39:30Correct.

39:30Yeah.

39:31Yeah.

39:31Yeah.

39:31Yeah.

39:32It's such a weird niche in itself and

39:34Um, I I'm a sucker for print publishing as well because I used to work in magazine design.

39:39So, you know, obviously I'm sat here not just making videos, but I've I want to open InDesign and lay it up myself and

39:45do all the stuff that I used to do back then with it.

39:47And that that that that would then become a distraction because again, like one of the things you have to learn is to delegate, right?

39:54If you were gonna write a book, write a book, let someone else do everything else.

39:59Well and editing your own stuff's very hard.

40:01Like I guess Oh of course.

40:02Yeah.

40:03Yeah.

40:03I used to blog a lot when blogging was a thing.

40:06And the one thing that helped me with all this a lot is I can't remember like some famous author had a course or something.

40:11The best thing ever was

40:13Write, don't try to write and edit at the same time.

40:16Like there's something you have to learn.

40:17Like just write it, then go back and ever yes.

40:21I remember reading Brandon Sanderson, who's now like a massive author, right?

40:27Like Brandon ran a Kickstarter last year, the year before that raised twenty five or twenty-seven million dollars or something.

40:33Talking about self-publishing.

40:34But anyway.

40:35As a famous author even, he said before he gives his book to his publisher, he edits it himself sixfold time.

40:41Like Can you imagine and he writes thousand page books, you know what I mean?

40:45And then he does six and he goes, 'cause I've got to cut

40:48This like I know it's garbage, but not when I'm writing it, only when I reflect on it.

40:52When you when you reflect you come back to it.

40:54It's true.

40:55Yeah, yeah.

40:55Um I can relate to that a bit.

40:57I have made, I don't know how many times I've made a video and just decided not to put it up.

41:02I did this just last week actually.

41:04I made a video.

41:05What was it about?

41:06Oh God, what was it about?

41:08Oh, it was around it's about um tableau job market.

41:10So I've made this video three times now.

41:12It's it's it's I've never released it.

41:15Um and basically I was trying to explain to people that when looking for Tableau jobs

41:19Um, don't believe the hype and you know euphoria around, oh, there are no tableau jobs or their tableau skills are going down or tableau jobs aren't valued.

41:26I was basically trying to dissect that question into sort of different pillars and say, look,

41:30You know, it's fine.

41:32What you actually want to do is get rid of the term tableau from your skill set and just understand what skills you have.

41:37Go look for those jobs and you'll find a lot of tableau jobs, funnily enough.

41:40Right, right, right, right.

41:45Tableau in front of it, they put power BI and then they just you read the d details and it's like we use Tableau.

41:50But they want you to have both skills.

41:52Anyway

41:53Um I I do that, so I totally I sort of sort of resonate and connect with that.

41:56Um but yeah, no, I'll I'll have to reflect on sort of my authoring.

42:00Um

42:01One my big project at the moment is rebuilding my website.

42:03So I'm I'm I'm maybe I'll ask your opinion on this.

42:06So one of the things it's it's a bit of bit of a bit of a bit I'm I'm not sure if it's controversial or not, but

42:12I make a ton of videos.

42:14I have been successfully able to transcribe these very well.

42:17So I use something called OpenAI Whisper, and it can translate my videos.

42:23With 99% accuracy to the point where I'm only correcting maybe four or five words in like a in like a what five-minute segment.

42:30It's gotten very, very good.

42:31If you use the large models, it's excess exceptional.

42:34So that gives you a transcript of an entire video.

42:36And then what I've played with is taking that transcript, putting it into something like ChatGPT, and asking it to reformat it in the tone of a blog.

42:47Yeah.

42:47And so the ethical bit of this is

42:50What what would people think of that if that makes sense?

42:53So here's a blog edited by me, written by AI.

42:57That's sort of been the tagline that I've been like playing around with in my head.

43:00To give people who don't like the video format an alternative format for the video that I've just made, but it's been edited by me.

43:08And so this is

43:09This is the thing I'm I'm trying to like understand.

43:11On one level, I think it's it's actually quite a good use of this ethically because I'm I'm only reposing my own content.

43:17Your IP, yeah.

43:19Um

43:19But on the flip side, I can't be certain that that's actually what the AI is doing because obviously the AI gets its knowledge and its its its sort of way of putting these together.

43:28from surprise, surprise, all the other blogs that it's scraped.

43:32So I, you know, you can give it your content as a direction, but I

43:37You can't be certain that it's not getting that sort of way of putting things together and it would lack your tone and everything.

43:44So in editing it, I'd end up rewriting it a lot of it.

43:47And I've done this with two articles and I spent about fifteen minutes editing each and I thought interesting.

43:52I I I didn't know sort of where to place that concept.

43:54So maybe I thought I I'll ask your opinion on that, because you've written blogs

43:57You've written a book, but you also watch my videos.

44:00Yeah, no, I think so Yeah.

44:04Um I think it's brilliant.

44:06So here's the thing, I think.

44:08Um

44:09We could use a ton of metaphors on this, right?

44:11Like uh if you if i if you were a musician and you ripped somebody else's song off, right, that would be um

44:19That would be uh uh obviously very unethical.

44:23But say we roll back to say 1978 and van, I think 78, 76, Van Halen one drops.

44:31And um Eddie Van Halen's using this topping technique, and he's not the first guy to ever do it, but he uses a lot more than anyone else.

44:37I think if someone else other people started um

44:41Using a lot of like finger tapping technique instead of picking technique.

44:45Yeah.

44:46Um I don't think very many people say

44:49Um, that's not fair.

44:51A lot of people go, that's a rip-off, so I'm not gonna listen to it because he's ripping them off.

44:55But it's not the same as or sorry, it's not so much a rip, but like I could why wouldn't I listen to Eddie if I should listen

45:00Right.

45:00So the parallel here I think is stealing the flavor, I don't think the flavor of the way people write blogs.

45:09is their IP.

45:10I guess it's interesting.

45:12I don't think it's interesting IP.

45:13Like the very first chat time I came across ChatGPT was someone

45:19It's funny, it's similar to this that said, clearly they were giving a speech.

45:23I don't know if they were actually giving the commencement speech, but um on on uh on the pair the paros of helicopter pairs.

45:30And I thought this was brilliant because he goes, he says, write me a speech on the perils.

45:35So right away the AI knows, well this guy's not pro welcome.

45:39He goes, like a commencement speech in the voice of Martin Luther King.

45:45And I'm like, that seems okay.

45:47Like t as long as you tweak it, you know what I mean?

45:49But in this case it's your input.

45:51So I think the IP is your IP.

45:54And the fact that it it I think it's insanity not to do it.

45:58That would be like

45:59All these people who I make fun of who are like have to write their own custom SQL because they don't trust Tableau to do it.

46:05Like why would you use Tableau to do it?

46:09Anyway, that's a long-winded way of thinking.

46:10I I think it's

46:12I think it's uh a brilliant idea and it's uh it's completely okay IP wise from

46:20my standpoint.

46:21I think so.

46:25I mean it doesn't mean it's bro like you won't know if you put it out.

46:30I have no idea.

46:36And this is it.

46:37I've I've ten I've I've I've written ten um blog posts already and I just I've just haven't gotten round to putting them up because

46:44I'm waiting for the website redesigned so that I can more clearly I do want to be very clear that this is what's happening.

46:49I you know, I want to have the video above and then say, okay, everything below here.

46:55has been done in this specific way.

46:57Have a separate blog post that talks about the technique.

46:59Have the transcript and have the blog post available.

47:01So, you know, people can look at both and sort of um just just be transparent about it.

47:06But I I want to see how that goes first and then I'll I'll try with a few I maybe call it out on something like Clinton and see what it is.

47:17Is that you're obviously a very intelligent guy and time's the one asset um

47:24uh you know the one currency that we can't get more of.

47:26We all down.

47:27I mean I'm like I'm a big fan of the only good of money is to trade it off for time.

47:31Because time's the only

47:32thing that's completely non-fun.

47:33So you going back through your own blog post and typing the most insanity.

47:37Do you know what I mean?

47:38It doesn't make any sense.

47:40Yeah, exactly.

47:41Yeah.

47:41The the transcripts themselves were in in itself like a revelation, right?

47:45Being able to go back and transcribe all my videos, like something like five hundred videos historically transcribe

47:51And I'm slowly going through YouTube sort of automating the process of uploading the transcript.

47:55So if you've ever if you've ever watched a video of mine, transcript is huge because it also enables other languages, because Google can then translate that transcript.

48:02into another language automatically.

48:05So and then I think um Google YouTube is also working on audio dubbing on your behalf.

48:10So it will take your voice

48:12another language and dub over what you're doing and do the lip syncing so that it looks like you're saying the thing.

48:18Oh really?

48:19On your behalf.

48:20Yeah.

48:20Yeah.

48:21Well if you're gonna do that, you can't do it.

48:23So it's it's it's pretty it's pretty wild and

48:26Um that that to me is like there's there's like levels to these things.

48:30But anyway, we we get we we digress.

48:32Let's get back to tabooing.

48:33Uh I mean AI is part of the discussion, isn't it?

48:36We'll come to that in a second.

48:37So

48:38I think the one thing I want to touch on then is um we've touched on relationships, we've touched on data modeling.

48:45Um, you know, how would someone

48:48One of the things I always explain to people, and this is maybe where that data modeling thing comes back, is I think Tableau has lots of layers and it's quite a it's quite hard to discover

48:57the full capability of what Tableau has to offer because I think the history of how they've brought the features together has been essentially building on the past, right?

49:06So

49:07Data modeling itself is built on relationships, is built on data modeling.

49:11Data modeling is built on the connections window and how you set things up.

49:15And so it all comes through time.

49:17And so

49:18The thing I was asked Tableau, how does someone uh the example I give is how does someone know to look for an LOD?

49:23They have to do something in the product that gives you the cue

49:28that they're trying to achieve something.

49:30And I've always felt the Tableau haven't taken that cue seriously enough to be able to then point the user to say, hey,

49:38You're playing with a filter and this table quite a lot.

49:41You keep trying to go back and forth between this thing.

49:45Maybe you need an LED.

49:47Can we show you what that is?

49:49And so I'm trying to understand what is what is the equivalent of that for relationships?

49:53How does someone know, okay, I'm having a real challenge here?

49:58I really need to look at relationships because that's the sort of that's the thing the user doesn't know.

50:02So like how if you don't know, how do you know that you need relationships?

50:06Like what could you what could you tell people to say, look out for to say, hey, actually, I need to investigate relationships?

50:13Um that's a great question.

50:16I mean my flippant answer is almost it should be information.

50:22Because on Snowways now.

50:24There's a couple of really fringe ones.

50:26But um uh uh it's a it's a great question.

50:31I do think they should and it should work both ways, because I s more often than not I see

50:36Back to your first example, I see LODs that didn't need to be written.

50:39I'm like, that's in the view.

50:41Like, that's already in the view.

50:42You didn't need to write an LOD to do that.

50:44Right.

50:45Um, but um yeah, I

50:49I wish it would do things like this is what I wish it would do.

50:56So imagine you've got a product table and you've got an org table.

50:59Let's make it really simple.

51:01And then I drag a product on, right?

51:04And I've joined those together.

51:07And I bring a product on, and then I bring sales on.

51:11I wish it would the first time prompt you to go, you know that you might have products without sales, but the way you joined your data together, you can't answer

51:20that question.

51:21Oh yeah, yeah.

51:23Right?

51:23That's a great one.

51:24Yeah.

51:24Yeah.

51:25Because that's the most obvious one to me is that like your joins are

51:30Unless you're doing full order joins and creating a bazillion nulls, um and good luck on performance then, right?

51:36Um then you're you're always making these conscious decisions to not let business users be able to answer questions, which is

51:43Which is kinda crazy, right?

51:45So I wish they would say right, or even and I think I think Thomas and team, by the way, are pretty

51:52acutely aware if they wish they could do it.

51:54You know what I mean?

51:55So you talk about Thomas Nan, the product manager.

52:01And if he didn't, he gets

52:03Which I'm sure I'd love to hear he he gets 30-page documents from Jonathan Drummond making suggestions for which is awesome.

52:10Do you know what I mean?

52:11Like Jonathan really thinks about this.

52:13stuff right of course yeah yeah in in a good way um and so i you know almost so what i don't know how much you've had a chance to play with multi-fact yet

52:23But it's a little dumb right now, but they know when it's going to get better in that.

52:28Not the model itself's very smart, but when you pull on

52:31If you pull on something from a table that doesn't have a direct relationship between what you have on it, it'll warn you.

52:36Yeah.

52:37Yes.

52:37But but um but sometimes it can make it.

52:41So the warning's not that smart right now.

52:42now so it'll get better.

52:43But I wish to go the next step and go uh to your point would be you know you drag something on say take my example from uh

52:52uh TC that simple one, right?

52:54Yeah.

52:55Maybe I go sales, um, sales, profit, discount, right?

53:00Boom, this puts unlike, and then it goes, I wish it would then go, hmm.

53:05You know, this might be inventory.

53:09Because they have the tech to do it, which is interesting in Tatlow Einstein, they showed this more front and center a little bit, but they've had the tech to do outliers.

53:20for a long time, but it's so buried in the product.

53:23Like the explained data's was my favorite feature for a long time that no one could find.

53:29So that that exists already.

53:30It's not even Gen AI.

53:32It's like it should bring the it should bring those forward, right?

53:36And the other thing is they call it predictive analytics, which it also isn't really.

53:41Correct.

53:42It's but it's outlier detection is what it is.

53:44And outlier detections

53:46incredibly important because people spend around spend their day dragging pills around and filtering trying to find outliers.

53:53Yeah.

53:53They've got this technology that could just surface outliers for them.

53:57So I wish they brought that tech together with

54:00So those are the things I'd like to see basically and be these UI gestures of right and and things like LOD would fit into that.

54:07Do you know what I mean?

54:08Are you actually trying to aggregate this number?

54:11Yes.

54:12Yeah.

54:21What's the word I'm looking for?

54:24Philosophy's not the right word, this close, and it'll come to me.

54:27Eat itic maybe, you know that sus.

54:29That like that you know what you're doing

54:32Like the products built on what you're doing.

54:34So all this interrupt and it's like in the flow, do you know what I mean?

54:38Like the tabloid flow is a big thing.

54:39Yeah.

54:39And now that would have to

54:41That's why they're very anti-wizard.

54:42You know what I mean?

54:43They got acquired by Salesforce.

54:44It must have been Curring because it's the wizard company of the world.

54:47Do you know what I mean?

54:48And then uh was like anti-wizard

54:51right so because they don't want to interrupt the flow so yeah these things would all interrupt the flow but I think I think um they have to get over that for the it's easy enough to have a dismiss do you know what like

55:03I got that.

55:03You know what I mean?

55:04Don't remind me anymore.

55:05But uh but those would just be brilliant.

55:07And they could do it.

55:08I don't think it would be that har I don't want to be flipping to the UX people.

55:12But it wouldn't be that hard for them to get the trigger to know.

55:15Like how they interrupt someone takes a UX expert, but but the but the understanding to do it.

55:21Because um

55:22The message is a good idea.

55:24People don't believe things.

55:25Like I spend a lot of time showing people what it's actually giving them in their model.

55:29Right.

55:30Like uh like uh and this would help with my other thing, which is I have this dream that I wish people that were looking for tableau people would say.

55:37understands data really well, can't write a line of SQL to save their life.

55:41Because they're because they're always like they want advanced SQL skills, but they never ask again.

55:46understand data.

55:47Do you know what I mean?

55:48Like trust it to write the sequel.

55:50Do you know what I mean?

55:51But um but you have to understand your data, right?

55:55Yeah.

55:55But you can make it a little easier to prompt you.

55:58Like for sure.

55:59And it's interesting.

56:00Tableau's always taken a view of I think you you nailed it there with like you're not stupid, you know your data.

56:07And sort of staying out of the way.

56:09And it's funny because one of the intrinsic parts of every authoring experience is something called show me.

56:15And I always feel like like here we are with Gen AI and Einstein co-pilot.

56:22And in my head I'm like, but you already had show me.

56:25Like all Show Me is missing is pairing how to do something with intent.

56:31So you always had a place that could have actually transcended a little side panel.

56:35It could have gone throughout the whole entire product.

56:37Show me how to use an LED.

56:40Show me how to build a chart.

56:42Show me how to connect to a data source.

56:45Show me how to build a data model.

56:47It's like right there, but it's stuck in the authoring experience.

56:51And that's the problem, because the people who really need Show Me are the people who engage with the analytics, the people who need the outlier detection, the people who need to drill down and don't know how to and need to build their own analysis, or people who need to get pulse.

57:05Set up a new filter and then get that, you know.

57:10But obviously the Einstein brand is super strong and you know they have to go with Einstein.

57:15But to me, Show me just sits there doing nothing for years now.

57:19And it's like just just pair them up.

57:23And copilot.

57:26And ISIN Copilots also, and it's not just them making what I think is this mistake too, which is I think the biggest reason people can't use Tableau today

57:36Um is not the technologies because they don't know how to frame good questions.

57:40Like all you have to do is watch people use Google.

57:42Like it's embarrassing.

57:44But it is what it is, right?

57:45So Einstein co-pilot

57:48assumes that people can't drag and drop pills.

57:52But they can answer good questions.

57:53I'm like, the good questions is the hard part, not drag and dropping pills.

57:56So Einstein copilists will be sitting there going, maybe you should drag this to run.

58:01Or maybe you should do this to create an LD.

58:03It should be like 180 degrees from what it is.

58:06And another thing you reminded me of is this this goes back a long time.

58:10I remember

58:12We've found the big banks in New York when I was working at Cognos.

58:14And every BI company's made this mistak if he was right this mistake ever since, which is

58:19He goes, you're pricing pyramids upside down.

58:22He goes, you should be giving me developers for free.

58:25Because if they're good, then I'm going to have all these consumers who consume it, and I only want to pay you if people are getting value out of it.

58:32But you're making it hard for me to create stuff because that's the expensive stuff.

58:37Whereas if I knew people were getting insights and doing that, like I'd pay a lot of money for that.

58:43And I think as a result

58:45All these features that consumers need or the business people who you know are technically inclined, they don't get because they're the low end of the price point.

58:54But that's hardly Tableau's fault.

58:56Every single BI vendor

58:58from the beginning of time.

58:59Like in Cogmas we were way worse.

59:01I think we had nine or eleven different prices.

59:03Yeah.

59:04And like you know the person who created the data model would be like $2,000 and the report author would be $700.

59:09And like coming down, do you know what I mean?

59:11It was like

59:13Such a fragmentation, yeah.

59:15I know it's a tricky problem to solve, to be fair, but Yeah.

59:21I'm I'm interested to see what companies like Sigma can do because

59:25In many ways, they uh they say themselves they follow in the steps of uh giants, right?

59:29So they they they have the benefit of hindsight in certain respects to be able to see where the industry is going and also um what

59:38How they could reconfigure the way things work to sort of work better.

59:41Um I I I you know, you know, full disclosure, I think they still have a lot of work to do in some areas.

59:47Sigma claims to be a platform, but it's not as fully featured platform as Tableau.

59:52And some of the simple things like you know not being able to connect to Flatfile still huge hurdles to lots of companies who just need to be able to, you know

60:01Get this Excel to connect to that Google Sheet to connect to that Snowflake table, right?

60:06That's still a huge use case for so many people.

60:09I haven't made up my mind on this yet, but the one thing I have to give

60:12like the Tableau founders a lot of credit for and they get a lot of credit for is they built something it was very Apple-like and they built something that no one was asking for.

60:23And and I know Henry Ford never actually said this.

60:27But I do wonder if Sigma are trying to build a faster horror.

60:30Like they're giving people what they want as opposed to giving people what they need.

60:34Maybe.

60:34I don't I haven't made up my mind on that yet.

60:36But I know they're like but like

60:38Like even when I'm a demo on your channel, there's a lot of well I'm a lot more comfortable starting with the data and you know aggregating it up.

60:46I'm like, well you might be, but is that really

60:52Microsoft built what right now might be the biggest company at any given day or the biggest company in the world by market cap.

60:58I'm doing nothing but building faster horses.

61:01Like that's Microsoft's whole

61:03Reason for being and they admit it.

61:05You know what I mean?

61:06Like they're cheaper because they're not the best typically.

61:08You know what I mean?

61:09But they're good enough.

61:10They have the scale, they have the connections that you can do that.

61:14Unless Sigma's extra strategies getting bought by Snowflake or something.

61:17Again, I'm I haven't spent enough time with Sigma to know if that's fair, to be clear, but I see a lot of faster horses products.

61:24Because it's hard to build the other product.

61:25Like it's hard to get it right.

61:28To to gain traction as well, you have to have a differentiator.

61:31And I I think they have the right differentiation.

61:33I think that like, you know, that bottom-um's bottoms up approach is

61:37The reason it resonated with me is because I think the point you're making, but from the other end, which is that it that feels to me it's like something people can understand.

61:46And I'm talking about the the the you know, everyone else, the tab I keep talking about more recently, you know.

61:52Not your developers, not your creators, not your data models, like everyone else in the business.

61:56People using Pulse essentially.

61:57And it's interesting Tableau have come out with Pulse.

62:00Um Altrix Auto Insights was the first time I saw something like Tableau Pulse Ultrax.

62:05uh you know, have their own challenges over the years.

62:07And uh they came up with auto insights out of nowhere and I was like, ah, this is actually really good.

62:12And I was like, a shame it's buried inside of the Altrex platform, because otherwise it would be great.

62:18Then Pulse came out and it was it was kind of a take on that, but actually very different in in the long term.

62:23I've ac I used to say they're the same, actually I've realized they're not the same at all.

62:27They they have a very different

62:30What's the word?

62:31Outlook um from the outset.

62:33Although they feel the same.

62:34They have a different outlook.

62:35Right.

62:36I think that spots a lot like Paul.

62:38Yes, yes, correct, correct.

62:41Absolutely.

62:42And it's interesting because the thing about ThoughtSpot is I feel like they're building

62:47They're starting with Tableau Pulse and heading towards where Tableau is now.

62:52And then Tableau feels like it's it's already got the platform, but it's it's it's it's sort of branching out into lots of these

62:59different places that it's it's not served in the past.

63:02Um everyone else uh being included.

63:04J just to close on Sigma, because I do think one place they could kill it, but I'm not sure they're getting this.

63:10Right.

63:11And I could be wrong.

63:12I mean they're in the device.

63:14I keep saying people keep saying the reason people use Excel.

63:17I think the peop reason people use Excel

63:20is to add more than lots of reasons, but the biggest is let's say I'm looking at everybody wants to do light what if.

63:27Like everybody wants to go, well what if I increase our sales headcount by five percent?

63:32But I don't want to

63:33But maybe ten percent of this region.

63:34And it's easy for like a non-technical person to do, right?

63:38And everyone knows why people don't like Excel, that's not governed, it's like whatever.

63:42If they could lean more into the right

63:45and give like an Excel server kind of feel that still had the chart instead of trying to compete with the tableaus of the world.

63:52Yeah.

63:52You know the problem with traditional BI is it gives you a lot of answers, but you can't model a future

63:57based on it.

63:59Future based on it.

64:00And we've got this high scale, you know, on top of scale.

64:03Everything else around top of it.

64:04Like executives want that, like for sure.

64:07Because we we tried to build it at Cognos with TM1 and Cognos

64:11too set like a planning product and but the yeah they were too monolith you couldn't get them together but i was always like if i started a bi product the single biggest differentiator i would have

64:20is right.

64:21I was on Francois about that all the time.

64:23I'm like, why can't we get right back to hyper?

64:25Because people want to do and he's like, there's lots of budgeting and planning.

64:28I'm not trying to replace a

64:30like a full on ERP or budgeting system.

64:32It's just people want to be able when they analyze something a lot of it to be different based on they don't want some crazy predict they might say they want some crazy predictable

64:41thing but what they really want is just to be able to type some numbers in for formulas and see what impact it has.

64:47And Sigma has that.

64:48I just don't I don't know how much they're putting that for I think that's the thing.

64:53And and to bring that full circle, I mean, if you think to Tableau, that to me feels like the perfect training set

65:02to really make the AI problem work.

65:05Help people ask better questions, right?

65:07Because if you can if you can see what they're trying to do with their data, like what they're trying to model, what are they trying to do?

65:13Yes.

65:13Um do what that is essentially doing is creating the the connections in the background to help enable better questions.

65:21And that's sort of the big problem with Ask Data in in the past is that you've had to go in

65:25And do a lot of heavy lifting to create lenses and then yeah, you go in, you write the questions, and the questions are really just ways to create charts without having to drag and drop.

65:35But with AI and with Pulse a little bit, I think they have a bit more of that.

65:39Th an opportunity at least to be able to sort of pair those two things up.

65:42But anyway, I we we we kind of sit here and um I'd say I call it wish casting when you kind of

65:48Uh us us as uh onlookers of the product, um maybe maybe maybe me more than you, because you've actually worked at these companies, but me more than you

65:56Wishing, wishing, wishing everything we'd love to see and uh not not not not being in a position to be able to like uh get it going as it were.

66:03But um I'm sure Tableau get it all the time.

66:06Um for sure.

66:08In terms of um in terms of sort of the future then, um I'm sure you've you've looked at Tableau Einstein, you saw the you saw the the big appeal.

66:16And I I think

66:17I put out a video um I put it out last night, but it feels like it was today.

66:21I put it out last night just you know saying I'm not actually sure what we saw.

66:25And and it's it's weird because you see stuff.

66:28And you you kinda have to allow a bit of time and I'm a bit guilty of this sometimes.

66:32I see something I immediately somewhere and I immediately get my thoughts.

66:35And actually

66:36You maybe people have noticed this with the most recent release of 2014.

66:40I've not made as many videos on multi-fact analysis or that stuff yet, because I'm still

66:44I'm still assimilating and I I want to be able to come back to it once I've really crystallized what it is.

66:50And I I kind of think that 24-3 might have

66:55the other shoe drop for multi-fact analysis and and the a few other things.

66:59And so in so in trying to teach them, I think that would be a better package to say, okay.

67:04So 24-2 has come out, then 24-3 has come out.

67:07Now I think you can make the video that really introduces the topic because the two things that need to be get there are right here together and you can kind of get it through.

67:15Um but anyway, that's the that digression.

67:18Um

67:18Where do you think, yeah, where where do you think, you know, Salesforce is is going with this Einstein sort of capability?

67:25I know that's a big grand question.

67:26I'm asking you to sort of guess the future a little bit, but from what you've seen and your sort of understanding of

67:32you know what's happened at Tableau in the past, but how things are going.

67:35Yeah, what do you think?

67:37So I think what we're seeing is a completely different product.

67:44Yes.

67:44And Tableau's going to continue as is.

67:46We're not seeing replacement product.

67:49We're going to see Tableau Legacy, whatever it's going to be called.

67:54Francois's term compared to C when it used to be called Tableau CRM or whatever, used to call it Tableau OG, which I kind of like.

68:00So it would be Tableau OG.

68:01Tableau OG, yeah.

68:02And then and then we're gonna see what I think will eventually be called either analytics or Einstein cloud.

68:09And let me share an anecdote first, I think, applies.

68:12I yeah after going to T I was at Tebo for two or three years and there's a guy that used to be on my team at

68:18Cognos went to SAP and got a big job, you know, VP job in their BI space.

68:23And I'm like, it's interesting.

68:24I go, we never bump into you anymore.

68:27And he goes, yeah, but he goes, I bet you don't have very many SAP customers anymore, do you?

68:32And I'm like, yeah, we don't know that you mentioned it.

68:34He goes.

68:35Yeah, we got sick of you guys kicking our butts.

68:38Like we couldn't compete on it, like Bob J was old, like you were killing us, right?

68:43Video, what we realized was

68:45If we embedded analytics and SAP, they weren't looking for some PurePlay vendor anymore.

68:51Right.

68:52And then and Oracle did the same thing.

68:54They bought like Hyperion, they bought a whole bunch of

68:57companies right and there's OBI and you I I think you can still they certainly support I think you can still buy business objects and yeah Oracle business intelligence right definitely support it but I think

69:08Minimal like you know, they'll continue to put new connectors and that kind of stuff.

69:12They'll keep it alive, yeah.

69:13Yeah.

69:13And then what we saw the other day, I think.

69:15So what they probably tried to do for the past four or five years, and I don't really have inside information.

69:20I've snippets, but I don't have the snippets.

69:22Yeah just to be clear.

69:23Um is Tableau became a monolith, but because they always do.

69:27Like I did think I commented on maybe something this they go I think we're both on the same th thread on LinkedIn where someone goes, Why can't they just release

69:35the fix for like hyper for multi-fact when it wasn't a temper one like because it's a monolith they can't just drop they're not like Amazon where they can just drop code on Amazon.

69:44Yeah, yeah it's like show up right so

69:47And I think what they did is they tried to embed it in Salesforce, and it was probably way too hard.

69:52And then what they probably did instead

69:55was went, but you know what, the Tableau UX is pretty good.

69:59It would be easier to rebuild it than it would be to try to put it in.

70:03So I think what we saw was completely new.

70:05But based on

70:07A UX that's based on the IP, yeah.

70:10And then I think it won't need something like prep per se, because I think every bit of data is going to go through data.

70:18So you will be able to bring in data breaks or whatever, but into data cloud.

70:23Right?

70:23And and hyper, I think they said this.

70:25I'm sure hyper still be there as an accelerator.

70:27So you won't be modeling in hyper, but hopefully automatically.

70:30It'll pull data you want, edit data cloud so it's faster, right?

70:35And then and it makes sense to me actually because

70:38If if I'm using say operational BI versus say strategic decision making, whatever you want to call, those words get overused.

70:45But say operational for sure.

70:46And then the other, I would say what what stuff like Tableau is good for.

70:50um is making decisions about the future.

70:53It's not that good for oh I just saw something take action.

70:55Like you know they always want to give that demo.

70:57It's not part of uh so for those type people, why wouldn't it be

71:04And then uh yeah, I so I don't think Tableau people have to panic or anything because I think Tableau's continue.

71:11Right?

71:12It will get as much investment

71:14as the people buying it allow it to support.

71:18And then people who use Salesforce who it's have a low percentage still that used to have low in the flow.

71:24Like it's almost like a new market for them

71:26them.

71:26So yeah, I could be wrong.

71:27That's what I think we're seeing.

71:29The only thing that's perplexing to me is why they're pitching, if I'm right, is why they're pitching the Tableau community so hard in seven cells

71:37Correct.

71:38It's the part I don't underst I can't wrap my head around.

71:40Do you know what I mean?

71:41But I wonder if

71:45They need that community to travel into the Salesforce ecosystem to use these things.

71:54Because they recognize that the Salesforce ecosystem doesn't have a body of people who have the heritage in doing this stuff, know the interface, and know all of that stuff.

72:03And so the reason they're pitching it to us is because they do want to get a bit of that rub off.

72:07Hey, you know the product, you understand it.

72:09Let's get you excited.

72:10Let's get you empowered to be the people that really know this stuff.

72:15And we can take that community and bring it into the self.

72:20It's it's it's it's you know, um I remember when um

72:24uh Mark Benioff after he'd finished making dolphin noises to Adam Solipsky on stage.

72:29Like he said that the Tableau community was, you know, one of the biggest

72:34Value propositions of the purchase.

72:37And I always think back to them.

72:39I think I think he meant that in in multiple ways.

72:42It wasn't just like community data fam, you know, look at look how much energy is in here.

72:46You also meant like the literal people, the skills, the skills that community has, and their ability to solve

72:54in my opinion, monumental problems in the Salesforce data system.

72:58Because if you've ever worked with Date Salesforce, it's like a mess.

73:02Like it's a

73:03It's so hard.

73:05Right.

73:06Well, actually and you know I can't believe I didn't think of this.

73:10But I think you're right.

73:10I think that's probably good.

73:11So as an example, I've got a really good friend who's like a

73:15Three or four time Salesforce MVP.

73:18Um the amount of Salesforce he knows blows my mind, like different Salesforce components and stuff.

73:23Yes.

73:24And about a couple of months ago, we got talking about data cloud

73:28Which I have a way better understanding what it really is now.

73:31I think I'm coming for my certification the next couple of weeks.

73:39But just to listen to him talk about it and how wrong he was, right, on his assumptions, because he doesn't understand data and databases from uh at least from an extracting data, you know, putting data in.

73:51Do you know what I mean?

73:52Yes.

73:52And I'm like

73:53He could run circles around me on like how to build a lightning web component, what security's like in sales for obviously, but I because he's in it, but he's also natively very smart at them like they do need people.

74:04Like they will need people, like beyond just Tableau, but like data people, literally.

74:09I think you're probably right.

74:10That would make sense someone for pitching it so hard.

74:13And that's why it has to be called Tableau, because without it being all that, the people would never go across.

74:19And so that's, you know

74:21I think that is that is it's an interesting balance.

74:25The other thing I don't understand is it feels like they have a schedule.

74:31And they have to hit certain notes to that schedule.

74:35I don't know if you get sort of get this.

74:36It's like Dreamforce is on this day and we will show you this and then

74:40the webinar that we just had, it in my opinion, it was no different to the conference talk, but it felt like they needed to do something in a specific window.

74:49To get ace like something out there and I think it will make sense in hindsight, but yeah, like If do you know the term PayPal Mafia

74:57About how so there's this term about people in the valley when it came to certain types of companies.

75:04They called them the PayPal Mafia because they were all in the executive team at PayPal.

75:08Right.

75:08So Edon Musk, uh, uh Peter Thiel, uh Reed Hoffman who went and LinkedIn.

75:17So these guys were the PayPal mafia

75:20My take on this is there's no one talks about it, but there's almost an Oracle mafia, which is all these execs that came out of Oracle.

75:30First, learn how to sell something before they build something.

75:33So Benioff is very and it doesn't mean he's not intending to build it.

75:38But he's like the talking about it at like Dreamforce is like Christmas.

75:43Do you know what I mean?

75:44Like we need something to dream force

75:46And then I need to show something.

75:47Was it you that said or someone said I heard, I thought I heard a GA date of aiming for February 1st.

75:52Maybe I made that up, but that's the start of the fiscal year.

75:55Do you know what I mean?

75:55Like so everything's very and like and we announce it before it's ready.

76:01Yes.

76:03It's a very

76:04It all came out of Oracle.

76:05Like Siebel was like that because Tom Siebel came out of back in the day.

76:09Like so and and you know, I think Mark was the youngest VP ever at Oracle or something.

76:13Like they had this

76:15Like so it's almost like an Oracle mafia to me, which is this thing, which is like, you know, announce product, hype it up before it's ready.

76:23Yeah.

76:23Which which culture wise was very because I'm I used to be very frustrated with Tableau, because they were the other extreme to that, which is

76:31Oh and it's available now.

76:33Like they wouldn't let anything before it came up.

76:35They were completely the other part.

76:37Yeah, true, true.

76:39It's part of why I was out of my channel, because I I'd be like

76:42go to a client and they didn't know something had happened and I'm like, no, it's it's just all here and yeah, yeah.

76:48Now it's the opposite.

76:49Yeah.

76:50But you can see like I would say data clouds

76:53only real for what data cloud's supposed to be about now to be honest and then talk about data cloud for a long time.

76:58It used to be CDP to get your um for marketing cloud to get because like you know as a marketer you'd be like

77:05they'd be like run a campaign to like our top customers.

77:07I'm like, who are they?

77:08They're in Salesforce, they're in this, they're in this, and they needed a way to consolidate customers.

77:13And then from that they learned, well, why don't we consolidate all our data but around customers?

77:18So it's smart.

77:20It just takes a longer to build software than people that are in software seem to acknowledge.

77:25It's like trying to build a

77:27Bridge under budget or something.

77:28It doesn't have to be much.

77:30Yeah.

77:30And I've also noticed a bit of an influx of people who left Tableau, you know, during the layoffs or whatever.

77:37Very notable people coming back as well.

77:39So I wonder if

77:41Like, yeah, I j I just feel like there's something going on.

77:43And I I think this time next year we'll sit back and we go, oh, okay, this was the play all along and they obviously can't tell you their plan.

77:50It's a company, they've got people to please and money to make.

77:53They want to help their competitors along the way.

77:55But there is something going on.

77:56Um But I think they probably want to keep their options open and they don't want to hurt

78:01Revenue strains in the meantime.

78:03Do you know what I mean?

78:03But Correct.

78:04Yeah, it's a delicate ship, yeah.

78:06But I missed what you were saying, which makes complete sense to me

78:09I mean the Tableau Consulting Community, I think I have this number right.

78:13Should be happy about it.

78:15I don't I never figured out, but for every dollar a Tableau sold.

78:19Like there's not a lot of consulting sold, right?

78:22For every dollar of sales force that's sold, I think the number's six dollars and nineteen cents the consultant.

78:27I know

78:28For sure.

78:29I hope I'm not making that number up.

78:30I know for sure when Lou Kirshner, who ran IBM and turned it around and then left, in the notes at the end of the book, he took shots at a lot of people.

78:38And he said, I'd like to thank Larry Ellison and

78:42Whoever's running SAP at the time, but you know, Larry Ellison and Oracle, he goes, because for every dollar of software you sold in global business services, we made seven dollars.

78:51Like it's it's a much more lucrative

78:54Consult Salesforce system because Tableau's because it's kind of easy to use and spam a lot of audiences.

79:01Yeah.

79:02Right.

79:02And the community is so powerful.

79:03Like it's not a great

79:05And you see all these tablet consulting companies get pulled into doing Databricks or Snowflake or something, because like the bigger, media kind of.

79:12But the Salesforce ones are

79:13But that's gonna be meaty because there's a lot of process in it.

79:16Like it's not a flaw in the product.

79:18It's like there's a lot of like um

79:21Not only integration work, but there's a lot of management consulting that goes into that.

79:26Do you know what I mean?

79:29I wouldn't be scared of this at all.

79:31I mean, being scared of a technology change is just

79:33Ridiculous thing anyway, because it's inevitable.

79:36But uh but I think this could have went a lot worse than this.

79:39I mean they could have said it wasn't what we thought and we're just gonna let it wither and die in the vibe.

79:43Yes, um, sold it off to like uh I don't know.

79:47It's only it happens.

79:50Right feeling?

79:51Well, the word that sell it off to some um what's the word I'm looking for

79:55You know those investments.

79:56Venture capital to squeeze it out.

79:58Yeah.

79:59Yeah.

79:59Just to like squeeze every penny but not in bet like.

80:03I think that's what's happened to Ortrix.

80:05Like I am fairly sure that's what's about to happen to All Trix, yeah.

80:09So we should be happy that didn't happen.

80:11Do you know what I mean?

80:11True, true.

80:14And yeah, you know, yeah, I'm I'm I'm an optimistic with with with change.

80:19You know, I I I do I do get the concern that people have, you know.

80:24Um people put a lot of time and energy building profiles and personas, but you know, that that's sort of like a a small subset of the whole landscape.

80:32And I think

80:33A lot of people who use Tableau today use many tools.

80:35They're not just wedded to Tableau and um they can and showcase frequently they are able to adapt and learn new things because new technologies that work with Tableau have come and they've learnt them, right?

80:46So

80:47Well my my advice to the community should be whether you like Salesforce as a product or not aside, their community that's the one thing for Synergy for sure.

80:57Very, very powerful.

80:59community and yeah and they have gotten that their success is in having all these MVD and I I went to Wondering Force and it is amazing in how many stories are around.

81:10people that were I don't know they were working in some other complete industry and through Trailhead they taught them to do this and they became a sales force of men and they're making 5x of what they used to make and I'm not sure how many of these stories are, but just the fact that there's any of them

81:23Do you know what I mean?

81:24Like how many companies like you don't see a lot of people who are tending bar one day who's an Oracle VBA three years.

81:30Do you know what I mean?

81:31Like it's kind of

81:33unique and cool.

81:34So I yeah in that way at least they can't you know you can I can't knock their community.

81:39I think it's great.

81:40Yeah yeah yeah

81:41Yeah.

81:41I'm excited.

81:42I'm excited.

81:43I think, you know, five years from now, um, we could be at the precipice of some really, really exciting sort of um opportunities with Tableau because

81:52At least this this new change they're pushing down allows them to take on I think things that have been on the cutting room floor at Tableau for years, like decades.

82:01We're only just seeing some of that stuff.

82:03actually get solved.

82:04And I think I always highlight that look, if you look at the last two years and some of the things they've they've solved, those those have been on the floor for a long, long time.

82:12Never made it onto a to-do list.

82:14And

82:15They're getting they're getting looked at, they're getting done.

82:17They're getting delivered.

82:18So um okay, it might be to synergize with the Salesforce platform, but nonetheless, it doesn't matter.

82:24Well the another thing Salesforce in Tabla

82:27share is like a clix knife code kind of of course it's a little bit of code but mostly they're in different ways.

82:34One's Dragon Ross, one's wizard, but most of them are.

82:37And a lot of the things Tableau needed to do

82:41to extend their ecosystem if they didn't go this way required code.

82:45Yes.

82:46And like it just doesn't feel right.

82:48Do you know what I mean?

82:50Well I'm even thinking just to integrate into someone else's workflow or something

82:54Like it's not as easy as people think it is, right?

82:57And and that's a Microsoft I keep saying anyway that the reason the people who like love Power BI always say they probably came from a developer background.

83:05Like I'd someone

83:06Open in front of me again today, I'm like, it's like I'm staring at Visual Studio, man.

83:10Like it just the feel of it even, right?

83:12Like it's familiar to the Microsoft people.

83:15Yeah, so but for them to build integrations, like a lot of them are probably more true developer focused, right?

83:22But in the Salesforce world, it's like

83:24Yeah, if I'm going to take action on something I see, they have an action framework that I could hook into if was built into the same UI that I wouldn't need to be a like a hardcore developer to

83:34write.

83:34Do you know what I mean?

83:35Like I'd be more like the same level of skill of it.

83:38Technical level of skill of a tabloid person, which I would say is uh

83:42um you know, a very smart person that doesn't have this huge syntax or maybe desire to sit in an IDE all day long.

83:51Do you know what I mean?

83:52Like

83:52Like I've written code for like a lot of years and I hate it.

83:55Do you know what I mean?

83:56Just because it's so inefficient of a thing.

84:01Yeah.

84:02Gosh, yeah.

84:03I was so much, so much to look forward to.

84:06I'm I'm I'm hopeful.

84:07I'm hopeful, I think.

84:08Um I'm I'm I work in consulting as well, so I always do wonder if our perspective in consulting

84:15Kind of skews our perspective on the product a little bit because I think if you work in a company and you just have one stack, right

84:22You don't intermingle so much with um some of the other ideas coming from different different platforms, different ecosystems.

84:29Yeah.

84:29And I wonder if that's where the concern comes from, right?

84:32Because you use Tableau.

84:33This is the one of the big things you use.

84:35And actually the product is telling you you're gonna have to change it.

84:38And that that can kind of feel like um not an attack on your character, but given the amount of energy people put into Tableau.

84:45When the software is telling you no, you need to change.

84:48It's kind of like you're staring at a mirror, right?

84:50Whereas in consulting, we're always used to change.

84:53Like our clients are telling us to change, you know, we're always wrong, right?

84:55Like could be right.

84:57I mean, it is the biggest

84:59It the biggest parado I think the biggest it'd be one of the biggest paradoxes of like our species, which I don't get is how we're in a constant state of change, but most people hate change.

85:08It's the interest.

85:12Absolutely.

85:13Gosh.

85:14Um what else have we not talked about?

85:17Um I don't think maybe that we didn't talk about because

85:22We got onto that which was great was just this like with the near-term future, right?

85:29Pulse and Einstein co-pilot and things that are gonna affect people today.

85:33Yes.

85:34Um and we talked about a little bit with pulse

85:36You are like so I used to say to people before, if you had a bad data model and Tableau, but you were really good at writing calculations.

85:45Forget about you'd have an inefficient workbook, but again, if you didn't have a lot of data, you could viz your way and calc your way out of a bad data model.

85:54Yeah.

85:55Copilot and pulse are useless.

85:58And and everything new coming without a good data because they're they're implying a good data model.

86:05It's inherent in the it's a s it's a built-in assumption into the way they built the product, even the interface, even the advanced features in pulse.

86:16Actually require you to have a good model because, and this is this is the bit I don't understand, is you can't see the model in pulse, right?

86:24You you could build a fantastic model.

86:26And you can add calculations, but you can't see the model the way you can see it in the relationships, you know, connection, right?

86:35You can't

86:35Sorry, I'll cut you short.

86:37No, no, sorry, I didn't mean to cut you up.

86:38But don't you know the other thing that I say to people that it takes a while to get around is you unless the new ver I haven't played with the

86:45No, no, no, it's really supposed to be so yet I should look.

86:48Um But all the calcs that you write in that awful interface, which I almost wish they didn't give you, and just say you have to go back to your data model, right?

86:56They have to be row level calculus.

86:58Yeah, you can't do table calculation because they're gonna try to aggregate it and do it over time, so it's gotta be row level, right?

87:04Like they don't but they don't express it that way, right?

87:07So an LOD will work because it's always a row level

87:10Right?

87:10Yeah.

87:10But you can't, and people are like, well, what?

87:12I'm like, there's no way they could do a table calculation though.

87:15Again, you can't table calculate edit this problem because

87:19That's an in-memory thing and it's loading directly against the model, right?

87:23So people like those concepts people

87:26kinda have to and once you do, it's a ma do you know what I mean?

87:29Like so ever in a workshop in Canada for this like a virtual workshop of a video on like in fifty minutes I show how to use pulse properly and like forty eight if it's the data model where to secure it.

87:40How I would

87:41How I would almost always create brand new models, because otherwise people are going to troll through fields you don't want them tro trolling through.

87:49Hide a bunch of fields to republish it.

87:51You know what I mean?

87:52Put it on the schedule, put it in its own project, secure that project for the people you want to see it.

87:57Boom, you're gone.

87:58You know what I mean?

87:59Maybe an entitlements table, bam.

88:01Right?

88:02Um

88:03Uh quick.

88:03So and I think but the thing that people don't realize is uh Tableau's actually always been that way.

88:09It just people fight it with crazy calculate.

88:11Like Tableau expects

88:13Like I don't think people think about this enough.

88:15Like when you connect to data, Tableau doesn't give you an opportunity to anything other than to define one of those feet, like you know, every column is a field.

88:24And it's going to try to aggregate that field.

88:26Right.

88:27So I say to people all the time, if and it's in the book, of course, if you have aggregate if you have like if you have a field called measure name and another one with value, basically, or metric name.

88:36And you conditionally say if it's equal to this, then return that like Tableau does not perform like some other BI tools are fine with that.

88:44It doesn't, right?

88:45Tableau doesn't make it, yeah.

88:47And even behind the scenes, what they were smart about is I think they're leveraging things like if it's in a column, it's an array, right?

88:53And so CPUs are set up for that.

88:56So if you say what's the median value, it knows it doesn't have to find it.

89:00Right?

89:00And if you say maximum, it knows because it's stored it.

89:04So it's off the top and pick off the bottom, right?

89:06So it's not running through that.

89:09And it's why some mistakes are so bad.

89:11Do you know what I mean?

89:12Yeah.

89:13Yeah.

89:13Yeah, it is true.

89:15And again, if you use relationships

89:17You don't need to use count distinct anymore, which people don't think about, because Pablo's smart enough to count it in the dimension.

89:23Correct.

89:26That's interesting, yeah.

89:27And and the last thing I would say is

89:29Is I'm not sure how good Pulse and these products are going to be for your typical business user.

89:36I can't believe this hasn't been a bigger thing in the tabloid community before it came out.

89:40is I want to see relations.

89:42I'm going to want to see correlations.

89:44I don't even think about them as correlations as a business user, but I want to see correlations across different parts of my business for sure.

89:50And Tableau's never been able to model that before.

89:53Like like you know, like we're discounting tables in the East, a lot of the time I'm like, people probably knew that.

90:00Do you know what I mean?

90:01If they didn't know that, I question, you know what I mean?

90:03It took them a long time to find it, but you showed it quicker.

90:07But

90:07You know, if it's because we don't have enough inventory, so you know, our support calls are up, like people have no idea.

90:13That's what they do not know.

90:14Like I would say Tableau's usefulness went up by 10x in real world when they added this

90:20Multi-facting.

90:21It's that big.

90:22I keep saying that people are like, oh you're hyperbole.

90:25Yeah, yeah.

90:26I know if they use it right, it is because as a business user, I want you to find correlations that I just couldn't find.

90:32Right.

90:33Exactly.

90:33And if it's if it's all like a set of measures that look the same, you don't have to shape the same.

90:39Like I probably I could find that some other way.

90:42Do you know what I mean?

90:42But it gets really powerful.

90:44Like uh I I hope I I gotta play with this a little more that like it's the explained data eve and feature over is gonna show up.

90:52Should find really cool things now.

90:54Do you know what I mean?

90:55Because

90:56Right.

90:57It just Yeah, true, true.

90:58It's it's almost ahead of its time.

91:00Yeah, it's a good point, actually.

91:02Because the but the big challenge to this point is it continues to be is

91:06It's hard to do multi-fact, not on real-world data, because you can't find fake data that's conformed to life dimension.

91:15Like it just doesn't exist.

91:16Like that was a lot of

91:18Using uh Adam Miko's data mock started before, like taking data going, could you turn it into

91:26Support data backing.

91:27Like I'm trying to fake that data out for TC.

91:29That's not that easy.

91:30You know what I mean?

91:31But I'm starting to work with clients with real-world scenarios now.

91:34Yeah.

91:35Right by that.

91:36It's a good point.

91:36Yeah.

91:37I'd it's it is uh it's an evolving

91:43It goes it goes back to what we discussed about earlier, which is how does how does how does someone know whether or not to do something?

91:50And I think

91:51I wish Tablet took a bolder steps to say, hey, look, I know you're using a join and a left join.

91:57Yeah.

91:58We've converted it to a relationship for you.

92:00Yeah.

92:01Right?

92:02Just just do that like

92:04Take that action not as an instruction but as um direction if that makes sense

92:12And so you build a model how you think you want it built, but we'll rationalize it when we you know there's that performance enhancer thing, right?

92:21Like one one capability of that could be, hmm, looks like you've done a left join.

92:26But you've really not done anything that our relationship can't do.

92:30Click this button and we'll turn it into a relationship and do all the switcheries for you.

92:34I know that's a big task.

92:35But that that's the kind of that's the kind of like that's the kind of, you know, like

92:41thing that I think people would go, oh, okay.

92:44Cool.

92:45All right.

92:45Yeah, sure.

92:46Right?

92:46And now they're using a relationship, they don't realize the benefit.

92:49But then later on, they'll try something.

92:52And that's when they'll realize, okay, this is what's going on.

92:55It's small things like that.

92:56I think it's a really good point because it's not very discoverable, right?

93:01Correct, correct.

93:03You have to like um you have to put it in the way of people to get it to work and even even in the visual, like it says

93:10dragging in doing the noodle and people are like but that's different and they double click in and they do the join and they're like ah familiarity you know like

93:20No for sure.

93:22And they uh yeah.

93:24And and I mean the last mind bendy thing on it I would say is

93:28Having been in VI now for 25 years, right, you have all these like small normalized tables underneath applications, right?

93:36So that have many applications, because that's an efficient way to write data into

93:40So you would never see a big fat like one great big table under an application that would have because it only updates the customer table and it needs to update the customer table.

93:48In in twenty, twenty-five years ago we said

93:51Oh, that's not very efficient.

93:53Um you have to you have to denormalize that table.

93:57So you have to take all those tables and munge them together.

93:58I'm sure it's going to explode, but you know, that's the only way that uh that you can expect expect to get any performance out of it.

94:05them right and then we also then said and you should pull cubes out of them but then and then cu and then so but then we gave up on cubes because only because computers got faster.

94:15And we went, um, and we went, no, actually, you know what?

94:19With Tableau, you can just build hierarchies.

94:20You don't need to force them into a cube, but there's only one way down.

94:23And Tableau can do that because computers run faster in that amount of time more than anything.

94:27But now what we're saying is effectively if you're using Salesforce as or any other app, so it's just we're talking about Salesforce, right?

94:37So

94:38Take your tables, right?

94:40If they're you know if you have good processors, your data should be clean anyway.

94:43Do you know what I mean?

94:45Snapshot those tables exactly as they are.

94:48We'll model them like that and dynamically query only what we need from them at runtime.

94:53That's what relationships are, right?

94:55So it's you know so at the extreme

94:59And again, it would never happen like this because combining data, but if you had one app, say you were a small company, you ran everything on Salesforce, but you had Sales Cloud Service Cloud, right?

95:07The dimensions are conformed anyway.

95:09And if you could process these, you could say you don't need data engineering, just when you connect Tableau to that, we automatically take a snapshot of that into hyper

95:19So we're not querying that Oracle database underneath.

95:22It would grind you to a hole, right?

95:24And then use drag and drop and we'll dynamically create the data model for you on the fly only for the questions you're asked.

95:30We're not going to pre-compute a whole bunch of stuff you might not ever need, right?

95:34And we're never going to eliminate answers.

95:36That's how powerful it is.

95:38And I don't think I'm not sure how many people understand that's what it is.

95:41Yeah.

95:42It's a hard one.

95:42Yeah.

95:43It's a hard one.

95:44I I don't even know how to I'm struggling to explain to people not what it does.

95:49I'm having a hard time explain trying to explain to people why it mattered and that they'd never had it before.

95:55So when I talk to say Tableau AEs, they're like, we could always do that.

96:00No, you couldn't just kid them.

96:04Or if you did it was messy and you didn't realize um you know what you're doing.

96:13Yeah, yeah.

96:14And oh god, yeah

96:16It's something I have.

96:18Oh yeah, someday I hope I have.

96:20Well I had one yesterday that was really simple.

96:22I had two tables.

96:24Um

96:24for it was calls and sales or something.

96:27And then he had a customer table that it conformed to.

96:30And it's natural to try to do this.

96:32He goes, but my numbers don't do, don't work up by date.

96:35I go, well, which date are you picking?

96:37Right?

96:37And he goes from one of the two.

96:39I go, well, the other one doesn't know that.

96:41Like the other table doesn't know that date.

96:43I go, all you have to do is just scaffold a day table to both of those baits.

96:47So he had customer

96:49Because you would have had to before.

96:50Customer, then the other two tables hang off it.

96:53I go flip that.

96:54So if the other two tables hang customer, now that the base tables also hang a date off those.

96:59Yeah.

96:59Like we did this in real time in 10 minutes.

97:01I go now bring that new date.

97:03I was lucky I already had a scaffold table, right?

97:05Good.

97:06I go now.

97:07Yeah.

97:08Um

97:08Now drag that new date on, boom, the cows work.

97:12Just like that.

97:12I'm like, you've been struggling with it for like hours and hours and hours.

97:15And he goes, I don't think I've got this to work.

97:17I'm like 10 minutes.

97:19And you touched on something there.

97:20I was like, even that, like what if Tableau just had an inbuilt date table?

97:27Right?

97:28So you didn't have to get the scaffold set up.

97:31Like specifically for this use case.

97:33Power BI has it, right?

97:34I Exactly.

97:36I talked to Thomas about this.

97:38I go, the two things that the two features I want more than

97:41And and I go we wouldn't have needed composable data sources nearly as big with these two.

97:45One is I want to be able to right-click in the cam or whatever the thing is and go give me a date table.

97:51And it just shows up, right?

97:52And it's just sitting in hyper anyway.

97:54Do you know what I mean?

97:55The other thing, I can't believe we haven't had this whole time is why do I have to imprep

98:00So if I'm I barely use CREP anymore.

98:02Like sometimes you have to for sure.

98:04But I often dumped a snowflake or something, right?

98:06Or a flat file.

98:08Because what I want to do is I don't want to publish a data source.

98:12I want to publish a table, because it's a table anyway.

98:15And why can't I get at hyper tables?

98:18I understand this, I think the performance impact of these composable data sources might be harder than people think.

98:24We'll see.

98:24But I guess you could have avoided the biggest use case for it if you let me publish hyper tables and then let me join hyper tables together rather than a data source.

98:34Because it's a database anyway.

98:36Like why not let me publish it to a namespace?

98:38in hyper on my Catalyst site and then let me see the tables on just like a database.

98:45Just create that restriction.

98:48Well because it's there anyway.

98:49They just don't expose it.

98:50Dude cause true.

98:53That's a good point.

98:54Yeah.

98:54Like that's all like it would have been easy to understand.

98:57So simple.

98:58Do you know what I mean?

98:59Because it's like why do I have to publish this thing as a data source?

99:02I want to publish it as a table and I

99:04I don't have right access to my company's Snowflake or SQL Server or whatever.

99:08Yeah, you don't want to go to the next one.

99:09That one seems like an easy one to me.

99:13Small things like that, I think, would just really help with the pickup of this kind of stuff.

99:17You know, the silly um yeah, like being able just to s specify like a sequential numeric sequence, like a date sequence.

99:25Specify the start and then how much further into the future you want to go and then just let Tableau run that um for you and then you don't have to bring it in.

99:34And then

99:36They could go to infinity and if they don't need to, but if they went to infinity on the date in the future, the relationship's never gonna call that data anyway, so it doesn't matter.

99:45Do you know what I mean?

99:46Like the whole human view of the relationship is it's that's never gonna come back in the query from that table.

99:51And you could do wonderful things like mess around with time zones and like all all of that beautiful stuff that could be done beautifully and elegantly, the way Tableau does it could just

100:00She just kissed for you.

100:04Yeah, yeah, yeah.

100:04You just get that for free.

100:05Um so be Yeah, that that to me is like it's those little things that I wonder like

100:12I know I know the teams like, you know, we sit on the outside looking in, oh, if you could just do this.

100:17But like, you know, Thomas and team, they're working at like like big problems, like really big problems.

100:24And then once they enable those things, we come flooding in with all these little requests that we go, just do this, just do this, do this.

100:31Well I think how do you it's great to be able to do that.

100:35A product manager for a lot of years is

100:38I think what people like that have to deal with that we don't see is Oh the in terms of well also but the but the amount of technical debt in the product

100:49That we don't see that when they try to build something it doesn't work like the way it's gonna work.

100:54Like the yeah.

100:55Luckily, the only reason I would do custom SQL right now

100:59Is in this multi-fact world, the only way you can filter a table on the number of years is if you do it by custom SQL, because otherwise the way the query stack is

101:10if you put any kind of filter and it forces a join an inner join across everything.

101:15Oh which is crazy.

101:17But um but they are going to fix that with something.

101:20There's going to be another thing in the order of operations above context filter called table filter.

101:25Oh sorry.

101:26They might call it something else, but it's effectively going to be a table filter, which is I only want to filter a table first before you do anything to it.

101:34Which makes sense in the order of operations.

101:36But again, it seems like an easy thing.

101:38It's relatively easy.

101:39Yeah.

101:40Maybe not.

101:40UX-wise, but definitely it's hard because they got this query stack underneath that they didn't write.

101:47Yeah, yeah, yeah.

101:48You know what I mean?

101:49Like that's the part.

101:50It's like a rule set.

101:53You know what I mean?

101:54Like, yeah.

101:55Yeah.

101:57Right.

101:57Well because the other thing is I I don't know why Tableau from day w I understand why they don't let us see BizQL, I guess.

102:04Although I really think they should have talked about it more.

102:07I talked to

102:08Andrew Beers, way back when we got I got hired in my first boot camp, he gives a presentation.

102:12I'm like, why don't why is that presentation not out for everyone?

102:15Like it's yeah.

102:16He's like, well, that's our secrets.

102:18I'm like

102:19They'd still have to copy it.

102:20You know what I mean?

102:21It's added to the and but at a minimum, like I don't know why they won't let you see the sequel they generate.

102:27Because they do some brilliant stuff.

102:29Like uh

102:30Andy Nelson had a great post because she traced it, right?

102:32Like lots of times, I think people don't trust relationships as well because it gives answers that joints effectively couldn't give without building temp tables first.

102:42And what Tableau's doing, which people don't know, is lots of times it's reusing the thing that's been in forever with um uh blends, which are a terrible experience, but

102:52What it does is it queries both, it brings them back, and it stitches the query together on the product.

102:59In the product.

103:00So it's like

103:01That's why I can do all that multi-levels of aggregation thing that would be nasty in SQL.

103:05Like a stages the subquery or view first and then join it to the other side.

103:10But they don't have to do that is why it's fast.

103:13And people are like

103:14I hear relationships are slower than joins, but I haven't had anyone prove it to me yet.

103:17I'm like, I don't know the use case of relationships.

103:20I've only ever had it faster.

103:21Yeah.

103:22I don't know.

103:26Exactly.

103:27Exactly.

103:27Yeah.

103:28Yeah.

103:28Like every yeah, I was explaining this to s to a colleague actually today and they were asking some questions and

103:34They were asking the entitlement tables and I was like, Yeah, relationships every day of the week.

103:38Just just just choose a relationship and they were like, Oh, what oh, and they kind of thought it through and they were like, Oh, okay, if I use a join

103:45it would explode the data by the number of people in my entitlements table.

103:50I'm like, exactly.

103:52And they did it and we did it and we kind of came to like a a a demo solution.

103:57We're like, oh, we have a thousand records in that table.

104:00And we have 50,000 records in that table.

104:02And at query, Tableau just takes the one person from the entitlement table and the rows they're meant to see and the result is less than the good like

104:13f less than what would have been no I don't know what it what it is because I don't know how many rows people see, but it you can go into millions if you're not careful.

104:1950,000 by a thousand people, right?

104:22Like tens of millions.

104:23And it comes out as just like a few thousand.

104:26Like and it's it's fast.

104:27It's crazy.

104:28Well, I think I have this these diagram.

104:30I think I put the diagram in the on the multi-fact blog I did on the Florida Twins again where I'm just like one's multiple

104:38Yeah, one multiplies together all your rows and columns, right, with a join and eliminates answers.

104:44The other one adds them up and doesn't eliminate answers.

104:47Which is crazy.

104:47Like the difference you're not trading anything off.

104:50You're getting the best of both.

104:52Like that the one where we started this whole thing with the Airbnb example, if I can show comments or whatever, I've tried in the demo to show how big.

105:01It would be if I joined it, forget about how long it would take.

105:04And I haven't been able.

105:06I've got like a MacBook with 60.

105:08I haven't been able to get the

105:09I haven't been able to get it.

105:13And it just installed me.

105:17I can use one of my tablet files like hire some AWS uh credits.

105:21I should try to get see if Tableau Cloud will do it for me.

105:24But um it would be so big.

105:26Do you know what I mean?

105:27And then and slow and I'm like the other one's like

105:30Because what it's doing is it's packing up what it wants and it's issuing the query against just to bring back what it needs to bring back.

105:35And your exact example of this

105:37Same client I have it set up, yeah, they part of the it's a call center up and they they outsource some of it and in-source some of it.

105:43So they need a vet a lock under outside vendors and what they can see.

105:47Yeah.

105:48And

105:49They could until relationships came they couldn't do it.

105:52They couldn't uh properly end they couldn't do it because it the Viz was so slow

105:57And then they throw this vendor, everyone's like, bro level security so slow.

106:00Like I've watched them have a new team member, we do office hours, and they'll build out a new one.

106:05And he's like, you need a vendor lock on that.

106:07I'm like, oh that's gonna slow it down.

106:08Like add one calculated field and the thing performs exactly the same.

106:12Or at least you can't perceive the difference.

106:15It's the same.

106:16Do you know what I mean?

106:17Like it's like it's But there's there's old blog posts out there that tell you that like doing it.

106:21like without the knowledge of the new capability.

106:24There are these sort of folklore things.

106:26Like don't do this in Tableau.

106:28Tableau likes the data in this track.

106:29You know, these things that were like really from way back when the products changed a lot since.

106:34Computers are faster as well.

106:36So actually some of those things, even if they're still true and they are slower, you wouldn't notice the difference 'cause your laptop is through it and

106:44The clouds mince minster it.

106:46Actually the large scale data you talk about is where you see these problems.

106:50And even at that scale

106:51It's really hard nowadays in the product to make that mistake because of some of the good practice they've put into um into some of the way you build things.

106:59But

107:00Still a long way to go with data literacy.

107:02Yeah, I know we can go on over this forever, but you just remind me of another good reason Tableau might want to do that joins versus relationship thing is

107:09Given that they don't charge for consumption on cloud, you think it would be in their best interest to drive compute down?

107:15True, true.

107:18Yeah, we should get on

107:19Trying a team of that alone to do that is like it would drop the AWS charges like crazy.

107:26That is a great thing.

107:27Is yeah, hey, look.

107:28Why didn't you sponsor community content on this?

107:31Because this is what it can do to your tablet cloud.

107:37Did you did you pick up on the hyperforce migration?

107:41Uh yeah.

107:43Yeah.

107:44I got an email like was it yesterday to my work instance of Tableau Cloud.

107:49And it talks about um moving Tableau Cloud and specifically US East, which is they use the same pods as AWS.

107:58So US East, they're moving it to Hyperforce.

108:01And Hyperforce is Salesforce.

108:03cloud sort of AWS instance.

108:06And going back to our discussion earlier on about, you know, what's going on with Tableau and everything.

108:12This move from Tableau for Tableau Cloud to move to this more central AWS ecosystems makes me think back to like what we were talking on earlier on about, which is

108:21What's going on behind the scenes?

108:23And this is a huge monumental thing.

108:25Like for context, you have to tell all your admins to change the IP addresses they've whitelisted.

108:30And for record, Tableau ones never change, the Salesforce ones change every month.

108:36So like they're not gonna love you for that.

108:41And then and then secondly

108:43Like the the the the potential for like just monumental screw-ups is going to be incredible.

108:51So

108:52I think this is another thing that people are going to sleep on.

108:54And then it's going to happen in November and December.

108:56The deadline is January.

108:58And that's when I think it will hit.

109:00But the Salesforce and Tableau have been saying it.

109:02Like the emails are coming out.

109:03And I think it's people just missing it.

109:05Well do you know what we uh we should definitely stay in touch on this because most of our clients are Canadian and we're Canada and Canada's the first pod going.

109:14So there we go.

109:15We're gonna get it first.

109:17Yeah.

109:18One of the last couple of pods to go up, first one to go over.

109:21And an upture.

109:22I have to spend more time understanding why they're doing it because I think the reason because I was at the Dreamforce where it was announced again five years ago.

109:29Like it's gonna be out in no time, Hyper Forge.

109:32The the idea was that then they could put it on Azure GCP as well as AWS.

109:39So your data egress cost would be nothing

109:42Right.

109:42So if you're paying BigQuery if you're paying Google and Negrest thing and you could drop it in it in the same pot.

109:49Now it's AWS only.

109:50I'm guessing they found out that they have way too much

109:53any dependencies on AWS specifically and back end monitoring and stuff.

109:57Which again and they've brought the data owners to the platform, Zero Copy, Snowflake, Google, um, Databricks, right?

110:06brought those people to the data source rather than going out to like platforms that might have that ecosystem.

110:12Yeah you're right all these bridge clients are gonna start I saw a bunch of bridge clients failed.

110:16Bridge clients bridge clients

110:20Oh I was like, well I because I'm on like I'm an administrator on a bunch of client sites and my inbox is filled up and I'm like the first thing didn't happen yesterday and then I went no no no Canada soon

110:32Like is it?

110:33Bridge clients.

110:34Geez, that is a painful I mean bridge client bridge is always fun.

110:38It'll be a good thing.

110:40Yeah, yeah, yeah.

110:42God, if you've got like a distributed um uh it's the consultancies that fill this pain.

110:47If you're managing or helping a customer and they will absolutely not have a clue, they're missing these emails, you're not getting the emails about it.

110:54And then you're gonna be like, right, I've got twenty clients to go through and change all their bridge instances.

111:00Nasty day.

111:00I'll keep you posted because I'm gonna get hit first.

111:02Yeah, let me know.

111:03Let me know how it goes.

111:04Canada custom.

111:06Oh god.

111:07I've I've had you on for way too long.

111:09Uh this was so much fun.

111:13We should we should have a chat again once all the dust is settled with um

111:18with Einstein.

111:19And I do think in a year's time we should probably revisit data modeling.

111:22Because I felt like data modeling has just been this moving target.

111:26Even with the data model, I feel like I just started to understand it.

111:29Then multi-felt fact analysis sort of has come online and I'm like, okay, even more's gonna happen.

111:35And I think with the way the ecosystem's going

111:37we are going to get to like, you know, a place where I think data modeling in Tableau, specifically the authoring experience

111:45is in a really, really good place.

111:47Yeah.

111:47And many of the things that we've had to go out to do in lots of other tools will become possible, but there'll be an approach to doing it in Tableau that will be specific to the ecosystem.

111:59And potentially with where Salesforce goes with data cloud and all that stuff, specific to how the whole sort of um setup works.

112:07So

112:08I think that would be an interesting transition to see um how it how it plays out and that your opinion on that would be hugely.

112:15So we'd love to have you touch on that as well.

112:17I'll leave you one to think about too is um when you start playing a little more with multifacts.

112:23Although the whole concept's hard to explain, the interesting thing is I have a feeling that it makes for people who haven't yet understood relationships, ironically, easier to understand relationships.

112:36And the reason is because it does things that they couldn't write SQL to do.

112:41So it forces them to think about what's going on.

112:44And then I think that my hypothesis is called.

112:47Oh, I gotta do without even these multiple

112:49base tables.

112:50But I never took the time before because I wanted to control it.

112:53But I couldn't SQL my way into this model.

112:55Right?

112:56So anyway that's interesting.

112:59Yeah.

112:59Yeah, that's a good point.

113:00That's a good point.

113:01I'll look out for that.

113:03Um we we have a we have a closing tradition.

113:06It's a very new one that's not that old, where the previous guest asks the next guest the question.

113:11It's copied off um a podcast by Stephen Bartlett, diary of the CEO.

113:15But I think it's really good in this ecosystem because we have lots of people working in data doing different things.

113:20The previous guest was Katrina Menna.

113:22You watched the video and uh at the end of that she asked um what is sort of your most impactful or memorable project that you've worked at that you you can share as it were.

113:31I appreciate that in consulting.

113:33You you can't often talk about your clients.

113:34So you can sort of allude to it or talk about the component that you found interesting in that sense.

113:40Uh

113:42Can I go way back on this one?

113:44Absolutely.

113:46No, no, it could be anywhere.

113:47Any any any context applies.

113:49You're the guest.

113:50You can choose.

113:51I'll I I'll make I'll sync this one I'll make this one as short as I can.

113:55So I I gr I actually I started programming I'll show my age on a VIC twenty when I was ten years old.

114:01But um I had this weird thing with my parents where they made enough money where they could

114:06Um they didn't have a lot of money, but they were gonna pay for my university, but they were gonna make me go to the one in town unless I found something.

114:13It's a terrible reason to pick a career, unless I found one that went away.

114:16So instead of going into computer science, I went into pharmacy.

114:19things.

114:19So I'm actually a pharmacist.

114:21And then first pharmacy I went into, this is now this is in the days pre-internet.

114:26Not a lot pre-internet, but a little bit pre-internet.

114:28And uh and the way they used to do inventory management in the pharmacy

114:33was um you know they'd you know if you if something ran out like pills ran out in the bottle you'd write it on the list right and someone would key those numbers in and hold it up to a phone and it would

114:44kind of go out or and then we had a pharmacy tech once a we could go around shape bottles to see if anything was missed.

114:49I'm like, this is insanity.

114:51Right?

114:52And then so um I brokered a deal with our owner and a wholesaler to go single supplier

114:58And we had automatic order.

114:59I worked with our software vendor to bring it in.

115:01And then this thing dialed out on a modem every night.

115:03But like we got we got our turns out by like

115:064x or something and our inventory down by 33%.

115:09And we used to store it about two people a day, and we got down to a half.

115:13And it was just the reason it was so fulfilling to me is it kind of set

115:17my career going, I'm gonna focus on using software and business to make processes better because it actually makes people's I mean it doesn't make lives better like a charity does, but it you know, people are doing these mundane useless

115:30It does, yeah.

115:31Do you know what I mean?

115:32And so it's been kind of ingrained in me since then that I'm just like always on the lookout for like what software and problems like kind of intersect that I can help people.

115:42you know, focus on the actual job instead of arguing about you know stuff that there's an answer that they could just get at it or whatever.

115:49So it happened really early in my career.

115:52Um and that's why like when I see this integrated Salesforce thing, I'm like, oh, that'd be awesome.

115:56Do you know what I mean?

115:57Like in the experience doing it.

115:59Yeah, exactly.

116:01Yeah.

116:02Seeing the opportunity in that, absolutely, I couldn't agree more.

116:04And it's a

116:05That is it you're fortunate to have had that experience.

116:08Because I think a lot of people in their careers get that a little bit later on because you know the world of work today is is sometimes quite prescriptive, right?

116:15Like it's

116:16you know, y y it's it's harder to be that kind of person who can go and suggest that kind of thing.

116:21So unless you work in a really small organization, you're not gonna get kind of get that exposure.

116:26So

116:26If you can get that kind of opportunity, I always recommend people do go work in the smallest company with the most exposure to the stuff you want to get closer to.

116:34'Cause that's the best way to learn.

116:36Oh I agree a hundred percent.

116:37Yeah, yeah.

116:38And then um the final bit is you you get to ask your random questions to the next

116:44The next guest.

116:44And I I answer it in this in this video, but they answer it obviously in the next one.

116:48So it can be anything.

116:50It doesn't have to be Tableau, just data related.

116:52Um it's really open to you.

116:54Right.

116:55So if um

116:56If a venture capitalist came up to you and handed you $100 million and you knew 20 developers that were ready to go.

117:07What analytic what would the analytics product you build look like?

117:10Oh Wow, that is a big question.

117:16What would the what would it look like

117:19See that is that the same question as what it would do?

117:22Or is that like Yeah, I guess I mean what would you yeah, what what would I build with it is really what you're asking.

117:28Right, because it's a little it's a little easy and too easy like we talked about because we don't know all the technical debt and bugs and priorities to say how we'd make Tableau better.

117:37But if you would uh if you had to go whiteboard from scratch, do you only mean would it be uh you know Figma with query in it?

117:46Like well I don't know, do you only mean?

117:47Like it's wide open for me.

117:49I think, and I'm gonna take a pr a bit of inspiration from um something I've seen, uh Cole, uh I never know how to say her surname.

117:58Uh oh yeah, I know what you mean, yeah.

118:00Um she's written a she's written a a book, a data book for children.

118:05And one of the biggest problems I think we have is

118:11People learn about data and the tools for data too late.

118:17So I think I'd try and build a product for

118:22education that helps teachers teach data and in it in itself it would be a data product if that makes sense right like you sort of learn by doing

118:32But it's contextual to the job of being a you know, in a class, it's contextual to the way a student goes about their education.

118:41And it it makes it makes the things they do in school and the data around them more contextual.

118:48And I it's you know, it's I don't know, I'm I'm not sure it would be a great business because, you know, schools are frugal places and

118:53Uh you know, a hundred million dollars it's I'd probably burn the money and not make a profit, right?

118:58No, uh there's a lot of people in the school space actually.

119:03Right, right.

119:04But I I think you just have to sell all the schools to make money.

119:11Correct.

119:11But you know, I kind of think like, you know, I used to remember my school used to use something called FileMaker Pro to as a date product data product to manage like records and stuff like that.

119:21And it was

119:21You know, fine, it was a database tool ish and it worked on Macs and yeah, the school had Macs and stuff like that.

119:26But it it was very initial.

119:28I'm not even sure if anyone still uses that.

119:29I'm sure someone does and

119:31you know, runs like uh the whole entire system off it.

119:33But you know, I think I think well what else do people have?

119:37And it's Excel and Google Sheets.

119:39And I think

119:40That's sort of part of the problem, right?

119:42You you're taught this conceptual way of working with data that I think doesn't open your open up your mind to like the wonderful creative ways that you can work with data that you learn later on in life.

119:53And so if you can instill that tool and that capability in young people, I wouldn't be surprised if they then go on to do incredible things with data in the future, like inherently, right?

120:05And so that to me is where I'd build a product.

120:07Really instead of trying to sell to businesses so they can go and do things faster and more and you know, create more, you know, nuggets or whatever.

120:14What if you could build a data product?

120:17for education, in education, that students can use themselves.

120:22You see university students using like uh GG Plot or um uh you know what's the other one that

120:30Uh that begins with S, S PSS, like I because this is what they're given, right?

120:34Like what if you could build something

120:37For that academic trade, like for that.

120:39Like having Python and put a D3 live.

120:42Are you kidding me?

120:42Yeah, yeah, exactly.

120:44Exactly.

120:44And and something that can abstract those

120:47complex things.

120:47It would sit on top of that, absolutely.

120:49It would need to be open source, absolutely, for it to make any sort of mileage.

120:53But that's that's the sort of space I'm thinking because at the moment.

120:56Those people only they live off education licenses or they live off other things that, you know, don't really quite fit their use case.

121:03So yeah.

121:04I love the getting people before they get to Excel and getting

121:09Exactly.

121:09Well the the one I do say to people all the time that's more obvious is would anyone in the world

121:15I thought about helping people with presentations going, you know how you should it get people's attention?

121:20It's a title and then a bunch of bullets attack

121:23That's the way they can power like had they not built it that way, people wouldn't have made all these terrible presentations.

121:28Terrible mistakes.

121:30Exactly.

121:30Tailwag that dog.

121:32And I often wonder how much Excel

121:34Um busy calc for or whoever that even predates me, but whatever logus one, two, three, whatever, like that spreadsheet style, like

121:43Do you know what I mean?

121:44Like we all learn because we I'm sorry, we all think that way 'cause how that's that's the first thing we're doing, yeah.

121:49That's where we learn

121:50But maybe there's another way.

121:52So anyway, that's a great answer.

121:54I love that answer.

122:02We should chat again.

122:03I'll make sure

122:04um I find some time and yeah I look forward to sort of all your comments and uh you know whatever what whatever other pieces of content you make.

122:11Maybe another book, maybe more videos.

122:13Highly, highly look forward to them

122:15Thank you, it was a blast.

122:18Good luck.

122:19All right, take care.

122:20Okay, it's back soon