Tableau's core audience is changing | Tableau 2023 Keynote Full Commentary & Thoughts

0:00Hey, it's Tim here. This video, there's not

0:03going to be any editing, I'm just simply

0:05going

0:05to be commentating or doing a commentary to

0:08the Tableau keynote. Now, I've watched this

0:11keynote now seven times over, it's

0:14ridiculous. So at this point, none of it is

0:17new to me,

0:17none of it is sort of familiar. And I've

0:19had quite a lot of time to formulate some

0:21thoughts

0:22and opinions. And what I'm going to do is I

0:24'm basically going to go through the entire

0:26keynote and share things that I couldn't

0:28really share in the summaries. It's always

0:30important

0:30realizing that the summaries and some round

0:33ups that I do, they don't often contain much

0:35of

0:36my actual opinion in there, because I'm

0:38simply trying to do a service, I'm trying

0:40to summarize

0:40them so that people who don't have the time

0:42to watch the whole thing can get something

0:44out of them. And so I have at the end of

0:46those videos inserted my opinions and

0:48thoughts.

0:49In this one, I'll actually go into a lot

0:51more detail and a lot more context. It's a

0:53long

0:54one, I'm sorry, but it's timestamped. And

0:56hopefully, you can jump to the bit you care

0:59the most about. And you can find out what I

1:01really think about the whole keynote. And

1:03I can also shed some light on to some of

1:06the thoughts and discussions I've had since

1:09actually,

1:09with various developers inside of Tableau,

1:11but also various people in the community as

1:13well. So I'm going to try and put all of

1:15that value into the next, what will

1:17probably end

1:18up being an hour and a half. But

1:19nonetheless, let's get stuck in. Okay, here

1:22we go. So Salesforce

1:23always starts these keynotes with this

1:26really irritating, forward looking

1:28statement. What

1:29I hate about this slide is it's a wall of

1:31text. And what do they always say about

1:34slides?

1:35Don't do walls of text. So I always think

1:38it's a little bit ironic, all the effort

1:41that

1:41goes into the keynote, and all the effort

1:43that goes into planning everything. And the

1:45first thing you see is a wall of text

1:48anyway. I always feel like you could you

1:50could just

1:51put like a QR code to say, forward looking

1:54statements, go here to read our full notice

1:57.

1:58We'll assume you've read it from this point

1:59on something like that. That'll be much

2:01nicer.

2:02Just please get rid of this slide. It's so

2:04annoying. And Tableau customers have seen

2:06this all the time. It's really irritating.

2:08Anyway, let's get stuck in. Should we watch

2:13it 2x speed? Maybe we should. No, we won't

2:16do that.

2:16Please welcome Chief Product Officer Table

2:20au, Francois Agenstadt.

2:22Hello, TC.

2:25Put a move on my mic, right?

2:29All right. It is so great to be here

2:32together. I mean, look at us. This

2:36conference is a celebration

2:38of this community, a celebration of the

2:42power of data, and what we can do together.

2:46And

2:46I mean, look at this, we have seven, eight

2:50thousand people now in Vegas.

2:52Now what is super interesting is I made a

2:54note about this setup. And this is a very

2:56Salesforce centric setup. And I kind of

3:00think this year, we kind of had to do it a

3:02little

3:03bit as well. But I think what we've got

3:05going on here is here on the front, you

3:08also can't

3:08see really where I'm pointing. But here

3:10where my mouse cursor is, I think are

3:13potentially

3:14keynote speakers with some really important

3:15people. They've got the really nice sofas

3:17up front. We have Pedro here, we have Lar

3:20issa, head of community services, you have

3:23Ryan

3:23Ater here at the top. We have obviously

3:26Francois Seat. And then here we have I

3:29think it's Is

3:30it Ken or Kevin? I can never remember which

3:32ones which I'm really sorry. One of the Fle

3:35ur

3:35Large twins. And we've also got Anne

3:37Jackson, these two made it into the Hall of

3:40Fame, Tableau

3:41visionary Hall of Fame, formerly the Table

3:43au Zen Master program. And then over here on

3:46the left, we've got two more people, I can

3:49't quite make out who they are. And but yeah

3:51,

3:51this is this is the new setup. And it's

3:54always interesting who goes into the circle

3:55, I kind

3:56of it always feels like there is some sort

3:59of hierarchy in the community. And based on

4:03the number of reservations that you know,

4:06get set up at any conference, I think my

4:09first

4:09ever conference, I didn't know about this,

4:11I rented the front because I thought I was

4:13being keen. I got there and there was just

4:15five banks of reserve seats. And I was like

4:18,

4:18oh, so like the closest I can get is like

4:21five rows back. And then I realized five

4:23rows

4:23back was like the worst seat, the best

4:25seats were actually in the crowd on the

4:28banks where

4:29you could see the whole keynote and get

4:30more of the atmosphere. Anyway, always

4:32think that's

4:32a little bit strange. But nonetheless, hey,

4:35that's that's just how it is. Together, and

4:3810s of 1000s more online. Welcome to the

4:42data party of the year.

4:50And this is an amazing event. We have three

4:53days of data goodness. Yeah. Over 200

4:59sessions,

5:00hundreds of doctor appointments. We've got

5:03the legendary Iron Viz celebration tomorrow

5:06.

5:06Oh, yeah. And we've got Sophie Tucker for

5:12data net data night out. It's going to be

5:14a data party. This is pretty interesting. I

5:17will say this year, I didn't go in person,

5:19so I can't speak to the in person

5:21experience. But I have to say virtually, it

5:23was actually

5:24quite disappointing. Only 20 of the

5:27sessions were on Salesforce Plus. And it

5:31was, it was

5:32a little bit lackluster, because, in

5:34essence, in between those 20 sessions,

5:36there was really

5:37nothing to do. You know, when we had the

5:40virtual conference, because of COVID, it

5:43was amazing

5:44how much effort Tableau put into a virtual

5:45conference. And I appreciate that it's

5:47probably

5:48hard for Tableau to put effort and budget

5:50into both an in person and a virtual

5:52conference

5:52simultaneously. And they did do that in the

5:56first year that it came back. But it felt

5:59like, hey, Tableau was calling on people to

6:01come and attend it virtually come and see

6:02on Salesforce Plus, but actually there was

6:04very little on Salesforce Plus during the

6:06event. That said, afterwards, I believe all

6:09the sessions are going to be made available

6:11.

6:11So that's a big plus. But I have to say,

6:14like, after the in person conferences kind

6:17of took

6:18a hit because of COVID. To me, it felt like

6:21in person conferences would never be the

6:25same

6:25again. To me, it felt like virtual

6:27conferences were here to stay. And the

6:29virtual audience

6:30was in fact the biggest audience of the

6:32Tableau conference. And if you compare it

6:34to experiences,

6:35for example, I use dbt as an example, dbt,

6:38in my opinion, ran a really successful

6:41mixed

6:41in person and virtual conference. Now, of

6:44course, dbt is much smaller, much smaller

6:47than it is. But their hybrid setup was

6:49really, really good. We had probably 70% of

6:52the sessions

6:53were available to watch live. And then the

6:55videos and the slides, everything was

6:58shared

6:58almost instantaneously on the platform they

7:01were using. We had slack, it's kind of

7:03ironic

7:03that Salesforce owners of slack didn't put

7:07up a slack instance for all of these

7:09sessions

7:10to happen. And it's also alarming that, you

7:12know, surely slack could host a virtual

7:14conference,

7:15like, why couldn't slack just host the

7:17virtual experience, you've got you own the

7:20product,

7:21you make the technology, I believe slack

7:23has this sort of meetings capability. And I

7:26'm

7:26sure you can build some custom integration

7:29for big events for customers to do live

7:31streaming

7:31of events that is surely a capability you'd

7:34love to, to have in the platform. So yeah,

7:36I kind of think there's definitely more

7:38Salesforce could do with the virtual

7:39experience going

7:40forward. To me, I still think it's the

7:42biggest audience of conference. And just

7:44looking at

7:44the number of people who've watched my

7:46videos, and that audience exists because

7:48there was

7:49nothing else to watch virtually. And people

7:51don't want to wait for anything else. So

7:53you

7:54know, me summarizing the videos, me summar

7:56izing all the other content, that stuff is

7:58still

7:59not on YouTube. It's on Salesforce Plus,

8:01but frankly, most people in the tablet

8:03community

8:03have never heard of Salesforce Plus and won

8:05't even sign up because they don't want

8:07anything

8:07to do with Salesforce. So nonetheless, I

8:09still think there's a massive opportunity

8:11there

8:12to really amp up the virtual experience

8:14really amp up the reach. The keynote, in my

8:16opinion,

8:17should be live streamed on every platform,

8:20LinkedIn, Twitter, Salesforce Plus, and

8:22YouTube.

8:23The whole point of the keynote is to really

8:26broadcast the message of tablet. And so why

8:28it was locked up inside of Salesforce not

8:30really clear. But anyway, let's, let's keep

8:33let's keep moving together. And we couldn't

8:35do this without our great sponsors who are

8:37making this possible. So please give a

8:40round of applause for Databricks, Deloitte

8:43Digital,

8:43Interworks, Starburst, and Kyvos. Thank you

8:51. Now you might notice that things look a

8:53little

8:53bit different this year. We're right here

8:56in the middle in this round because we

8:57wanted

8:58to be together and celebrate together. And

9:00we're going to talk this week about the

9:02latest

9:03and greatest innovations in the Tableau

9:05platform, how we're going to move this

9:07platform forward

9:08and help you do more with data. Because at

9:11the core of Tableau is innovation. This is

9:14the 20th year anniversary of Tableau. 20

9:18years. It's amazing. And I will say there's

9:24something

9:24about the audio in this recording that

9:26doesn't quite amplify what's actually going

9:28on in

9:29the audience. So I think there was way more

9:31applause way more round. But I will say

9:33this,

9:34like the 20 years kind of caught everyone

9:36by surprise. It was actually a few months

9:38ago, it caught me by surprise, I didn't

9:39know until two weeks before there was a 20

9:42th years.

9:42And maybe that's a little bit about culture

9:44. Maybe that's a little bit about some of

9:45the

9:46other needs that was going on at the time.

9:48But nonetheless, yeah, for many people in

9:50the in the room, they won't know the Table

9:52au has been around for 20 years. But also,

9:55more

9:55importantly, if that's a surprise, then it

9:58's also potentially an opportunity to

10:00highlight

10:01how far Tableau has come. And so you can

10:04kind of see the whole the whole setup. Now

10:07what

10:07is really interesting in this chart is they

10:10they have sort of marked out the Salesforce

10:13era here, you can kind of see 2019 2020

10:172021, you've got my AI annotation tool

10:20picking up

10:20the text here, those 2019 to where the very

10:22end of 2019 is when Salesforce announced

10:25that

10:25we're going to buy Tableau. And I think the

10:27deal finished in 2020, and then 2021, and

10:29so on and so forth. So you can kind of see

10:32the Salesforce logo there in 2018 2019. And

10:35that is a super interesting thing for them

10:38to highlight. And it's, it's, it's maybe

10:40potentially

10:41a subtle communication. But look at this

10:45meteoric rise, since our Salesforce

10:49acquisition. And,

10:51you know, Tableau was a much smaller

10:52company before, I always think of version

10:5410. As the

10:55version where Tableau really came alive, I

10:58started learning with version eight. And

11:01the

11:01innovation. There's there's a there's a

11:04really famous data visualization by Jock

11:06McKinley,

11:07where he talks about the different points

11:09of innovation in Tableau. And this is kind

11:11of a recreation of that. But I think that

11:13is a real chart to go to, I'll try and link

11:15to it in the in the description below. And

11:18that to me is is the real sort of viz to go

11:20and look at because there you can really

11:22see that the pace of innovation has always

11:24been

11:24consistent. There's basically been no gaps

11:26has always been some sort of innovation

11:28pretty

11:28much every year. And it's been really

11:31exciting to see. What it does call out is

11:33the innovations

11:34that Tableau call out since the Salesforce

11:37acquisition. So let's have a look ask data,

11:40prep conductor Tableau catalog, then there

11:43's like a whole year of no innovation. But

11:46of

11:46course, these are just versions and

11:47releases. So they get sort of marked out.

11:49Then you've

11:50got data stories, accelerators, data

11:52mapping. And so it's actually, you know, I

11:55think many

11:56people in the Tableau community will easily

11:58look at those and say, yeah, those are

11:59those

11:59are very Salesforce centric like things. If

12:02you look at the Salesforce platform, you

12:04see

12:04similarities and synergies with the way

12:06Salesforce thinks about products with the

12:08way that these

12:09innovations are starting to happen. Whereas

12:12if you look before the Salesforce

12:13innovation,

12:14what's considered innovation to me is real

12:16innovation. So hyper Tableau prep, LEDs,

12:19Tableau

12:20calculations, blending data engine, mapping

12:23, all of these things are sort of the core

12:25of

12:25what we know is part of the experience

12:28today. And in many ways, the community is

12:31has been

12:31submitting ideas way back since 2012. On

12:35some of these even basic innovation, some

12:37of which

12:38were announced at this keynote, some of

12:40which are yet to come. But those are the

12:41innovations,

12:42you know, they're not really innovations

12:44because they've been ideas for a while. But

12:46frankly,

12:46those are the ideas that people want to see

12:49. So let's, let's, let's, let's keep going.

12:53Innovation has been at the core of Tableau

12:55since the beginning. And we're continuing

12:57to advance our mission to help people see

13:00and understand data. And last year, we

13:02delivered

13:03over 120 new capabilities to the Tableau

13:06platform, new capabilities for analysts,

13:09consumers,

13:10IT and developers, new capabilities like

13:13data stories and HIPAA certification, over

13:1650 new

13:16accelerators. Just listen to that back.

13:21Those roles, analysts, consumers,

13:24developers, and

13:25IT. Just just keep those four in your head.

13:28We'll come back to that point later. Okay.

13:31And community favorites, like image roles

13:34and dynamic zone visibility. And there's so

13:37much more coming today. Now, in August 2019

13:40, we joined forces with Salesforce. And this

13:44is the customer 360. Yeah, that's right.

13:47You can applause for Salesforce. Now the

13:50customer

13:50360 is our portfolio.

13:55That is a great little moment. And someone

13:58said, woo in the crowd. Okay. That's

13:59probably

14:00the one Salesforce admin who absolutely

14:02loves the fact that Tableau is now part of

14:04the Salesforce

14:05family. And he can use it to help solve

14:07problems, he would probably he or she was

14:09probably using

14:10it already. Nonetheless, nonetheless, that

14:13was a very interesting sort of transition.

14:16And you know, this is a tough crowd as well

14:18. Like just just going back there made me

14:19think

14:20that a lot of people came into this

14:21conference with a little bit of tension has

14:24been a lot

14:24of sort of, let's say upheaval in the last

14:26six months, whether it's to do with the

14:28Salesforce

14:28layoffs, whether it's to do with, you know,

14:31a lack of features, whether it's to do with

14:33this sort of what appears to be a rising

14:35tide of, you know, Power BI versus Tableau

14:37again,

14:38and Power BI kind of winning the battles at

14:41C suite. And there is a lot of that tension

14:43sort of all bundled up into this room. And

14:46you can kind of feel it. And you can tell

14:48that you know, as Francois presenting, he

14:50can feel it, you can tell just as he's

14:52looking

14:52out for people, and they're just looking

14:56back at him cold. And maybe maybe it's the

14:58camera

14:58angles. I'm not in the room, so I can't

15:00tell. But yeah, it's it's a tough crowd,

15:02definitely.

15:03Products and services that helps customers

15:06or companies connect with their customers.

15:09And at the heart of the customer 360, well,

15:12it's data in AI. And what do you need when

15:14you have data? You need Tableau. And so

15:18Tableau is unlocking the value of the

15:20customer 360

15:22and helping every company get better

15:24insights into their customer data. Now, one

15:27of the

15:27reasons that Salesforce and Tableau is such

15:30a great match together is that we share

15:33values.

15:33We believe that business is the greatest

15:35platform for change. We follow what's

15:37called the one

15:38one one model where we provide 1% of our

15:41equity, 1% of our time and 1% of our

15:44product to help

15:46do good in the world. And you're gonna see

15:48that in our keynotes around data for good

15:50or vis for social change. But data is a

15:52force for good, and we're doing that

15:55together. And

15:56as a result, we can do well while doing

15:59good. You know, we're a leader in

16:01philanthropy.

16:02We're a leader in innovation. We're a

16:04leader in culture. And as a result, it's

16:07fueling

16:07our momentum because we're putting our

16:09community and our customers at the center

16:12of everything

16:13we do. And that's fueling our momentum.

16:16That's fueling our success. Let's just go

16:19back here.

16:20This is gonna be really tricky. Let's just

16:22can I go back into this clarification? High

16:23end of a fly 24. I think these are sales

16:27force figures. I think it's a selfless

16:29figures.

16:30I'm pretty sure that's the case. It's

16:32fueling our momentum because we're putting

16:35our community

16:36and our customers at the center of

16:38everything we do. And that's fueling our

16:41momentum. That's

16:42fueling our success. At the end of Q four,

16:45we actually achieved $2.5 billion in

16:48revenue

16:49run rate for Tableau. Is that amazing? So

16:54that's okay. So this is another really key

16:56point. In the last three years, the keynote

17:00and the developer slash authoring message

17:04or whatever, whatever analysts like to see

17:07have all been crammed into this hour and a

17:09half. And what that means is that

17:11previously you could come to the keynote

17:13and kind of

17:13switch off in this session because this was

17:17not the place as a user analyst of Tableau

17:21to pay attention. The time to pay attention

17:24was devs on stage. Those two messages were

17:26nicely decoupled because we're now one

17:30keynote. What we're getting more of is what

17:33seems to

17:34be like an investor message. This is a

17:36message to the market because of course

17:38this is live

17:39there. Press releases coming out of this

17:42and Tableau have to say these things for

17:44the market

17:45to understand how it's performing, how to

17:49assess Tableau as a leader or player in the

17:52market. It also allows Tableau to give

17:56analysts and journalists ammunition to go

18:00and ask the

18:01same question of other analytical platforms

18:03. If you're doing so well, what is your

18:05revenue

18:06run rate? What is your custom account? What

18:08is your year on year growth? How many

18:11activities

18:12and data management? All of these things,

18:14they're kind of giving you these five

18:15talking

18:16points. And the really interesting thing

18:18about these five talking points is they all

18:21show

18:21growth. They all show growth in a really

18:23interesting way. Yet there's a general

18:25sentiment in the

18:26communities, general sentiment in the

18:29industry, let's say, the Power BI is sort

18:31of gaining

18:32ground. And so it's difficult to know where

18:35the truth really lies. Of course, the only

18:39people who really know the answer is in

18:41each and every organization, as people make

18:43those

18:43decisions, they know the answer to that

18:45question, they know which product does best

18:47, they know

18:47how much they spend with Tableau, and they

18:49know whether they're getting value or not.

18:51But here, we have to take it at face value,

18:53and we have to take it and assume that

18:55Tableau

18:55is really doing as well as it says it is.

18:58And these, you can't lie about these facts.

19:01These are things that you are held to

19:03account for in your accounting when you

19:05report financially.

19:07When Salesforce does its annual accounts,

19:08these will be the same figures that you'll

19:10hopefully see as well. So really

19:12interesting to see this here, but I think

19:14it's worth highlighting

19:15that this is a message for investors, the

19:18markets, and the analysts who write about

19:20this industry. And so it has to come here,

19:23because this is what they're listening for.

19:27And we're continuing to add more customers

19:32to the Tableau family, over 100,000

19:37customers

19:39around the world, from small businesses to

19:42nonprofits, public sector organizations to

19:46the largest enterprises in the world. And

19:48our products continue to be adopted at

19:51scale,

19:51and we're continuing to see customers

19:53moving to the cloud at a faster clip. The

19:56cloud is

19:56not just another deployment platform for

19:59Tableau. It's actually become the preferred

20:02option

20:02for Tableau. We're seeing new customers

20:05start in the cloud. We're seeing existing

20:07server

20:08customers transition to the cloud. And you

20:10know why? Because they're seeing better

20:13performance.

20:14They're seeing better scalability. They're

20:17able to spend more time doing analytics

20:20instead

20:20of fussing with infrastructure. And so the

20:23activities on Tableau cloud is just skyrock

20:26eting,

20:26and it's the best place to experience the

20:29innovations that you're going to see today.

20:33So I always think when Tableau says

20:38something like this, what they're really

20:45saying is we're

20:46heading this way. This is becoming the core

20:50product. This will be where our focus is.

20:54And I always hope and pray that there are

20:56people in the room reading the tea leaves

20:59because I've been saying this for a while

21:02now that it's really obvious that Tableau

21:04wants to move to the cloud. This was, I

21:06think, the reason they went and got Adam

21:09Solipski

21:10who helped build up AWS services and just

21:14essentially turned the company into a cloud

21:19first company. And now that he's gone back

21:21to Amazon to take over Andy Jesse's role so

21:24Andy Jesse can step into Jeff Bezos role,

21:27the company is still on this trajectory.

21:30And

21:30of course it would be because Salesforce is

21:32a SaaS company and one of the core tenants

21:35of SaaS is the cloud. So unless Tableau

21:38suddenly decides to rewind time, the cloud

21:42is a direction.

21:43And I think customers are understanding

21:45that. That's not to say that Tableau should

21:48be excused

21:49for not having feature comparability

21:51between server and cloud because there's

21:54still some

21:55areas where that's not the case. But I

21:57think that means that customers can put the

21:59fire

22:00under Tableau a little bit and say, "Hey, I

22:02want to do these things, but I still can't

22:05do them." But nonetheless, those hurdles

22:06are coming down thick and fast. Literally

22:09every

22:09release there's features that are cloud

22:11centric. And yeah, I think the writing's on

22:14the wall

22:14here. This is the direction that the

22:16company is moving in. And if you're a Table

22:20au server

22:21customer and you have a use case that's

22:23exclusively Tableau, of course Tableau will

22:25continue to

22:26serve you. But it's pretty clear here, they

22:28're saying that that will be the secondary

22:30experience.

22:31You know, there'll be features that come

22:34first to the cloud and then later on a

22:36Tableau server.

22:37And that's actually been the case for a lot

22:39of releases now. Tableau server's only

22:41released

22:41every six months. That catches people out

22:44still. Tableau cloud gets releases

22:46quarterly.

22:47That in itself shows you there's a

22:49preference to Tableau cloud getting

22:50features faster if

22:51that's the only preference that was exist,

22:53but it's not. There's other preferences as

22:55well. And so it's definitely happening. And

22:58yeah, you just have to sort of get on board

23:02with the program. And if not, then

23:04obviously there's other competitors in the

23:06market that

23:07use other products. And so, you know, does

23:11Power BI have the same sort of cloud

23:13offering?

23:15You could argue that Microsoft does, but it

23:17's not set up in the same way. And

23:19potentially

23:20that's where there's some differentiation

23:23to be had. We'll come back to sort of serve

23:25the features and you know, whether this is

23:27all making sense, but nonetheless, let's

23:30keep

23:30going.

23:31The thing we've done is we've actually

23:33brought in a new leader for Tableau to help

23:35steer

23:35our ship forward. And so I'm pleased to

23:38welcome to Tableau conference, Ryan Aitay,

23:42the new

23:42CEO of Tableau.

23:44All right. Very good. Great to be here.

23:48Welcome to Tableau conference. Thank you.

23:51And by the

23:51way, I have to say it is a fantastic place

23:53to be in the center with you and with our

23:56community and with our customers and our

23:58partners and all of our employees. And I

24:00just want

24:01to say thank you for all of your support.

24:04It's amazing. Now, Ryan, you spend a lot of

24:08time with customers. What are you hearing

24:10from customers? What are they telling you?

24:13What should we do?

24:14They're telling us a lot and we're

24:15listening. And I think the big thing that

24:17customers are

24:17speaking to right now, you talked about it,

24:19is cloud, right? And the reason they want

24:21to be focused on cloud and migrating to

24:23cloud is this opportunity to go faster,

24:25this opportunity

24:26to have better performance, the opportunity

24:28to also make sure that you're getting cost

24:30efficiencies in your business. And of

24:32course, the opportunity to take advantage

24:34of some

24:34of the great things you're going to talk

24:36about today, which I won't reveal right now

24:37, but

24:38you can reveal later.

24:39All right. Is there one thing you're very

24:40excited about for the future ahead?

24:42I would say I'm excited about a lot of

24:44things. And one other thing I want to talk

24:46about really

24:47quickly is one other thing I heard.

24:49One?

24:50Sorry, one thing.

24:51Okay.

24:52One thing is when we talk about other

24:54things we hear from customers, they also

24:55talk about

24:56support and the need to help our customers

24:58across the board. And so one thing...

25:00I'm sorry, that was a really labored one

25:03thing. I hear one thing, one thing, one

25:05little thing.

25:06It's about customer support. Tableau does

25:09not hear about customer support. Tableau is

25:13being bombarded about customer support

25:17nonetheless. It's definitely something

25:20where something

25:21has gone wrong with the experience and it's

25:23definitely not how it was designed to be.

25:26And it's definitely broken in certain parts

25:28. And I don't know if that's because of just

25:30the lack of manpower. Maybe it's actually

25:32that growth we're seeing. Maybe the fact

25:34that

25:34the growth has happened so quickly,

25:37accelerated, and yet maybe the support

25:40structure hasn't

25:41changed to sort of meet that growth. Maybe

25:44that's potentially what it is. But it's

25:45definitely

25:46obvious if you've been experiencing Tableau

25:48support over the last, let's say, decade,

25:51you can definitely plot a trend. And in the

25:53last couple of years, something's not right

25:55.

25:55And Ryan here is essentially acknowledging

25:58it on stage. It's definitely what customers

26:00are talking about. Let's hear what he has

26:02to say.

26:02We're going to go through today is the

26:04Tableau success plan. We're going to be

26:06launching

26:06that for all of our customers later this

26:08year. And I just think it's something I

26:10want to

26:10actually call out to make sure you're all

26:12aware of it. We are listening to you.

26:14Okay, good. All right. The other thing you

26:20asked me about the future. I would say the

26:22future is amazing in the sense that we have

26:24so much to look forward to with generative

26:26AI. And that way we're trying to do things

26:28in a safe and secure way. You may have

26:30noticed

26:30last week we announced Slack GPT. Previous

26:33to that, we announced Einstein GPT a few

26:36months

26:36ago. There's some really great things

26:38coming this week, and I'm excited about

26:40that. You're

26:40going to hear about it. If I then go... H

26:43int hint, Slack GPT, Einstein GPT, drum roll

26:48,

26:48please, Tableau GPT. I don't know. I really

26:52enjoyed to sort of listen to this during

26:54the

26:54keynote. Like it was pretty obvious what he

26:58was saying. But yeah, every company under

27:03the sun is going to do something GPT. Chat

27:06GPT is kind of showing people that this is

27:09a really powerful technology that everyone

27:11should be paying attention to and done

27:13right

27:13with the right sort of guidance rails and

27:15the right sort of deployment. It can

27:17actually

27:18save a lot of time in certain use cases,

27:20and those use cases need to be explored. So

27:23yeah.

27:24And I think about where's the future,

27:25Horizon 2, and where we're going beyond

27:27this. It's

27:28about Data Cloud because Data Cloud is a

27:30great opportunity, and it's powered by

27:33Tableau,

27:33which is fantastic. We love data. This to

27:37me is like incredible. Salesforce Data

27:40Cloud

27:41is powered by Tableau. Just listen to that

27:43for one second. Salesforce Data Cloud is

27:45powered

27:45by Tableau. This is to me the biggest

27:48signal of why Salesforce purchased Tableau.

27:52We might

27:52think it's because Salesforce needed an

27:54analytical product. Yes, of course it did.

27:56Einstein Analytics

27:57was nowhere near as capable as something as

28:00the Tableau platform. But more importantly,

28:04Salesforce Data Cloud is now powered by

28:06Tableau. Later on in the conference, this

28:09is termed

28:10as hyper. Salesforce Data Cloud is being

28:12powered by the innovation around hyper and

28:15some of

28:16the capabilities around that. And so it

28:18kind of shows to you that Salesforce didn't

28:20just

28:20acquire Tableau for the sake of it. There

28:23is an IP, there's innovation inside of

28:26Tableau

28:27that was pretty unique to Tableau that made

28:29it a valuable prospect to go and solve a

28:31potentially

28:32huge problem inside of Salesforce. And it's

28:34clear that those investments are now

28:36starting

28:36to pay through for Salesforce. Many Sales

28:39force users may never notice that something

28:41changed.

28:42Other may potentially things getting faster

28:44and more features coming out. But for Table

28:46au

28:47users, it's really important to call this

28:48out that now Salesforce is powered by Table

28:50au.

28:51That's the heart of the setup. And the last

28:54thing I'll say really quickly is that our

28:57community,

28:58all of the people here in this room, all of

29:01our customers, all of our partners, we need

29:03you. And we will continue to invest in you

29:06and we will continue to bring it to life.

29:08And I think that's really important to talk

29:10about. I want to accelerate that in our new

29:12environment together in partnership with

29:14you and others in the business to make sure

29:17that

29:17we're bringing you forward. Right? And that

29:19relates to the future as well with data

29:22skills

29:22and literacy. And that is the future of

29:24Tableau. Love it. Well, welcome to Tableau

29:27conference

29:27and welcome to the Tableau team. Very good.

29:35Now, a really important aspect of the

29:37economy

29:38of the future and for all of us is data

29:40skills. We know that investing in data

29:43literacy is

29:44critical to the success of our companies

29:47and to the future success in our jobs. That

29:49's

29:49why the end of 2021 we launched a pledge to

29:54train over 10 million people on data skills

29:57over the next five years. 10 million people

30:01. This is so critical. And I'm pleased to

30:03say

30:04that we've already exceeded our year one

30:06goals. We've already trained 1.9 million

30:09people in

30:10the first year and a half alone. That's

30:14amazing. I just wonder what is that number?

30:16What is

30:171.9 million people? Like trained who? Like

30:23students? Users? How is that number

30:28comprised?

30:29And to be honest, when they said 10 million

30:32data learners by 2027 in like the previous

30:35conference, I was kind of intrigued by how

30:37they measured that because I thought it was

30:39actually quite an ambitious plan. But

30:42nonetheless, it's happening. So I'd be

30:44really curious to

30:45see how that number is broken down. Like

30:48what are the groups? What are the different

30:51areas?

30:52And a big part of that is because of all of

30:55you, because of this community and our

30:57community

30:58is stronger than ever. We now have more

31:01than 3 million people on Tableau Public,

31:04over 7

31:05million visualizations on public. It's

31:07amazing. The work that you guys do is

31:10inspiring. We

31:11have over 350,000 people on the forums

31:14helping each other, making each other more

31:17successful,

31:18supporting each other to be great. And we

31:20have incredible user groups all over the

31:23world.

31:24Right. The metric that really matters when

31:26you look at these numbers is in fact active

31:28users, active profiles, active visual

31:31izations. How many of these visualizations

31:34are published

31:35within the last year? And how many of them

31:38are getting visited? That to me is the kind

31:41of really important metric. 350,000 Tableau

31:45forum users. Okay. How many of them are

31:47active?

31:48And by active, I mean, logged in in the

31:50last month or in the last year. Come up

31:52with some

31:52sort of measure and tell us about sort of

31:55how you define that because the aggregate

31:57is one thing, but there'll also be many

32:00millions of users here that overlap many

32:02other analytical

32:04tools. And so that is less unique than

32:07potentially quoting these sort of high

32:10level numbers.

32:11Oh, yeah. 70 countries. The community is

32:16thriving like never before. And so to tell

32:20you more

32:20about the Tableau community, I'm pleased to

32:23bring up Larissa Amoroso, vice president of

32:26the Tableau community. Thank you, Francois.

32:32And hello, Data fam. I am so thrilled to be

32:37here with all of you today in Vegas at

32:40Tableau Conference. Now, every single

32:43person in this

32:44room and all of you tuning in online have a

32:47unique connection to the Tableau community.

32:50I've heard from many of you that Tableau

32:53and this community have had a profound

32:56impact

32:57on you. Now, beyond your personal lives,

33:00many of you have changed the course of your

33:03organizations

33:04and the world for the better. Big change

33:08like this takes exceptional leadership. And

33:13that's

33:13why I'd like to take a moment to give a few

33:16shout outs. Thank you to the more than 100

33:19news group leaders right here in this room

33:21and hundreds more tuning in from around the

33:24world. Thank you for creating a space to

33:31connect, to learn, and just to talk all

33:35things data.

33:37Thank you to our ambassadors who champion

33:42Tableau and our mission. They inspire

33:47creativity.

33:48They foster and nurture technical growth in

33:51our community. So ambassadors and user

33:53group

33:54leaders, everything that you do brings

33:58Tableau to life. Now, there is one more

34:01group of people

34:01that I like to recognize. >> I was going to

34:04pause it here and say the

34:06people who play these roles, user group

34:09leaders and ambassadors, it's very much a

34:11thankless

34:12task. They do so much work behind the

34:14scenes, whether it's coming up with new

34:16ways of using

34:17the product, organizing opportunities for

34:19people to get together. User group leaders

34:21in particular, they act as the catalyst for

34:24companies, groups, specific communities,

34:27sectors,

34:29industries to share knowledge in a safe and

34:31collaborative way. That is always sort of

34:33an understated thing. When you go to a user

34:35group and someone gets up and shows you how

34:38they're doing something in their

34:39organization with a bit of anonymized data,

34:41that in most

34:42industries is completely unheard of because

34:45that's sort of considered competitive

34:47advantage.

34:49This happens every single week. You only

34:51got to go to YouTube and see all the user

34:53group

34:54recordings that are being put out. If all

34:57you did is watch those user group

34:59recordings

35:00for one year, I promise you, you'd learn so

35:02much more about Tableau than any course,

35:05any

35:05YouTube channel at all. So go ahead and

35:07really engage with those as much as

35:09possible. Then

35:10ambassadors sometimes do straddle the user

35:13group. So the user group leaders are

35:15sometimes

35:15ambassadors but nonetheless, ambassadors

35:17also tend to make a lot of great content, a

35:19lot

35:20of really unique content. They also deserve

35:23a big shout out to a huge group of people.

35:25You can see this is a photo of just the

35:27people at conference. There'll be many more

35:28who couldn't

35:29make it in the whole setup. So let's keep

35:32going.

35:33These are the people who raise the bar for

35:35what's possible with Tableau. We call these

35:38leaders Tableau Visionaries. Nominated by

35:47the DataFam at large, this year we are

35:49honoring

35:4950 Visionaries, our largest and most

35:54globally diverse group to date. And I'd

35:57like to take

35:58a special moment to honor our two newest

36:00inductees into the Visionary Hall of Fame,

36:03Anne Jackson

36:04and Ken Flurlidge. It was Ken. Apologies,

36:11Ken. I'm so sorry for getting it wrong.

36:14Now Anne and Ken joined 14 others who have

36:17served as Visionaries for five full years.

36:20They will hold the title of Tableau Vision

36:23ary in perpetuity. So congrats to Ken and

36:27Anne.

36:27Now all Visionaries, please come join me up

36:33on stage.

36:34Okay.

36:36Okay. Get in here. Don't be shy.

36:51Now Visionaries, you inspire us. You teach

36:59us. And you push us to do more. You empower

37:04us to deliver on our mission to help people

37:07see and understand data. And for that, I

37:09just

37:09want to say a huge thank you. So thank you

37:13Visionaries and thank you DataFam.

37:15Now where's Francois and Ryan and Pedro?

37:18Get up here. We're going to take a quick

37:21selfie.

37:22Okay. Gosh, if I was at conference, I think

37:26I'd be faced with the dilemma of taking

37:29part

37:29in this myself. On one hand, you know, let

37:33me put myself in my own shoes as a Tableau

37:36Visionary. On one hand, you know, I do what

37:40I do because I enjoy it. I think it helps

37:42a lot of people. I hope it helps a lot of

37:45people. And if it does, I don't expect

37:47anything

37:47back for it. That's sort of always been the

37:50mantra. If you do, if you do go into these

37:51things expecting something back, what tends

37:54to happen is you burn out quite frankly,

37:55because

37:56you're kind of looking for the recognition

37:58in a different context and it doesn't work,

38:00right? It just, there's just something

38:03about our con. I can't really explain it

38:04well. That

38:04to me is just a really important factor.

38:06And so just stepping back and just say,

38:10okay,

38:10Tableau Visionaries, they all do

38:12sensational things. Don't get me wrong. I

38:14envy every single

38:15one of these people for a skill that I don

38:18't have because they're all able to do

38:20something

38:20that is incredible. And I think everyone in

38:23this group who's been a visionary before is

38:25a visionary this year absolutely deserves

38:28it. But there's one small but I think this

38:31moment on stage is a little bit awkward.

38:34And if you're not part of the community and

38:37you're

38:37not part of the 1% who actually gets what's

38:41going on here, this just looks a little bit

38:43strange. And I think at this point when

38:47they got up and you know, everything

38:50happened fine,

38:51but then it's it's the selfie. And I know

38:54it's great to be part of the selfie, but

38:56like,

38:56I don't know if you were in the crowd or if

38:59you knew in the audience, let me know how

39:01was this part of conference? Did it feel

39:03strange? Is it for weird? I don't know. But

39:05to me,

39:05I just think, you know, I've never I've

39:09never been at a client and I've never gone

39:12anywhere

39:13in like a general setting with everyday

39:16people that use Tableau where they even

39:19know that

39:19there is a visionary or an ambassador or

39:22any sort of recognition program. They just

39:23use

39:23Tableau. When someone goes and watches a

39:26video explaining what Tableau is, I

39:29guarantee you

39:29there'll only be like hundreds of people

39:32potentially who even know that there's a

39:34recognition

39:34program in the Tableau ecosystem. But there

39:37are hundreds of thousands of people who

39:39just

39:39want to use Tableau and just get on with it

39:42. Quite a few of those people come to this

39:44conference

39:44for the first time. I think this moment on

39:47stage, I think it's just a little bit

39:50awkward.

39:50I don't want to be the person to say that

39:52it shouldn't be done or whatever like that

39:54is that is an absolutely you know, Tableau

39:57's own decision. And it's absolutely sort of

39:59up to them how they orchestrate this. But I

40:04just do think that I personally I would

40:07really

40:07struggle to be part of something like this

40:09and say I don't know, I'd probably have to

40:10sit it out because I just it just it's just

40:12a little bit weird.

40:13All right.

40:17You guys ready? Ryan, Pedro, come on in.

40:23All right. Ready? One, two, three, Data!

40:29Amazing.

40:31Congratulations.

40:32Pretty cool.

40:33Off to you. Okay. Thank you everyone.

40:35This community is just so inspiring. It's

40:43amazing. I mean, look at what they do and

40:45look at how incredible they are. Together

40:49we are doing amazing things. And today we

40:53're

40:53using data in completely new ways. I mean,

40:57think about it. 20 years ago, data were

41:00boring

41:00tabular reports and together we've made it

41:04visual, fun and interactive. You know,

41:08today

41:08data is everywhere. It's gone mainstream.

41:12We saw it through the pandemic where we

41:13were

41:13looking at data every day to see where the

41:16cases are. But data is also powering new

41:19experiences.

41:19Think of Netflix. The recommendation feed

41:22is powered by data and AI. And data is

41:26changing

41:26sports. Just look at baseball. Baseball is

41:29now completely data driven. I mean, data

41:32has

41:32completely changed the game. And with new

41:36technologies like the cloud and generative

41:39AI, we're going to change the game again

41:42together. But we all know that data is

41:47still too hard.

41:48It's not accessible to everybody. There's

41:52still too many people that don't have

41:54access

41:54to the data.

41:55So at this point, we've come straight off

41:58the community event and we've gone into

42:00what I called glue in my keynote summary.

42:04Somehow we need to, Francois is building a

42:06keynote. He's building a bridge from all

42:09this community fanfare, celebration of the

42:10past, celebration of the people, everything

42:13. He needs to build a bridge, ironically,

42:15over

42:15the chasm to the core content of the

42:19keynote. So this slide is both like a

42:23perfect illustration

42:24of what is explaining, but also ironically

42:27explaining the situation in the keynote

42:29that

42:29he's in. And so what he's trying to do is

42:32say, hey, look, these are the amazing

42:33things

42:33that we can do with data. Here are some

42:36opportunities for the future. But here's a

42:39core problem

42:39in the business. And the problem is that

42:42data is not accessible to all. People have

42:44limited

42:44skill, missing context, hard to trust. Only

42:4730% say actions are driven by data.

42:49Now, what I find super interesting about

42:53this particular slide, in fact, let me say

42:55this

42:55after this slide is finished.

42:56To make better decisions. In fact, only 30%

43:00of people in organizations use data to make

43:02decisions. So what do the other people do?

43:05They just make it up or make decisions

43:08based

43:09on gut feel? Well, we have to help everyone

43:12. And part of the challenge is a lack of

43:14skills.

43:16They lack context to make the right

43:18decisions.

43:18So to me, I'm stopping right here because

43:22this is the moment. And this is going to

43:24come

43:24up in the theme that I think has been

43:27shared by lots of people at conference,

43:29which is

43:29I didn't see anything that helps me as a

43:33user of the Tableau desktop authoring

43:36experience.

43:36But I think Tableau is calling it out right

43:39here. You are that 30%. And this slide is

43:44kind of saying, hey, data is not accessible

43:47to all data's over here. And only 30% of

43:50you

43:50and little circles of the people in the

43:53community. Yeah, you people, you're doing

43:55it well.

43:56You're happy, you're successful. But

44:00everyone else can't cross this gap because

44:02they have

44:02limited skills. They're not data analysts,

44:05they're not data engineers, they're not

44:06dashboard

44:06authors, and they all can't be for the

44:08record, you can't possibly train up

44:10everyone in an

44:10organization to use Tableau desktop to get

44:13to an answer. That's just really not

44:15realistic.

44:15And even if you did do that, you'd have too

44:18many people using tools they don't

44:20understand

44:20and therefore it becomes hard to trust. So

44:24in a way, I kind of feel like this is,

44:28Tableau

44:28literally saying it here, and they've been

44:31saying it for a while. Everything we're

44:33about

44:33to tell you is not for this 30%. Everything

44:37we're about to tell you is for the 70% that

44:41come after. This is for everyone else. And

44:45so when the keynote then starts to become

44:47about Tableau, past all these other great

44:49technologies and everything coming up, I'm

44:51not surprised by all of that because Table

44:53au calling out here, and I feel like they've

44:55been calling out keynote after keynote,

44:58session after session, feature after

45:00feature, release

45:00after release. And we'll get onto that a

45:03little bit more, but I think it's here in

45:06black and

45:06white. They're literally saying it right

45:09here. They want to help everyone else, not

45:12just

45:12this 30%, which is where they've sort of

45:15mostly focused their effort in the past.

45:17They're lacking trust in the data. Well,

45:21Tableau is about empowering everyone. And

45:23this is why our mission to help people see

45:27and understand data is...

45:29I feel like Franzo just underlined people.

45:33People, not like you ninjas in the room

45:34here,

45:35not people who are mostly attending this

45:38conference, literally everyday people,

45:40people who will

45:40never open a calculation window, people who

45:43will never open desktop, people who will

45:45never

45:45open prep, people who will never administer

45:48server, those people. That's who I think

45:50they're really trying to emphasize. Without

45:52saying it, they're really trying to

45:53emphasize

45:54that. And actually they do spell it out a

45:56little bit more throughout the keynote.

45:58More relevant today than it's ever been,

46:00because our mission is about empowering all

46:03people to get insights and generate meaning

46:07from data. It's about getting the best out

46:09of everyone so that we can have impact on

46:13the world together. But we still have more

46:17work to do. We still have to push forward

46:20to help everyone with data.

46:22You're saying it again. I can't stress it

46:26enough. You said everyone, more people,

46:28everyone, not just the people who have

46:31access. I feel like he's just saying it

46:34again

46:34and again. He's trying to really drum the

46:37point home. If I'm wrong about this, let me

46:38know, Francois. I feel like that's what you

46:42're saying.

46:42This is why today is a new day for data.

46:47And there you go. We want to help everyone.

46:51We want to help everyone. This is why today

46:53,

46:53this keynote is a new day for data. Today,

46:57we focus on the 70% of people we've not

47:00helped in the past. If you walk away from

47:03this keynote thinking, "Hey, I didn't see

47:05anything from me as an author." Well, this

47:08is why. This is their focus.

47:10And that's not to say it's the right focus.

47:12That's not to say that everyone at

47:14conference

47:14should be happy and celebrating and ringing

47:17off cheer bells and everything. Absolutely

47:18not. If you're an author who's invested

47:21years in Tableau and you're sitting here

47:23in the audience and Tableau is saying to

47:25you, "Hey, you're not our key focus for the

47:27next few years." That is a tough message to

47:30take and that is a tough thing to hear.

47:32But if I'm paraphrasing this correctly,

47:38that is what's being said here. And I don't

47:38think it could be more clear if anyone in

47:41this keynote tried to understand it that

47:43way.

47:43Today is a new day for Tableau. And

47:48together, we're going to build an amazing

47:49future together.

47:51Now, it's a new day for consumers. We're

47:55going to deliver new…

47:56The first thing, a new day for who?

47:59Consumers. That's the first thing. Never

48:03has that happened

48:04in a Tableau keynote. In the past, if this

48:08was a new day for data, it would have been

48:10for analysts, for dashboard builders, for

48:13creatives. Those have been the headlines of

48:16the past. It would take me way too long to

48:19go back in the keynotes and find out what's

48:20the core, overlying message. But if I think

48:23back to the Christian Chabot days when we

48:24were talking about data storytelling and

48:27visualization, the focus there was on

48:30enabling people like

48:31me, people like you, probably watching this

48:34video, on how to build beautiful things and

48:35get those things into the business. But

48:38here, it's saying a new day for data.

48:39Consumers.

48:40Consumers don't build anything. They just

48:43consume. They're viewers. They are

48:45explorers,

48:45potentially. They just consume.

48:48Experiences that helps everyone close that

48:52insight to action loop and bring data in

48:54the

48:54flow of everybody. It's a new day for

48:59developers.

49:00Group number two. It's a new day for people

49:04who want to build on top of Tableau, people

49:05who want to extend Tableau to places we're

49:07not interested in going, but we'll give

49:09you the tools to go and do that.

49:12Tableau enables every developer to bring

49:16the power of data and analytics to every

49:18single

49:18application. It's a new day for IT, to help

49:23IT build trusted data at scale so that we

49:26have a single source of truth. And it's a

49:29new day for analysts as well.

49:32IT, people who manage these infrastructures

49:36, but also notice he talked about metadata.

49:38He talked about setting up servers. At the

49:41beginning, he called out that IT are doing

49:44less of that because of Tableau Cloud. So

49:47when he talks about IT, he's really talking

49:48about unearthing the metadata, making sure

49:51things interconnect and work better

49:53together,

49:53making sure they can do governance and

49:55apply their policies and everything in a

49:57nice, simple

49:58and clear way. And then when we go to

50:00analysts, I think this is the most

50:02interesting thing

50:02because I think this term analyst is

50:05confused throughout the industry.

50:07Why does everyone uses the term analyst too

50:09much? Because what is an analyst?

50:11An analyst could be someone who works

50:14inside of a hedge fund analyzing a specific

50:16market.

50:16An analyst could be someone who works

50:18inside of an organization sifting through

50:20data day

50:20to day, but that doesn't describe what an

50:23analyst does. An analyst is a role.

50:25And so when Tableau calls analysts, I think

50:27it's really important to realize they're

50:28not talking about authors. They're

50:32fundamentally talking about people whose

50:34job it is to take

50:35data, process it in some way, whether it's

50:38mental through analysis or through some

50:41sort

50:41of reading, some sort of consumption

50:44process, and then relay it back out to the

50:46business

50:46in some way or form so they can make a

50:49decision. That's what fundamentally

50:50analysts do.

50:51It didn't say data analyst, didn't say any

50:54of that, just analysts full stop.

50:55And so it's an incredibly broad term. I

50:59think it gives Tableau a get out of jail

51:00free card

51:00to say, hey, we're talking about you and

51:02everyone in the room. But actually, I just

51:04don't think that is fully the case. When we

51:08get to devs on stage, yes, you could say,

51:09yes, there's some stuff for the analyst

51:11there. But fundamentally, the stuff they

51:12lead with

51:13is more consumption, more analyst sort of

51:19focused than it is someone, an analyst who

51:22's

51:22going to go and build stuff and do stuff.

51:24But anyway, it's not clear sort of what

51:26that word means. And I think it's often

51:28taken for granted. Anyway, let's keep going

51:30.

51:30We have to continue to make Tableau easier

51:33and more powerful so you can answer richer

51:35questions. We have to remove the burrs and

51:38keep you in the flow. And you're going to

51:39see us double down on our core community

51:42this year. It's going to be awesome.

51:44So it's a new day for data.

51:47At the end of the day, he said he's going

51:50to double down on the core community.

51:52Fine, that makes total sense. Again, what

51:55does that actually mean? We'll get back

51:57to that later. So here we are. Here are the

52:00new tenants.

52:00In this keynote, we'll show you the latest

52:03and greatest…

52:03New day for data. Data augmented by AI,

52:06data embedded everywhere, and data trusted

52:09by all.

52:09Those are the three key tenants. It's all

52:13about data. It's not about storytelling

52:16as it's been in the past. It's not about

52:19analytics. The term analytics in Tableau

52:22has always meant dashboarding, calculations

52:25, all of that stuff, the analytical pipeline

52:27.

52:27This is all around how do we coalesce

52:30around this term, data, and how do we make

52:34all these

52:34sort of core activities happen?

52:36Innovations coming to the Tableau platform.

52:39First, we're going to talk about how we're

52:41going to augment data with AI so that we

52:44can bring the power of data to everyone.

52:46We're going to talk about how every

52:49application is going to become an

52:51analytical application.

52:52And we're going to talk about how we can

52:55share data at scale in a trusted and

52:58scalable way.

52:59And of course, we're also going to have

53:02devs on stage.

53:05And there's the core community. So can I

53:18just say, I think that to me, just tie that

53:18in a bow and present it.

53:18You've got these items on stage, a new day

53:21for data, augmented by AI, embedded

53:23everywhere, trusted by all.

53:24Silence. And then as an aside, we've got

53:28devs on stage too. Then the immediate crowd

53:31goes wild.

53:32To me, that just says it all. This stuff

53:36here, it's not for the core community.

53:38The majority of this keynote is about to be

53:41not for the community.

53:42For you, we've got devs on stage right at

53:45the end.

53:45And I think that, again, just to draw my

53:48message home, this is how it is.

53:50Oh yeah. And let me just say there's a few

53:55mic drop moments in devs on stage.

53:57Or should I say pineapple drop?

54:01All right. So let's get it started. Let's

54:04talk about how we're augmenting the data

54:06experience with...

54:08So that pineapple reference, you'll get it

54:04when we get to the Chirinomiya demo at the

54:10end,

54:12because Matt brings out the pineapple. Hint

54:15, hint. This is the bit to pay attention to

54:17anyway. Let's carry on.

54:18And for that, I'd like to welcome Pedro A

54:21riano, the new Senior Vice President of

54:24Product.

54:25Welcome to the stage, Pedro.

54:27Thank you, Francois. Thank you so much.

54:30Thank you, everybody. I can't tell you how

54:35special it is for me to be here.

54:37I want to give a special personal shout out

54:40to the Latino community, comunidatos.

54:42Thank you for this shirt.

54:44It's so special for me to be at Tableau

54:48Conference as part of Tableau,

54:51because I remember my first Tableau

54:54Conference years ago.

54:55I was a spectator. I was sitting in the

54:58audience, just like all of you,

55:00and I was surrounded by so many of you,

55:02thousands and thousands of very passionate,

55:04very excited, very loud people.

55:07And I remember you all losing your minds

55:11because somebody on stage showed this

55:14little tiny new option on a menu bar.

55:15I was like, "What is this?" It's like, "I

55:18've never seen this before."

55:19And that's exactly what people are thinking

55:22when everyone's on stage, you know,

55:24whatever.

55:24There's a whole bunch of first experiences

55:26that people don't get, and I don't know.

55:29But it was just so special and so exciting,

55:35and it made me think about how truly

55:37innovative and disruptive Tableau has been

55:40over the years.

55:40Tableau empowered millions of people like

55:44you from all kinds of backgrounds,

55:45different types of expertise, to work with

55:48data in a way that just wasn't possible

55:51before.

55:51It ushered in a new era of analytics. It

55:55allowed you to express yourselves

55:57creatively through data.

55:58It actually made data fun. Think about that

56:01.

56:01You know, we came from financial statements

56:04to somebody building a viz that shows the

56:06history of Fleetwood Mac.

56:07That's a real viz, by the way, if you're

56:10out there, shout out to you. That viz

56:12exists.

56:12Well, now we are entering a new era of

56:18disruption.

56:20And that is the emergence of generative AI.

56:23Generative AI has the potential to become

56:28the biggest, most transformative technology

56:31that we see in our lifetimes.

56:33And it's just so early, we're only

56:36beginning to understand the potential that

56:39it has,

56:39but we're already seeing glimpses of how it

56:43can alter everything around us.

56:46And that is exciting. It's also a little

56:49bit scary. However you feel about it,

56:52there is no question that generative AI is

56:55an incredible technology achievement.

56:57So, I think Badro is basically

57:01acknowledging the anxiety around AI,

57:04how quickly it's moving and how I think

57:07businesses feel like, well, they want to

57:09get in on this,

57:09but they also want to do it well and they

57:11want to do it safely and they also don't

57:12want to break any laws.

57:13The laws around this are pretty immature,

57:17if let's say the least.

57:18Regulation is always quite late to react to

57:20this kind of stuff.

57:21So it's interesting that Badro is sort of

57:24just acknowledging that,

57:25teeing people up to say, hey, you know, all

57:27these things you're thinking it, we're

57:29thinking it too,

57:30but we've taken all of that into account

57:33and that's sort of the implication.

57:34And yeah, we're going to show you how we

57:37think we can do this well in the context of

57:39a business.

57:39What does it mean for all of us?

57:41What does it mean for Tableau and its

57:43community?

57:43Well, all of you know that Tableau has been

57:47investing in artificial intelligence

57:48and augmenting analytics for a number of

57:50years.

57:50You're familiar with capabilities like Ask

57:53Data, Explain Data, Data Stories.

57:55What generative AI does for us is it allows

57:59us to take these investments to the next

58:01level.

58:01And that is why today we are unveiling to

58:06the world

58:09Tableau GPT.

58:11Tableau GPT is a new...

58:20This is the moment Ryan Atay called out.

58:22He said, we've done Slack, we've done

58:24Einstein, here is Tableau GPT.

58:26And you can even see Einstein's on stage

58:28here.

58:28... suite of capabilities that allow us to

58:31infuse the power of generative AI across

58:34all experiences in Tableau.

58:36Experiences like analytics that feels more

58:40like conversational Q&A and less drag and

58:43drop.

58:43Experiences like anticipating the questions

58:47that a user might ask because of having an

58:49understanding of the data.

58:50Or an experience like taking dozens or

58:54hundreds of insights and explaining them

58:56using very easy to understand summaries.

58:59Tableau GPT is a game-changing innovation

59:03and we're going to disrupt the industry

59:05once again.

59:06This is going to help us fulfill our

59:09mission of helping everyone see and

59:11understand data.

59:12So it's a pretty bold statement there

59:16because what Pedro is saying is he's kind

59:19of calling out the success before it's

59:21happened, right?

59:21And I think as a leader you do that because

59:24you're confident about what it's going to

59:26enable you to do.

59:26And you're also confident about the skills

59:28you have inside of the organisation to

59:30deliver on that.

59:31And everything you just said there was a

59:34vision.

59:34It's not something that's come to reality.

59:37And actually we do see some demos but I'll

59:39get into the demos a little bit later on.

59:41Everything you said there is teeing up for

59:45the kind of expectations, essentially

59:47setting expectations actually, for the

59:49features you should expect to see.

59:50Nonetheless I think those kind of messages,

59:52people don't remember that because what you

59:54remember is exactly what they show you

59:56after.

59:56So he's saying it here, we'll see it in the

59:59demos but you'll remember the demos and I

60:02think you'll walk away from the demos

60:03thinking, ah yes, this is what it's like.

60:06But in reality none of this stuff has fully

60:09been fleshed out, still in development,

60:12still being worked through.

60:13And yeah, we'll see a bit more of the

60:15detail soon.

60:16[Applause]

60:21So how does it work?

60:23So Tableau GBT brings together the power of

60:26Salesforce's own proprietary AI models, our

60:30large language models from our ecosystem of

60:33partners and the power of Tableau Analytics

60:36.

60:36Surrounding all of this is a very important

60:39layer of security and governance.

60:41Because we understand that as we invest

60:44more and more in generative AI, our

60:47responsibility as trusted stewards of your

60:50data becomes even more and more important.

60:53So now I'm going to introduce one of the

60:57first places where we're going to leverage

61:00the power of Tableau GBT.

61:02Our new groundbreaking innovation,

61:06introducing Tableau Pulse.

61:07Okay, so I'm going to stop this here. I've

61:11done a video on Tableau Pulse, go watch it.

61:13It's probably going to be much shorter than

61:16the blurb I'm about to go through here.

61:17And for that reason I'll probably skip

61:19through various bits here. We'll see the

61:21demo anyway.

61:21I'll have to skip the customer parts

61:24because there's a small issue with me doing

61:27videos about customers and putting it on

61:29YouTube.

61:29We'll get the story for another day.

61:31Anyway, that whole teeing up just there,

61:35what he just did here, I think where he was

61:38talking about the innovation.

61:39This is a bit of a nothing diagram. I call

61:43these a nothing slide.

61:44A slide with something on it but you haven

61:46't really got anything to show.

61:47So you make something and it's just two

61:50things with a spinning arrow going to each

61:52other and you put them in some boxes and

61:55you say concept and you go, "dunno, this is

61:57how it's going to work."

61:57But all it really means is the language

61:59model is working in partnership with the

62:01analytics stuff.

62:02So the stuff you know, the analytics, great

62:04. We're adding language models right next to

62:06that and we're going to help those things

62:07understand.

62:07And by the way, these things run in a

62:10container. The container is security and

62:12governance and data sources are also part

62:14of that vision.

62:14And so, yeah, I think that Pedro felt he

62:17had to sort of tee that up to say this is

62:20how it's going to work and don't worry, we

62:23're doing this, we're thinking about this.

62:24A repetition again from the very beginning

62:27but nonetheless, yeah, here's where we are.

62:29We're going to leverage the power of Table

62:32au GBT. Our new groundbreaking innovation.

62:36Introducing Tableau Pulse.

62:37So the pilot is in half to 2023. So

62:41basically just after summer.

62:42So think of this as going to be released in

62:46either the 23.3 or 23.4 as a pilot and here

62:52's the thing, as a pilot.

62:54So Surface Personalized Metrics, Inter-R

62:57otation to Workflows, Get AI-Powered Ins

62:59ights at Scale.

63:00Now, when we go through this feature set

63:04and in my video, I called out straight away

63:08, I bet you this is going to be a Tableau

63:10Cloud Only feature.

63:11I don't have confirmation of that. I don't

63:13have confirmation, I don't know the answer

63:15to that.

63:15But my hunch is that this will be a Tableau

63:18Cloud Only feature straight out of the gate

63:20.

63:20Why? Well, number one, server only gets two

63:24updates a year.

63:24So if there's a pilot in the second half of

63:27the year, it's definitely not going to be

63:29in server.

63:29So the pilot will be in Tableau Cloud as a

63:31starting point.

63:32Again, extrapolate that a bit further.

63:35Tableau Cloud will be the place where Table

63:38au can do what it thinks it needs to do to

63:41power this, i.e. not have you worrying

63:43about large language models and data and

63:45everything running because you're in charge

63:47of that infrastructure.

63:48You can do that in a way that's secure and

63:50trusted and governed.

63:51And as Tableau Cloud does, it has all the

63:54compliance and capability setups already.

63:57So that is a much better place for this

63:59stuff to be done, as Tableau would say.

64:01That's an important thing. Now, if Tableau

64:03said actually we're going to do it on

64:05server as well, there's just a whole bunch

64:07of other questions that suddenly need to be

64:09answered.

64:09For example, if this stuff is really

64:11running AI models and everything, AI models

64:14have completely different compute

64:16requirements to everyday servers and

64:19systems.

64:19Just even a simple activity of training up

64:23an AI model takes ridiculous amounts of

64:26compute power.

64:27We're talking really ridiculous amounts of

64:30compute power. And to do that for every

64:33single organization at scale is much more

64:34cost effective.

64:35That's why services like ChatGPT and Google

64:38Bard and everything as it comes out are much

64:41more cost effective to use because all of

64:45that is being centralized.

64:45They're sort of in economies of scale.

64:47But for every organization to suddenly

64:49invest in the capability and beef up their

64:52servers to run this kind of stuff, it just

64:54would never work.

64:54So that's why I think this is going to be a

64:57Tableau Cloud only feature right out of the

64:59gate.

64:59But again, we'll come back to this a little

65:02later.

65:02[Applause]

65:07Tableau Pulse is a completely reimagined

65:11Tableau experience.

65:13It's going to help you, it's going to make

65:16you change and expand the way that you

65:18think about working with data.

65:19It uses AI to understand your data, to know

65:23what your goals are, and it helps you stay

65:26on top of them.

65:27And it sends you these surfaced insights

65:30automatically through very personalized,

65:33targeted, very contextual summaries.

65:35It helps you know what to do next.

65:38And all of this is built on this brand new

65:41innovative metrics layer that adds very

65:44valuable context and business understanding

65:47to the insights that you receive.

65:48Okay, so actually there's two things in

65:51there.

65:51There's Tableau Pulse, there's Tableau GPT

65:54sort of working in there, there's AI

65:56working in there.

65:57And then there's also a new thing in there

65:59called a metrics layer.

66:00And the metrics layer in itself is just

66:03simply the capability to define these

66:05metrics.

66:05I highlighted this in my video.

66:06So Tableau Pulse is a bit of a strange one

66:09because even in my video, I present it as

66:12one thing.

66:12I present it as a landing page.

66:14But actually it's actually more than that.

66:16It's almost four or five technologies

66:19coming together to work.

66:20And it's actually five technologies because

66:23there's one technology that I don't mention

66:25here whatsoever that is actually powering

66:27this.

66:27And I think wasn't that obvious in the

66:29keynote, but having talked to a few people

66:31behind the scenes, including at Tableau,

66:33they've confirmed that actually what's

66:36powering Tableau Pulse is ViscuHEL data

66:38service.

66:38And so all of these innovations are coming

66:42together to form an experience.

66:44And I think that's a super important thing

66:46to bear in mind because it really shows you

66:49the scale and the level of effort going

66:52into this direction.

66:53And by this direction, I'm not talking

66:55about Tableau Pulse.

66:56I remember this 2331 in this direction, I'm

67:00talking about what he said here, a new day

67:03for data.

67:03Here we go. Here we go. Here we go. Where

67:05is it? Where is it? Where is it? I shouldn

67:06't have done this, should I?

67:07This problem. It's talking about this

67:10problem.

67:10That feature Tableau Pulse is to help solve

67:13this problem.

67:13And it's not for this 30%. You've got to

67:16keep remembering that.

67:17Like you see these people on screen and you

67:19kind of do this sort of thing where you

67:21morph yourself into one of those people.

67:23I'm one of those 70% that's not being

67:26helped.

67:26No, no. For most people in the audience,

67:30you are the authors, you are the dashboard

67:32builders.

67:32This slide is actually saying to you that

67:35Tableau is trying to help the people you're

67:38serving with this feature.

67:39I just wanted to sort of draw a line

67:42between those two things. Let's go back to

67:44where we were.

67:44Here we go. I think we go back here.

67:49Targeted very contextual summaries. It

67:53helps you know what to do next.

67:55And all of this is built on this brand new

67:58innovative metrics layer that adds very

68:01valuable context and business understanding

68:04to the insets that you receive.

68:06Now let's learn more about this with one of

68:09our customers.

68:10John Lewis Partnership is a brand that is

68:12known--

68:12>> Right. So I have to skip this clip.

68:13Skip, skip, skip, skip, skip, skip, skip,

68:16skip, skip, skip, skip, skip, skip, skip,

68:18skip, skip.

68:19Also, keen faces.

68:21There you go.

68:26>> [APPLAUSE]

68:32>> So who is ready to see Tableau GPT and

68:34Tableau Pulse in action?

68:35>> [APPLAUSE]

68:37>> All right.

68:38>> Sounds like me.

68:38>> Please join me in welcoming to the stage

68:40Vice President of Product Management,

68:43Caroline Sherman.

68:44>> Thank you, Pedro.

68:46>> [APPLAUSE]

68:48>> Hello.

68:49I am so proud of what my team has built and

68:54so excited to finally share it with you.

68:57We are revolutionizing the analytics

69:00experience for everyone.

69:02We're on the cutting edge of generative AI.

69:05>> Just going to say, every time they say

69:07everyone, I just keep thinking in my head,

69:09but not you.

69:10>> With Tableau GPT, and we're empowering

69:16you to put data in every corner of your

69:18company.

69:19All right. Imagine--

69:21>> All right.

69:22She did actually sort of clarify it there

69:25by saying, we're empowering you.

69:26You're going to be using this tool to help

69:29everyone else.

69:30I think that's an important context setting

69:32here.

69:32And it's actually an important ask of the

69:36audience.

69:36They want to show you how they're going to

69:39amplify your capability.

69:40Now, it's interesting because I think the

69:43people in the audience aren't asking for

69:45this, right?

69:45No one who builds dashboards today was

69:48asking for Tableau Pulse.

69:50What they were asking is for the ability to

69:53build things like Tableau Pulse very easily

69:55and quickly in Tableau Desktop.

69:56But instead, Tableau sort of come, yeah,

69:59you wanted that, well, here you go.

70:00But by the way, we're going to scale it up

70:02and you're going to do it in this specific

70:04way

70:04so that we can run our technology and

70:06everything on top of it.

70:07Maybe that's a big summarization, but I

70:09think that's essentially what's going on.

70:11>> I'm a store manager at John Lewis

70:13Partnership.

70:14My job is to drive results.

70:18I need to make informed decisions fast.

70:21Enter Tableau Pulse, a whole new way of

70:26experiencing data.

70:28Right at the top, I get an overview of the

70:32metrics that need my attention with the

70:35most relevant insights.

70:36This whole experience is personalized and

70:40smart.

70:40>> Now, in my sort of rundown, I went

70:42through this.

70:43I did a bunch of things.

70:44What I'll do is I'll sort of break this

70:47down.

70:47First and foremost, Tableau, this demo,

70:51what we're seeing here is an interface we

70:55've not seen anywhere in Tableau.

70:56And so what you have to remember is I'm 90%

71:00certain that everything you're seeing on

71:03screen is a UX demo.

71:05So it's running in something like Figma,

71:07but it represents where Tableau is going,

71:12essentially.

71:12And at the top, you have Tableau Pulse.

71:14It might be an experience you click into.

71:15It might not even be part of the server or

71:18cloud setup.

71:18Instead, it will be something else, a

71:20different interface, a different delivery

71:22portal, somewhere where you don't have to

71:24log in.

71:24It could live inside of Slack.

71:25It could live inside of your phone.

71:27It could live in your emails.

71:28Wherever you go to to experience this, I

71:31don't think it's the standard Tableau

71:32server or Tableau cloud experience.

71:34The second thing is it's actually quite

71:37heavily personalized.

71:38So it starts off with your name.

71:39It tells you here's your pulse.

71:41And it says one unusual, 11 normal.

71:44The unusual thing is actually a feature

71:46that already exists today.

71:47There is an anomaly detection capability

71:50inside of Tableau that looks at your data

71:52and alerts you to anything that goes

71:54unusual in your data.

71:55It's actually two ways.

71:56There's explained data, which can do it for

71:59a data point.

71:59And then there's another feature which I

72:02can never remember the name of.

72:03It always escapes me.

72:04But in essence, it's taking snapshots of

72:07your data and it's plotting specific

72:09behaviors and looking for changes in your

72:11data so you can go and take action on them.

72:13So it's basically telling you there's one

72:15unusual thing, 11 normal things.

72:16So that's not new, although it looks part

72:18of past.

72:18That's not terribly new.

72:20And you've got three sections following for

72:22you and all metrics.

72:23So following the things you subscribe to

72:26for you or maybe things that have been

72:28specifically assigned to you.

72:30And then metrics is just everything,

72:31everything that you've ever looked at and

72:33maybe everything you're ever interested in.

72:35And by the way, that one unusual is this

72:37one unusual thing you've got here.

72:39So although it's unusual, it's actually

72:42unusual, but it's good. It's green. Right.

72:43I'm wondering if there's an unusual but red

72:46, unusual but orange or whatever.

72:47Tableau uses for sort of things that I'm

72:49doing too well.

72:50And it's basically saying, hey, something's

72:53gone up 27 percent compared to the previous

72:55week.

72:55And it's got this little star icon and that

72:59star icon throughout the interfaces we've

73:01seen throughout all these demos has been

73:04Tableau GPT.

73:04So that's where the A.I. is coming in,

73:07building up sentences, doing analysis for

73:09you.

73:09And the chart is just sort of that story.

73:11But the A.I. is building out that sentence

73:14and kind of giving you context.

73:15And it's all sort of integrated into this

73:18set up.

73:18So hopefully that's a decent narrative of

73:21how this is all working.

73:23So it tells me exactly what I need to focus

73:27on today.

73:28Next, these are three of the metrics I'm

73:31following.

73:32Each card shows me the latest value of the

73:35metric, quick visual of the trend and A.I.

73:39generated insights.

73:40Like this one.

73:42The week to date appliance sales has seen

73:45an unusual increase and it's now above the

73:47normal range.

73:48Well, that is unusual.

73:51I should know what's going on.

73:54And when I click, Tableau Pulse helps me

73:58understand this specific metric.

74:00There you go.

74:02Metadata from Tableau is right at the top.

74:05So I know I can trust this.

74:08And this is a I mean, that's an assumption

74:11just because the metadata is there doesn't

74:13mean you can trust it.

74:13And I think what she's saying is, look,

74:16when you click on the metadata, you can

74:18find out who the author is, what data

74:19source it is and how the definitions work.

74:22In reality, businesses are not keeping all

74:26of this up to date.

74:27So as long as there is a pretty thorough

74:30process that actually pulls all this

74:32information, then that's going to work.

74:33There is something called the metadata API,

74:36which allows you to surface things like

74:38this.

74:38But if you look at capabilities like data

74:41guide, if you look at capabilities of the

74:43data management add on, that's that's all

74:46created work for someone.

74:46And that work is not necessarily being done

74:49because people don't have the time.

74:50So what also has to happen, in my opinion,

74:54is a concerted effort by Tableau to help

74:57populate metadata for you in a meaningful

74:59way and assign authors, assign ownership in

75:03a way that's smart, but also kind of

75:04reflects the business.

75:05Because if I'm if I'm looking at this and I

75:07'm the business user and it goes and says,

75:09hey, Tim, if I'm in an organization of

75:1110000 people, I might not know who Libby H

75:14ickey is.

75:15Right. I might not know.

75:16So just because you've got that information

75:19, that doesn't mean people in a 10000 person

75:23org knows exactly what this means.

75:25What it should say is maybe their role and

75:27their responsibility with this data set.

75:29That might be something a little bit better

75:31to get.

75:31But then again, there you're feeding into

75:34what's called organizational metadata that

75:36never lives in Tableau, that lives in H.R.

75:38systems and a bunch of other things.

75:40And so you start to see why this gets

75:43really, really difficult for you to really

75:45have trust in this information, for this to

75:48be the point of trust.

75:49You have to know a lot more about the

75:51organization and that data just doesn't

75:52live in your data source.

75:53It doesn't live in metadata.

75:55It lives in systems that, to be honest,

75:58Tableau has no right to be in or pulling

76:01data from for lots of other really good

76:04reasons.

76:04Anyway, let's carry on.

76:06And if I want to learn more, I have a

76:09couple options.

76:10I can take guidance from Tableau GPT or I

76:13can ask my own questions.

76:15Let me show you.

76:17These are questions Tableau GPT has already

76:21suggested.

76:22So I choose one.

76:23What is driving this change?

76:26And Tableau Pulse answers me in plain

76:30language and with a vis.

76:32This new insight is telling me Air Fryers

76:35is driving the change.

76:37Well, they're all the rage these days.

76:41You may recognize some of this A.I. from

76:45other Tableau features, but Tableau Pulse

76:48is taking it to new heights.

76:50It is comprehensive and personalized with

76:54insights like drivers, outliers, trends and

76:59much more.

76:59So what what you're saying is that you've

77:02seen our state and you've seen explained

77:05data.

77:05You've seen all these other capabilities.

77:08Yes, we're leveraging some of that here.

77:10But actually, this is a lot more because of

77:14the integration with Tableau GPT and also

77:16sort of a reimagining of what's going on.

77:19Now, you could argue, well, why didn't I do

77:22this before I and Chanté Petite took off?

77:24And I think actually there's a very simple

77:28answer to that. And Chanté Petite did

77:30something to the whole world.

77:31It made A.I. and machine learning a

77:34household term.

77:35And what that meant is that you went from

77:38users being very skeptical and nervous and

77:42anxious about a capability to Chanté Pet

77:44ite showing how useful it could be for

77:47people to suddenly.

77:49Now, there's huge acceptance of this

77:52technology, but no one deployed it in any

77:54sort of meaningful way to capitalize on

77:56that.

77:56So now. Now the floodgates have been opened

77:59and now people suddenly trust this

78:01technology to do stuff.

78:02Everyone's clamoring to kind of open the

78:05floodgates of their tools.

78:06Google have kind of been caught in this

78:08position as a as a technology company.

78:10Google's A.I. systems have been well

78:12advanced of what open A.I. has done with

78:14ChatGPT.

78:14The difference is, is ChatGPT shipped

78:17something that worked first.

78:19Google were too cautious, too nervous, and

78:22they held off shipping A.I. capabilities.

78:24They've demoed them at Google I/O many,

78:26many years before.

78:27But, yeah, Microsoft and ChatGPT pushed

78:31something out first and suddenly that broke

78:33people's barriers down.

78:34And now they're behind because they didn't

78:37take that risk.

78:37They didn't take that sort of leap of faith

78:40in helping sort of users trust the product

78:42by showing them a compelling use case.

78:44So that's that's that's I think what's sort

78:46of happening here as well a little bit.

78:48You know, every single company has recalibr

78:51ated what their perceived tolerance for A.I.

78:54is in their tool,

78:55because there's now much more mass

78:57acceptance and mass adoption of A.I.

78:59and machine learning, generally speaking,

79:02through large language models.

79:03Next, I can ask a question my own way.

79:07What else should I know about air fryers?

79:12And in response, Tableau Pulse is pointing

79:16me to another metric that matters.

79:19Inventory fill rate.

79:21It's getting too low.

79:23Now, this is a problem I need to be sure

79:26headquarters is solving.

79:28So I share it with my colleague at

79:30headquarters.

79:31Let's call him Pedro.

79:36Now he and I can work on this problem

79:39directly.

79:40Hold on.

79:42Are you seeing this?

79:44So this is the other thing.

79:46A lot of this interface is just completely

79:49new.

79:49There's nowhere in the Tableau product

79:52where there's any interface like this

79:54whatsoever.

79:55So this is definitely to me a completely

79:58separate experience, completely new

80:01experience.

80:01And it could speak to the fact that, you

80:04know, the thing we've used to build charts,

80:07you know,

80:07the vis QL, as it were, that takes a query

80:11and sends it to rendered visualization.

80:13That stuff is too slow for this kind of

80:16really quick responsive Web design.

80:18React is typically what's used for this

80:20kind of stuff.

80:20And so it's needed a whole rethink of the

80:22Tableau interface.

80:23And maybe this is what's going on with the

80:26native product Tableau desktop.

80:27And it's probably maybe what's causing this

80:30sort of backlog of features that aren't

80:31being addressed,

80:31because the rewrite to the Web and all this

80:34technology to sort of make Tableau Web

80:37friendly,

80:37generally speaking, and cloud friendly, is

80:40just taking so much time.

80:41And in some places, it's better to start

80:43from scratch.

80:43And in other places, you have to lift up

80:46the experience to the standard that's

80:48required.

80:49I just got a pulse on my entire business,

80:53saw the one thing that was out of whack,

80:56and acted on it in minutes.

81:00Yeah, she's right.

81:02Penny didn't go to met people for a while.

81:04Actually, yeah, that is what you did.

81:07We can't forget about Pete.

81:08What I think people just, maybe what the

81:11silence shared is that, yeah, and you did

81:13that without a dashboard.

81:14I think that's what people actually, that's

81:17what was going on in people's head.

81:19Wow, you did that without dragging or

81:23dropping anything.

81:24It was just a search box and, whoa, okay,

81:29like, wow.

81:29And I think that's literally what was going

81:31on in the crowd.

81:31And once people sort of caught up with that

81:33thought process, they started sort of

81:36applauding.

81:36But nonetheless, yeah.

81:38He'll be notified in Slack with the same

81:43information.

81:44And the best part is he can do the exact

81:47same things.

81:48Drill down on insights, ask questions, get

81:52those aha moments that close the loop from

81:56insight to action.

81:58So how do I build these metrics?

82:01Three easy steps.

82:02Okay.

82:03One, select a data source.

82:06Right.

82:08First thing, and again, I can do this with

82:11hindsight.

82:11These to me are all published data sources,

82:13100%.

82:14All published data sources confirmed.

82:19The other thing is if these are all

82:21published data sources, then they also

82:23probably include things like the data model

82:25.

82:25So the data model capabilities, you've now

82:28been able to publish a data model.

82:30But if you're doing things like blending,

82:32that's not going to work in this space.

82:34If you've got data sources that you're

82:37trying to put up here and do all that stuff

82:40, that's going to be great.

82:41But fundamentally speaking, there's no viz

82:44in between these data sources and the

82:46metrics.

82:46That was actually the big hindrance of

82:48metrics in the past.

82:49You're going straight from data source to a

82:52metric.

82:52So again, another sort of subtle message.

82:56Hey, you don't need to build a dashboard.

82:58You don't need to build a dashboard.

82:59And not only that, Tableau Pulse over here

83:02on the right hand side is telling you the

83:05potential metrics people might be

83:06interested in.

83:07It's doing some sort of analysis to say,

83:09hey, I've identified 11 or 5, 14 or 4

83:13things that might be interesting in this

83:15data set.

83:15And you can kind of use that as a way of,

83:19you know, intent, sort of signaling intent

83:21and saying, okay, yeah, let's go to the one

83:23with the most.

83:23Let's take a look at that and see what it

83:25says.

83:25Let's see what she does.

83:26Leverage all that hard work you've already

83:30done in Tableau.

83:31Two, define the metric.

83:33Select--

83:34So if we pause it here, right at the top,

83:37it says suggest the metrics.

83:38So I'm assuming when you click that, it

83:40might pre-fill some of these fields.

83:42It might name them.

83:43It might give you some context.

83:45It can only really be doing that one of two

83:47ways.

83:47Analyzing metadata and coming up with the

83:49metrics or, more importantly, it's looking

83:52throughout the business and seeing what

83:54data sources people are using.

83:55And it's coming up with metrics based on

83:58usage.

83:58There's some sort of-- some information

84:00driving that suggestion.

84:01And then you've got the name, the

84:03description, everything else.

84:04That's all perfectly fine.

84:06But nonetheless, you can kind of see that

84:09this is not a desktop altering experience.

84:11To my point earlier, this is not for

84:14desktop authors.

84:14This is for authors to build experiences

84:18for everyone else.

84:19And the key thing, the everyday analyst,

84:22the analyst who will never open desktop,

84:25the analyst who will never open prep, to

84:28come in, choose a data source, and without

84:30seeing any drag and drop interface, select

84:33what they want, find and search what they

84:36need.

84:36And instead of going to Excel to do the

84:38same thing, they can just do it here.

84:39Define it, set it up, and enable it here.

84:42Measures and dimensions and three, share it

84:49.

84:49So there's something briefly there.

84:52Oh, I can't move the cursor fast enough.

84:55The time dimension is really hard because I

84:58think what you have to bear in mind with

85:00time dimension is some metrics require some

85:02sort of capability of snapshotting because

85:05as time moves, the metric changes and the

85:08metric evolves.

85:09And so to be able to do sort of a

85:11comparison to the previous month, you might

85:14actually need to take a snapshot at a point

85:16in time, store that number, and then show

85:18people what that number was like.

85:19And so I don't see that here.

85:21I simply see something saying, what's the

85:24granularity monthly?

85:25What's the period, whatever.

85:27But I think in the future, what you might

85:30have to do is take that into account when

85:33building the data model or fundamentally,

85:36this will have to have a service to take

85:38snapshots.

85:39And again, Tableau already does do this.

85:41There is a service that looks out for

85:44pieces of information and essentially it's

85:46taking snapshots every single week on

85:48certain data sources, spotting what's

85:50unusual and giving you the feedback.

85:51So it's not net new capability, but it's an

85:55important thing to bear in mind.

85:57And three, share it.

86:01And you'll probably want to make many more

86:04metrics on that same data source.

86:06Well, Tableau can do it for you.

86:09With one click, Tableau GPT will scour your

86:13data source and pull out the relevant

86:16metrics.

86:17From there, you control who sees what.

86:22Now, metrics and insights go directly to

86:28you and your colleagues in their email,

86:34phones, dashboards.

86:37I feel like that was the bit that people

86:40didn't clog.

86:40Because again, we were still like going,

86:42hey, how does this work?

86:43And people were generally just processing

86:47this. So when she says dashboards and you

86:48can see it there, oh, it's in a dashboard.

86:48And the funny thing is like, look how stark

86:51it looks compared to a dashboard.

86:52I mean, who wants to build what's at the

86:55bottom here when you can build what's at

86:57the top?

86:57Honestly. And the thing is, you have to

87:00drag them in.

87:00So I'm assuming this will be a dashboard

87:02object called metric.

87:03You'll drag it in and it will just render

87:06this.

87:06But the problem is, is that you look at

87:08this and then you'll go, but I can't build

87:11anything like that.

87:12So what am I supposed to do here?

87:15So, yeah, that's a pretty interesting sort

87:18of juxtapose there between those two.

87:20Slack, wherever they work today.

87:25Yeah.

87:27You've just seen a brand new analytics

87:31experience powered by generative AI.

87:34Everyone is supercharged to make better,

87:38faster decisions in ways they never could

87:41before.

87:41September, October.

87:42And DataFam, this is a whole new way for

87:46you to be a superhero.

87:49This is Tableau Pulse.

87:52And she said it again there.

87:54This is the whole new way for you to be a

87:57superhero.

87:58Not dashboarding, not building anything.

88:00This is different.

88:02And that's why I think this is going to be

88:04a different experience.

88:05Like when you watch this keynote over and

88:08over and over again,

88:09you can't help but notice the different

88:13callouts to the change that's coming at

88:16Tableau in terms of focus.

88:17And they're literally spelling it out.

88:19I can't stress that enough.

88:22Powered by Tableau GPT.

88:25Thank you.

88:32Floor is yours, Francois.

88:35Okay, I'm going to pause there and make

88:37sure my recording is still working.

88:37So one second. Let's figure this out.

88:40Okay. We're good.

88:43It's a 40 gig recording so far.

88:47I'm so glad I stopped it because 100%, we

88:51've still got an hour to go on this.

88:53And yeah, I needed to move it to another

88:57drive.

88:57So I'm glad I did that.

88:59Let's carry on.

89:00All right. What do you guys think of Pulse?

89:02Pretty cool?

89:03100%.

89:04I mean, Pulse is going to be the heartbeat

89:08of business and will really enable everyone

89:11to work with data.

89:12But you know, people work with data

89:14everywhere.

89:15They're in applications.

89:17We need to deliver data within context.

89:20And this is why our second chapter of the

89:23keynote is about bringing data everywhere

89:25and making every application an analytical

89:28application.

89:29So Pedro, come on up.

89:35Thank you, Francois.

89:36All right. So you all just saw how we are

89:39reimagining the Tableau experience for

89:42consuming insights.

89:43And hopefully you're super excited about

89:45that.

89:45But now let's talk about how we are reimag

89:48ining it for building, for developing.

89:51We love our Tableau data devs.

89:53We love our Tableau developers.

89:54And you've all done so much with embedded

89:59analytics.

90:00And we responded.

90:02So we introduced, for example, a brand new

90:06embedding API that makes it much, much

90:09easier for you to bring embedded analytics,

90:12interactive visualizations, or the full

90:15exploration experience into the

90:17applications that you build.

90:19We introduced connected apps to set up

90:22secure and seamless authentication between

90:25Tableau and the applications that you're

90:27building.

90:27But we're not done.

90:30We're rolling out the new Tableau embedding

90:33playground, which is in developer preview

90:36right now.

90:36It is an experience built just for you, for

90:40our developers, to more easily and quickly

90:43build embedded experiences.

90:45So this is not technically a new thing.

90:48I'll put a link to the original Tableau

90:50embedding playground built by a former

90:52colleague of mine, André de Vries.

90:55He built this to help enable people to

90:58learn how to use embedding.

90:59I think what happened, though, is he's

91:02moved on from the information lab now and

91:04he's not updating that playground.

91:05And so because lots of people have been

91:08using it, Tableau have had to kind of build

91:11a new version with a new version of the API

91:14so that it's more relevant for everyday

91:16people.

91:16So it's really good to see this, actually,

91:18because it also shows you how easy it is to

91:21use embedding.

91:21And it maybe brings new developers into the

91:24community a lot faster because it shows

91:26them how to go and do things.

91:27And by immediately showing people how

91:30things work, they get a much better hint of

91:32how things are supposed to work and how

91:33they're supposed to set these things up.

91:34So it's actually quite a useful tool

91:36because documentation is one way of doing

91:38things.

91:38But if you can show people the thing they

91:41're trying to do working right out of the

91:44gate, that is just a much more sort of

91:46easier to understand situation.

91:48And so actually, a lot of the embedding

91:50capabilities have been building up momentum

91:52.

91:52And ephemeral users is actually quite an

91:54interesting one.

91:55Ephemeral users are basically users that

91:58you can create when they need to be created

92:01.

92:01So let's say you have 200 seats, and ep

92:06hemeral user is simply a user that you don't

92:10need to specify who they are upfront.

92:12But when they turn up, you can create that

92:14user, set them up, they come right through

92:16the system.

92:16And when they're done, that's it, you get

92:19rid of that user.

92:19And that's it, they exist once and they use

92:22the instance of whatever they're trying to

92:25use.

92:25And when they're done, they're done.

92:27This means as an application builder, what

92:30you have to worry about is about provision

92:32ing that user with access rather than trying

92:34to license everyone upfront and do

92:36everything upfront, which is a whole sort

92:38of pain in the neck.

92:39I'm sure I'm not describing this well

92:41enough, but nonetheless, it's something

92:43that's very welcome.

92:44Anyway, let's keep pushing.

92:46We're introducing ephemeral users, which

92:49makes it very easy to centralize user

92:51identity and access management.

92:53You don't have to worry about maintaining

92:57or setting up users in two different places

93:00.

93:00And with the introduction of usage-based

93:03licensing, we're making this feature even

93:06more powerful.

93:07Usage-based licensing is a new payment

93:09model.

93:09It's paid for what you use.

93:12Think about how easy it'll be to scale up

93:16or scale down as you launch new

93:18applications.

93:19And now...

93:21So user-based licensing is pretty important

93:24because up until then, you had to know how

93:26many people were going to use your platform

93:27, which meant you had to kind of price that

93:29in right from the gate, even if those users

93:32didn't exist or even if that business use

93:34case hadn't been proved out.

93:35So with usage-based licensing, you can

93:37basically say, hey, we want to use Tableau.

93:39We don't know how much we're going to use,

93:41but we're going to buy a bunch of credits.

93:42And when those credits run out, we'll buy

93:45more.

93:45But those credits allow a certain amount of

93:48interaction.

93:48So those interactions are essentially

93:50visualization loads.

93:51Those are counted as interactions by Table

93:54au.

93:54And so you can buy a bunch of tokens that

93:57count with those interactions.

93:58When you hit your million, you've hit your

94:01million, you can buy more and you can scale

94:03up, add more credits much, much easier than

94:06you can predict the usage patterns of an

94:08individual and or a group of individuals in

94:11a particular application.

94:14We're going to introduce the next big leap

94:18in capabilities for embedded analytics.

94:21Introducing Tableau ViscQL Data Service.

94:25It's pretty cool.

94:29So I've gone through a whole wave of

94:32feelings with this.

94:33To me, this stands as the most important

94:36part of the whole entire keynote.

94:38And as I've been talking to a few people

94:41behind the scenes, it's it's I can only

94:43emphasize that more.

94:44This is by far the most important thing in

94:47this keynote one, because it's actually the

94:50technology enabling Tableau Pulse.

94:52Tableau Pulse itself is using this API to

94:56build all those metrics infused together

94:59with capabilities like Tableau GPT, the AI

95:03capabilities to understand intent.

95:05But also fundamentally, I think this is

95:08also a signaling of Tableau really

95:10understanding that it's a platform and

95:13really showcasing that it's it's got the

95:15breadth of a platform because most

95:16platforms want to get to a certain scale,

95:18allow developers developers to build on

95:21specific things.

95:21If you take Apple and the iPhone, they

95:23build a bunch of features into the iPhone,

95:25it could be a camera, it could be a face

95:27sensor, it could be face tracking, it could

95:29be face scanning.

95:30And once they've got all those features,

95:33they enable developers through APIs to take

95:35advantage of those features, whether it's

95:37the three cameras, the sensors, all of

95:39those things, AR capability, there's so

95:41many things in your phone, so many

95:42capabilities in your operating system that

95:45you think belong in the apps, actually,

95:47that Apple enabling those technologies to

95:49make those apps allow them to do those

95:51things.

95:51And so here, this is the same thing. So

95:52Tableau taking Viscuel, the process that

95:54goes and gets data, and instead of

95:56rendering the visualization, actually, it's

95:58stopping short of doing that and just

96:01serving up the data to you.

96:02So you can build whatever you want on top

96:05of that, and Tableau themselves have built

96:07Tableau Pulse, but you could build anything

96:09like Tableau Pulse on top of that.

96:11First, you just need to have a good

96:13understanding of what you need. That's

96:15often the hardest thing. Bring it all in,

96:16process it in whatever tool you want to use

96:18and go off and build your experiences.

96:20I think this is a huge, huge thing. Tableau

96:23will use the term decoupling. To me, I

96:26think this is potentially the first signal

96:28that Tableau is really, really taking its

96:31platform sort of capabilities seriously and

96:33scaling them out.

96:35Viscuel Data Service is us popping the hood

96:40on Tableau and giving you access to the

96:44magic, the power of the core Viscuel

96:48Analytics engine. Now, you may be familiar

96:51with this concept. You may have heard of it

96:52as headless BI or composable analytics.

96:53Essentially, all it means is we are decou

96:57pling the Tableau front end business from

96:59the back end, the magic, the power of Table

97:02au. And imagine how powerful that is.

97:05When he just said that, I was like, oh,

97:09here we go again. Here's another sort of

97:12assault on the authors and stuff like that.

97:14Here we go.

97:14For all these years, you've known the magic

97:17of Tableau, but to access that magic, you

97:19've had to have these people who build dash

97:22boards. Guess what? We're decoupling those

97:24things.

97:24Don't worry about the dashboard building

97:27stuff. Here's the data. Here's the magic,

97:29as they call it. Now go build things on top

97:32of that.

97:32Again, yet another message here that Table

97:36au is really not catering to the author in

97:39this keynote. They're asking the authors to

97:42change their efforts.

97:43Instead of building dashboards, come and be

97:46a developer. Instead of building dashboards

97:48, come and build metrics. The message can be

97:50clearer.

97:50Giving you a very simple programming

97:53interface that puts that power in your

97:55hands so you can do whatever you want with

97:57it.

97:57You can create a brand new way to interact

98:00with Tableau that's not about drag and drop

98:02or business. Maybe you have some automated

98:05workflow.

98:05You want to infuse Tableau Viscuel Engine

98:07into those workflows. Maybe you want to

98:09build a chatbot that interacts with Tableau

98:12.

98:12You can do all these things now with Visc

98:15uel Data Service. Truly, the only limitation

98:19is your creativity.

98:20And this is a massive step forward in

98:23helping you put the power of Tableau

98:25anywhere you want people to experience data

98:28.

98:28So who wants to see this in action?

98:34Please join me in welcoming to the stage

98:38software engineer, Kate Gaiman.

98:40So I'll preface this by saying, thank you,

98:46Kate.

98:46I will go through a couple of features that

98:48have existed in embedding for a while.

98:49So what I'll do is I'll try and call out

98:51the things that are new.

98:52Embedding demos can have this sort of

98:55challenge of being a little bit sort of

98:57challenging to see because the only thing

98:59they can show you is a coded experience to

99:01try and simulate what it might be like.

99:03But ultimately, Tableau is nowhere near as

99:05creative as the people who build these

99:07tools.

99:07So let's listen to what Kate has to say.

99:10I have traditionally made this hard. And

99:13today I'm going to show you how.

99:21As a developer at Cumulus Bank, I love to

99:25craft beautiful, creative solutions.

99:29Check out this personalized banking app I

99:33built for our customers with Tableau embed

99:36ding.

99:36So the point of this is to highlight to

99:41people that this is a kind of thing you can

99:45build as a bank with an embedded

99:48visualization with data coming from Tableau

99:50.

99:50This chart, this account overview is a

99:52Tableau chart.

99:53This chart on the bottom says spend, spend

99:55versus budget, spend versus current year is

99:57coming from Tableau.

99:58The stuff on the left is not coming from

100:01Tableau.

100:01That's coming from the application.

100:03The stuff across the top, the stuff on the

100:05left, that is all the application.

100:06The only two things coming from Tableau,

100:08the two charts in the top and in the middle

100:10, those are Tableau charts.

100:11And that's sort of the whole point of embed

100:13ding.

100:13You look at something like this and you don

100:15't think there's any Tableau in it, but

100:17actually it's just a highly well

100:19orchestrated embedded application.

100:21Specifically, we're looking at Francois Sp

100:25ending, who conveniently just happens to be

100:28a loyal Cumulus customer.

100:30There is so much interactivity built into

100:34this viz.

100:35I can filter by money in, money out,

100:39overall cash flow or change the relative

100:43date range.

100:44I can pull in key metrics from Tableau.

100:48Oh, no, net worth is down?

100:52Why is that?

100:53Francois has spent so much money in May

100:58already.

100:59In fact, Francois, did I see you with the

101:02slot machine last night?

101:04It's all starting to make sense.

101:07Lucky for him, if he's really concerned, he

101:11can take action on his data right in his

101:14workflow with this custom context menu that

101:17will send a report of his spending directly

101:20to his financial advisor.

101:22So that is a really subtle thing.

101:26I didn't even notice that until now.

101:28I've watched this so many times.

101:30I've only just clocked that that action was

101:34something I've not seen done anywhere else.

101:36If I just can I go back a few frames?

101:38Yes, I can.

101:39I can use the.

101:43Think I sit here and just keep going back.

101:46We're doing it in reverse.

101:49Come on, keep going, keep going, keep going

101:51, keep going, keep going.

101:52Here we go.

101:53Send to my financial advisor actions.

101:58Either that's new or these are actions that

102:02have existed as part of the dashboard

102:06experience, but that I was under the

102:08impression those were Salesforce actions.

102:10I didn't believe them to be actions here.

102:12So it's super cool that you can you can

102:15build in actions into that contextual menu.

102:18I did not know that was a thing.

102:20Maybe it's a new feature coming.

102:21Completely missed it.

102:23Super interesting.

102:24Report of his spending directly to his

102:31financial advisor.

102:34Today with Tableau embedding, you can even

102:39embed web authoring.

102:42Clicking explore data will empower Francois

102:47with our first class drag drop self-serve

102:50analytics experience.

102:52How cool is that?

102:55Very good.

102:57As a developer, it was so much fun to

103:01integrate all these bells and whistles.

103:04But do you guys know what's even more fun

103:07than building things?

103:09Building things fast because forget weeks

103:14of reading documentation with Tableau embed

103:20ding Playground.

103:21I built this web portal in just minutes.

103:26Let's slide on into Playground and check it

103:31out.

103:31Good little transition.

103:32All I need to do is paste a link to my vis.

103:34More like a dance floor made of that.

103:37And Playground chalks out a prototype for

103:41me.

103:41Code, written, and all.

103:44Using this text editor at the bottom, I can

103:47play around with the API and get instant

103:50results.

103:51It's basically like playing on the monkey

103:54bars.

103:54But no broken bones this time.

103:57Now, over on the left is where the magic

104:00really happens.

104:01The embedding API offers a wide range of

104:04functionality from customizing the display

104:07to adding all sorts of user interactions.

104:10Like this context menu we saw earlier.

104:14I can browse through and when I --

104:16You know what would be really cool in here?

104:19Tableau GPT.

104:21If you could just have a text box saying I

104:24have this vis, here are the details.

104:26Show me how to embed it in this particular

104:31context.

104:31And I'd also like you to generate the HTML

104:34for the buttons and filters to filter this

104:38particular data set based on how this works

104:41.

104:41I mean, chat GPT could do that if it knew

104:43enough about your data source and you

104:45prompted it correctly.

104:46Tableau GPT should surely be a part of this

104:49and be doing that for you.

104:51Never mind showing you how to do it.

104:52What if it just did it for you with a

104:55search box?

104:55That would be incredible.

104:57See what I want to add?

104:59Say a relative date filter.

105:01It's as easy as dragging and dropping and

105:07the code just appears.

105:10That is the next best thing.

105:12And you probably have more confidence that

105:15this is correct.

105:15Re-run Playground and you can see the

105:18dashboard is now filtered to only show data

105:22from last quarter.

105:23How cool is that?

105:25My question is why would you beep back when

105:30you could just drag drop?

105:33Once I'm satisfied, it's as easy as copy

105:37the Playground code, paste back in my web

105:40app and deploy.

105:41And that's how, with just a few clicks, I

105:46seamlessly added analytics for our banking

105:49customers right in their workflow.

105:52With Tableau embedding, we can bring robust

105:56interactive components into any web

105:58application or web portal quickly and

106:01easily.

106:02But when we talk about analytics everywhere

106:06, we're not just talking about visualization

106:10.

106:10Because it's fundamentally about the...

106:13Oh, God, I'm just going to keep stopping

106:17every time they say it and be like, she

106:20said it again.

106:20We're not just talking about data.

106:22When we talk about analytics, we're not

106:24talking about what you think.

106:25We're talking about this.

106:27Couldn't be stressed more.

106:29That powers the vis.

106:31So there's one last tool that, especially

106:35as a developer, I am most excited to show

106:39you guys.

106:40With VisQL data service, we are opening up

106:44direct access to our data modeling

106:48capabilities.

106:49So data modeling capabilities, that is not

106:55just the data sources.

106:57I think that's another subtle thing.

106:59You can essentially publish up a data

107:02source with a data model and then query

107:04that data source using visQL data service

107:06and have the model an intrinsic part of

107:09that query, which is, again, super powerful

107:11.

107:11Because with most other applications, what

107:14you'd have to do is you'd have to stand up

107:17a flat table of flat data structure behind

107:19that for something to query.

107:21But here instead, you can just query the

107:24data model.

107:24And if your data model has been built by

107:26someone who understands the concept of data

107:29modeling, then, yeah, you don't have to

107:32worry about all the intricacies of

107:34relationships and everything, because that

107:36's all done for you.

107:36And you can leverage that work all the way

107:39through your stack in your applications, in

107:41your dashboards, in your metrics.

107:42All of that is going to take advantage of

107:45that single sort of version of the truth.

107:47Your data model can power automation,

107:51unlocking insights and actions anywhere.

107:55To see how this works, let's check out

107:59Francois' Spending by Category.

108:02Through our newest API, I can now access

108:06these numbers outside of the context of

108:10this vis or even in a completely separate

108:13application.

108:14Over in my developer tool, I've written

108:18this request to visQL data service to get

108:21customers spending by category.

108:24I just added the dimension and the measure

108:28that I wanted, and clicking send returns me

108:32the exact same data we just saw in our

108:35application.

108:36So this interface, I don't think will be an

108:38interface you'll ever really actually see.

108:40It's simply just a bit like graphic URL

108:43where you can help build queries and

108:46visualize an API to understand better how

108:49it works.

108:49But in reality, the query you'd build would

108:53be in some sort of script in a back-end

108:56system, potentially paired with something

108:59like the REST API, calling Tableau, getting

109:01the data.

109:01And that data would come back in a JSON

109:03array in some sort of web format that then

109:06you can quickly deploy and use in your

109:08application, essentially.

109:09So this demo is a conceptual demo as much

109:12as it is to just show you what happens when

109:14you send a query.

109:15This is what you get back and this is what

109:18it looks like.

109:18But in actual fact, it's just a

109:20visualization of the feature.

109:21But visQL data service is flexible, so I'm

109:24not limited by what the vis was showing.

109:26In my second request, I've added another

109:29dimension, subcategory, to my query.

109:32In the response, I can see a more refined

109:36breakdown of Francois' spending.

109:38Whoa, Francois has been spending a ton of

109:44money at restaurants.

109:47Thanks again for dinner.

109:50Let's use this new intel to programmatic...

109:54What I would have loved is for this demo to

109:59showcase something using the data model,

110:02because I think that was called out.

110:03And I think that that would have been

110:05something super powerful to show, that to

110:08get this value correct, you'd have to rely

110:10on the capabilities of relationships and

110:13the data modeling setup in order for this

110:15to sort of work.

110:16It's a little bit tedious to show on stage,

110:19to be fair, and by the time you've done

110:20that and the pennies dropped, that's like a

110:2215-minute demo.

110:23They don't have that kind of time.

110:25But that, I think, is sort of the kind of

110:27demos that people will sort of...

110:29Those are the demos where people will

110:31realize how powerful this is.

110:32We sent in an alert back in our banking app

110:36that...

110:37Congratulations!

110:39You've qualified for a 10% cashback bonus

110:43at all future restaurant spending.

110:46Way to go, Francois.

110:49This is just one single example of how Cum

110:53ulus Bank could leverage VisQL data service.

110:57What else could Headless BI do?

111:00Maybe fraud alerts?

111:02Yeah, you bet.

111:04Dynamic budgeting?

111:06Why not?

111:08Automatic investment?

111:10Let's build it.

111:12That's the thing. These examples are all...

111:14These must be use cases.

111:16Someone's come to them and said, "Hey, I'd

111:18love to do this kind of thing.

111:19Querying the data sources in the old way is

111:23too slow.

111:23If you could stand up some other way for me

111:26to query this data,

111:26I could more easily identify these kinds of

111:29issues and challenges and use Tableau as

111:34sort of a backend system

111:34for storing all this information where dash

111:37boards, metrics and also applications can do

111:41their work, if that makes sense.

111:41In many ways, you could almost look at this

111:43service as being the first time an

111:46application can run on the Tableau platform

111:48.

111:48In actual fact, it's running on top of the

111:52Tableau platform,

111:53but you could say it's running on the

111:55platform because it's using data from the

111:57Tableau setup.

111:58Or whatever other creative solutions you

112:03all come up with next.

112:05That's all from me. Back to you, Francois.

112:08[Applause]

112:14Great job, Kate, and I'm glad you didn't

112:17dig into my shopping purposes... purchases.

112:20All right, so we've seen how we're bringing

112:23data to everyone and how we're putting it

112:25everywhere in every application.

112:26But really, Tableau needs data and we need

112:29to trust the data.

112:30And Tableau continues to be for all data.

112:35That is not changing.

112:37No matter where your data is stored,

112:39whether it's in the cloud or on-premises,

112:41in Snowflake, SAP, SQL Server,

112:44if you have data, we want to connect to it.

112:47This is why in the past year, we've

112:49expanded our data connectivity options.

112:51We added more connectors through the Table

112:54au Exchange to help you connect to the data

112:57that matters to you.

112:58And we're just about to release a new

113:01connector to Amazon S3.

113:03Oh, yeah, so if you've got data in S3,

113:06whether it's in Parquet files, CSV, or

113:10Excel, it's just one click away with Table

113:14au.

113:14So we're continuing to add more connectors,

113:19but more and more data leads to more and

113:23more data chaos.

113:25So we're well and truly into the third

113:28section of this keynote.

113:29And now we're about to talk about Sales

113:31force.

113:31So at no point in this keynote, in the

113:34three key areas,

113:35is anything for a dashboarding author,

113:38anything for the typical analyst of the

113:40past, be in Showcase.

113:41That you get devs on stage, that's the

113:44stuff they call true to the core, right?

113:45That's the stuff that, you know, the core

113:47user up until now.

113:48But in terms of where Tableau is heading,

113:50this is it.

113:51You've got Tableau Pulse, Tableau GPT,

113:54embedding, developers built on top of the

113:57platform.

113:57And then you've got Salesforce Data Cloud,

114:00essentially the capability to ingest data

114:04all into Salesforce

114:04and then have a data model built for you in

114:06the background.

114:07This is essentially what we're about to see

114:10a demo of.

114:10I'll let sort of Francois speak a bit more

114:13about it.

114:13We'll skip the F1 bit and then we'll keep

114:16going.

114:16Data is hard.

114:18But we want to make data easy.

114:20We want to make data available and

114:23accessible and trusted for everyone.

114:25This is why late last year we launched the

114:29Salesforce Data Cloud.

114:30The Data Cloud enables you to do amazing

114:34things.

114:34It lets you combine data together from

114:37hundreds of sources.

114:38It then lets you harmonize that data into a

114:42single source of truth.

114:43So you can model it all in one place.

114:46And you can then use that data and enrich

114:50it with rich algorithms with AI built in.

114:53And you can use the Data Cloud not just

114:57with Data Cloud.

114:58You don't have to move all your data in

115:00there.

115:00This platform is completely open, which

115:04means that you can zero copy that data into

115:07other data.

115:07Into other databases like Google BigQuery.

115:10Which we're just announcing at this

115:12conference.

115:12So now the data in the Data Cloud which

115:16comes in in both real time and batch

115:18can automatically be available in BigQuery

115:21with no transformation, no duplication.

115:24It's instant.

115:26Now I have to say the audience at this

115:31keynote, I think very few people will get

115:33what's going on here.

115:35And that's because it's just a completely

115:37different world.

115:37This is almost the world of cloud and Dev

115:40Ops basically.

115:40Zero copy, zero ETL.

115:42To be brutally honest with you, that makes

115:45no sense to me.

115:46Zero copy means you're not having to

115:49literally get the files over there.

115:50You're not having to copy it over.

115:51Zero ETL means you're not having to run a

115:55whole bunch of ETL setups to get the value

115:58from your data.

115:58So for that to be available in Google Big

116:03Query.

116:03Either it's already running in Google Big

116:06Query or there is something else I don't

116:09fully understand.

116:10So if you know the answer, drop a comment

116:13below.

116:13Educate me on what zero copy, zero ETL

116:16means in the context of Google BigQuery.

116:18But I simply don't understand.

116:20Vice versa, the Data Cloud data can come

116:23into the Google BigQuery data

116:25can come into the Data Cloud without any

116:28data movement as well.

116:29It's instantaneous.

116:31And it processes massive volumes of data.

116:35But best of all, the Data Cloud is

116:39optimized for Tableau.

116:40So the only way that could work is if Sales

116:44force is just doing the work on the data

116:47itself.

116:47Like zero copy, zero ETL means wherever the

116:51data is, it's already in the right place.

116:52And it can already process that right there

116:54.

116:54But you're not doing ETL.

116:55You're not moving it.

116:56You're not reshaping it, standing it up as

116:58something else.

116:59You're literally processing it from where

117:01it is in the Salesforce Data Cloud, native

117:04ly speaking.

117:05And then this goes back to the point that

117:07Ryan mentioned earlier,

117:08which is that Salesforce Data Cloud is now

117:11powered by Tableau.

117:11And here you can see Hyper.

117:13Hyper is a technology that's really running

117:15all of this stuff.

117:16We've taken the Tableau Hyper engine and we

117:20've put it at the heart of the Data Cloud.

117:22And what that means is that when you

117:24connect to the Data Cloud,

117:25it's like if you were connecting to your

117:27local Hyper extract.

117:28But it's running at cloud scale.

117:31It's fast and it's analytically ready on

117:36day one.

117:37In addition, we've also introduced instant

117:41analytics.

117:41So you can go from data to analysis in

117:45Tableau instantly.

117:47There's no reconnecting to the data.

117:49There's no remodeling the data.

117:50It's automatically available.

117:53And so to show you the power of the Data

117:56Cloud, I'd like to...

117:57So essentially what he's saying is that

117:59once you've done all the work in the Sales

118:00force Data Cloud,

118:00you can just connect to it in desktop, do

118:03whatever you need to do it.

118:03You don't need to do or repeat any of the

118:06work you've already done,

118:07again, in order to start visualizing and

118:09doing stuff with it.

118:10So I need to skip this F1 segment.

118:12Let's go ahead and do that.

118:14Normally I'd never skip F1, but hey, let's

118:18skip ahead.

118:19Here we go.

118:20Then we go to Francois' demo.

118:21He's a McLaren fan.

118:23Fun fact, I actually grew up in McLaren's

118:26hometown of Woking.

118:26And that's literally where I grew up when I

118:29was young.

118:29So McLaren is also my home team, as it were

118:32.

118:32They are my home team, although I support

118:34Mercedes and Lewis Hamilton,

118:36because Lewis drove for McLaren.

118:38That's actually how I started supporting

118:40Lewis.

118:40And when he moved to Mercedes, I kind of

118:42moved with him.

118:43But I've always remained a McLaren fan,

118:45Lando Norris, everyone else.

118:48Yeah, so great that Francois is a McLaren

118:51fan.

118:51Francois, if you're ever in the UK, give me

118:54a shout.

118:54I'll go over to the MTC, and I'll take you

118:57around Woking.

118:57And maybe Lando Norris will be zooming

118:59around in his MP1 or something like that,

119:02and we can bump into him.

119:04...speed for speed, and the combination of

119:07Data Cloud and Tableau

119:08just brings data to life in brand new ways.

119:11So here we're looking at a Tableau

119:14dashboard,

119:15and we can see all of the various touch

119:18points in which Formula 1

119:19engages with their fans, from emails to F1

119:23experiences to online purchases

119:25or drive to survive on Netflix.

119:27It's amazing.

119:28Now, think about it.

119:29This data is coming from hundreds of

119:32sources,

119:32and Formula 1 needs to aggregate it all

119:35together

119:35into a single view of their customers

119:38so they can engage with their fans in new

119:41ways.

119:41Now, this dashboard is powered by the Data

119:44Cloud,

119:44and so let me show you how to put together

119:48the data behind this dashboard.

119:50First, here we're looking at all of the

119:53various data streams

119:54that we're bringing into the Data Cloud.

119:57You can see that we have data coming in

119:59from Salesforce,

120:00from the Marketing Cloud, from Amazon S3.

120:03This is hundreds of data sources coming in.

120:06And you can bring in new data with a single

120:10click.

120:10And there it is.

120:12And now we can connect to our Salesforce

120:15data really easily.

120:16I can connect to my Cloud storage data,

120:18or we can bring in data from any number of

120:23sources

120:23just really, really easily.

120:26All right, now that we have data, we then

120:29have to model that data.

120:31So here we have what we call the harmon

120:34ization layer.

120:35This is where we take all of those various

120:38data sources

120:38and combine them together into a single

120:41model

120:41so we can get a single view of the customer

120:44.

120:44And so you can see how we're creating all

120:47of those mappings

120:47between the source and the target.

120:50Next, you'll see on the right,

120:53our data is coming in from all these places

120:56,

120:56but we then have to unify that data

120:58because Francois that's engaging through

121:01Netflix

121:01or through the website or through the Grand

121:03Prix,

121:04I mean, I'm the same person,

121:05but there are individual different records

121:07in all the sources.

121:09And with the data Cloud, we have the

121:11ability to resolve the unified profile

121:14and turn 63 million individual customers

121:18into actually 23.9 million rows.

121:21That's it.

121:22All right, so now we've modeled all that

121:25data

121:25and you can see how all of the various data

121:28sources

121:28come together in this graph view.

121:31Now, do I have to do anything else to bring

121:33this into Tableau?

121:34Absolutely not.

121:35Here we have instant analytics for Tableau.

121:38I can go ahead and download this as a Table

121:41au data source.

121:42Now, when I saw this demo, I was like,

121:45"Huh, that is pretty powerful."

121:47That's literally what I spend all my time

121:50doing as an analyst,

121:51building dashboards.

121:52All this activity that we've just gone

121:54through

121:54up until now that is done in the Salesforce

121:56,

121:56that is literally the bread and butter of

121:59being a data analyst sometimes.

121:59Just bringing the right data source

122:01together

122:01is sometimes 90% of the challenges.

122:04Once you've done all of that and you've

122:06formatted and scoped the data

122:07and it's all perfect, the rest is easy.

122:09The dashboard just comes naturally.

122:10If Tableau is a good tool, it will let you

122:13do that.

122:13But the hard stuff is always this.

122:15It's really interesting to see that, look,

122:17Salesforce actually thought through this

122:19for its own challenges.

122:20I'm pretty keen to try this out.

122:23I'll bring in Superstore, we'll try mapping

122:26orders and sales and regions

122:27and just see how well that works.

122:29Put the data in Dropbox, we'll put the data

122:32in S3,

122:32we'll bring it all together and see what

122:35happens.

122:35But, yeah, this is actually pretty

122:38compelling.

122:38If it works as advertised, and I have to

122:40say this is fairly new

122:41and I think I heard from Tableau themselves

122:45that they're using this internally to help

122:48them solve problems

122:49inside of the organization.

122:51So if they're using it and it's hopefully

122:54helping them,

122:54I get a sense they'll get a really good

122:56feel

122:56for how it could be made better because

122:59they're actually using

122:59what they're selling, as it were, as a tool

123:01.

123:02So they'll experience it, all the same pain

123:04points,

123:04all the same issues, the same you and me

123:06will experience it.

123:07That's actually quite promising.

123:09I'm always worried when a company doesn't

123:12dogfood its own product

123:12to help run its own business.

123:14Or TDS, or soon you'll be able to open it

123:18up directly in the browser

123:20and start exploring it.

123:21That's pretty good.

123:22And when you open it up, well, you get the

123:25entire data model

123:26pre-created in Tableau.

123:27You don't have to redefine your

123:29relationships,

123:30redefine your joints, it's right there for

123:33you.

123:33I will say that's a monster data model.

123:35[laughs]

123:37Look at how many tables are in there.

123:39It's absolutely nuts.

123:41And yeah, it doesn't seem to be benefiting

123:44from some of the new capability

123:46of shared modeling either.

123:47So if this is, you know, I really hope that

123:52with shared modeling,

123:52you know, something like this kind of works

123:55really well

123:55with several base models and then a bunch

123:57of other sort of,

123:58you know, related things.

123:59Also allowing you to change your

124:01perspective.

124:01So maybe I don't want all of these things.

124:03I'd love to be able just to choose what I

124:05need and bring that in.

124:06Even though the model is much bigger,

124:08I'd love to just prune down to the few

124:11things I need

124:12and have that in the connection rather than

124:14having everything else.

124:15And then I can start analyzing this data.

124:20Now every time I drag and drop, I'm

124:24connecting to the data cloud.

124:26I'm slicing billions of records across all

124:29of those sources

124:30and I'm able to visualize it and use the

124:33full power of Tableau

124:34because remember, we're using hyper in the

124:38backend.

124:38We're using hyper at scale and powering

124:42these experiences.

124:43And that's how simple it is.

124:45And I can go from data to insight and

124:49engage with my data in brand new ways.

124:51And so this is the power of data cloud and

124:55Tableau.

124:56Really good demo.

124:58Really good demo.

124:59And this is available today for all of you

125:02to use.

125:02Whether you're using Salesforce data or not

125:05, it's available for everyone.

125:07And so today you've seen the next evolution

125:11of Tableau,

125:11how we're bringing new...

125:12That was almost... I completely missed that

125:15actually as well.

125:15Every time I watch this, I learn something

125:17new.

125:18It's available for everyone whether you're

125:20using Salesforce or not.

125:21So you can go and use Salesforce data cloud

125:24even if you're not using Salesforce.

125:26I'd love to see how that works in practice.

125:28Of course, it's an opportunity for them to

125:31upsell you to Salesforce.

125:32But nonetheless, yeah, if that's really how

125:35it works, amazing.

125:36I wonder why they didn't call it Tableau

125:39data cloud.

125:39That would have been a much better branding

125:40because to me,

125:42I sort of think Tableau is the data product

125:44within Salesforce.

125:45So why didn't they just call it Tableau

125:47data cloud

125:47and then tell everyone in Salesforce that

125:50this is going to help you clean your data,

125:51prep your data, model your data.

125:53And it would have been a bit more of an

125:56accepted sort of solution

125:56within the Tableau community more broadly.

125:58But anyway, you don't build features for

126:01this one or two percent.

126:02You build it how you're supposed to build

126:04it

126:04and you worry about the problems later.

126:06Experiences for business users, how we're

126:08empowering developers

126:09to build new kinds of applications powered

126:12by Tableau

126:12and how the data cloud is going to help

126:15everyone deliver data trusted

126:17for your organization.

126:19But are we done?

126:21Oh no, we're not.

126:22It is time for the one and only Devs On

126:29Stage.

126:30And so I've actually done videos on

126:34everything in this.

126:36I've gone through pretty much all these

126:38features.

126:38I've summarized them already.

126:40You can check out those in the channel.

126:42I'm not going to go through them now

126:43because I sort of don't think

126:45they're really sort of part of the keynote.

126:47This is almost a separate session.

126:49They are fantastic.

126:50I've broken all of them down and the only

126:53thing I haven't broken down yet

126:53is visQL data service.

126:55I'll do that in a separate video.

126:56But as for everything else you're about to

126:58see,

126:58including Tableau Pulse, Tableau GPT,

127:00I've done videos on all of those in the

127:03same playlist as this video.

127:04So we're going to stop this video here.

127:06I'd love to know what you thought of this

127:08commentary.

127:08It was pretty long, a lot longer than I

127:10thought it was going to be.

127:12But nonetheless, yeah, thanks for watching

127:15and I'll catch you in the next one.

127:16[END]

127:28[ Silence ]