Tableau Conference 2023 Keynote Roundup

0:00Hey, it's Tim here. In today's video, I'm

0:01rounding up the Tableau Conference Keynote

0:04in 2023 in just under 15 minutes.

0:07We've got no time to waste. Let's get stuck

0:09in right now.

0:10The keynote was an hour and a half long

0:11broken down into six key sections. Let's

0:14just go ahead and list them out now.

0:16First, we have the introduction. This

0:17pretty much set the tone for the whole

0:19conference and answered a few questions

0:21about Tableau, Salesforce and the community

0:23in general.

0:24The second section was actually a little

0:26bit of context setting for the whole of the

0:28rest of the keynote. It was a little bit of

0:30context setting talking about the industry

0:31in general.

0:32The third section was about Tableau and AI.

0:35This was the first main section. Tableau

0:36really showcasing what they're going to be

0:38doing with AI going forward.

0:40The second section was really a story about

0:42telling people how Tableau fits into any

0:45part of any analytics stack through embed

0:47ding.

0:48They announced a couple of new features

0:49here. We'll break them down soon. The third

0:51section was about Salesforce Data Cloud.

0:54Tableau termed this as trusted data for all

0:56, but in essence, they showed how Salesforce

0:58and Tableau work together.

1:00Last but not least, we had Devs On Stage.

1:02Devs On Stage is, of course, something that

1:03has always been separate, but in the last

1:05two conferences, it's been tagged onto the

1:07end of the keynote.

1:08We had an absolute blockbuster range of

1:10features showcased here. What I will do

1:12over the course of the next seven days is I

1:14will make a video on each of the features

1:16that were announced there just to keep this

1:18video nice and short for you so you can get

1:20a hint of what happened in the keynote.

1:22Right, let's get into the first section,

1:23the introduction.

1:25Right out of the gate, it's worth

1:26recognizing the new layout for conference.

1:29It was a very different kind of feel. It's

1:30the Salesforce feel that we're familiar

1:32with seeing at things like Dreamforce.

1:34But nonetheless, I think Tableau did a good

1:35job of making it feel like a Tableau

1:37atmosphere. We still had a big screen and a

1:39stage up front, but then you had the middle

1:41space where the presenters were walking

1:43through.

1:44Really interesting look. Let me know what

1:45you think about that in the comments.

1:47After that, we kind of went on to a couple

1:48of other interesting things. First of all,

1:50we obviously had the 20-year celebration of

1:52Tableau.

1:53We also had Ryan Atay being announced as

1:55CEO at the conference. That makes a ton of

1:57sense because that's where people are. This

1:59is where the headlines come from.

2:01If you've seen my recent video, I kind of

2:02highlighted that this would probably happen

2:03.

2:04Really good to see it here and obviously

2:06hope to see the news and outcomes coming

2:08from that.

2:09It was really interesting when Ryan himself

2:10was talking, he actually called out the

2:12issue around customer support that has been

2:14referenced in lots of different places,

2:16including the Gartner Magic Quadrant.

2:18That was super interesting to hear. He

2:20announced something called the Tableau

2:21Success Plan that's going to be coming in

2:22the later months.

2:23This might be a direct response of not just

2:25trying to resolve the problem, but trying

2:27to go ahead and do better than we've ever

2:29done before maybe with the product.

2:31Let's wait and see what that has in store

2:33for customers coming forward.

2:35Tableau then moved on to celebrating the

2:36community. This was really important

2:38because I think it really helped break the

2:40ice of the conference.

2:41This conference to me had a little bit of

2:43attention in the keynote because everyone

2:45is really keen to see what is Tableau going

2:47to do? How is Tableau going to innovate?

2:50In order to answer that question, first I

2:52think Tableau has to put you in the right

2:54frame of mind.

2:55Talking about the community was a really

2:56good way of doing that.

2:58Putting familiar faces, people that you've

2:59hopefully interacted with in the community.

3:02If you haven't, calling out all these

3:03people to show that the Tableau community

3:05is really a community. It's like a family.

3:08By getting the visionaries on stage,

3:09getting the ambassadors called out, user

3:12group leaders, highlighting all the

3:13different ways that Tableau is helping

3:15people develop data skill.

3:17The overall message is that we're on this

3:19journey together and it was really good to

3:21see everyone there in the middle.

3:23This sort of circular setup, I don't think

3:24that was coincidence. It's very deliberate

3:26even though it maybe speaks to a setup that

3:28Salesforce already has.

3:30So it was a really good setup and I think

3:32it maybe brought everyone's barriers down a

3:34little bit and opened people up for the

3:37next section where Francois started teeing

3:38up some of the things we were going to see

3:41in the keynote.

3:42Let's talk about that.

3:43Now to set the context for the rest of the

3:44keynote, there was sort of a small little

3:46bit of almost like glue.

3:48Essentially Tableau needed to set the scene

3:50and they talked about the chasm between

3:52data and people in every organization and

3:55how organizations still feel like not many

3:57decisions are data driven.

3:59And so this was really interesting because

4:01I think it's still true.

4:03Tableau has been on this mission for 20

4:04years and they've acknowledged many times

4:06there's still a lot of work to do.

4:08They've helped sort of start this journey

4:09but organizations need a lot more help.

4:12Now they did this by announcing a new day

4:14for data.

4:15This is kind of like a tagline to say look

4:17from this day onwards we're taking a new

4:19direction and based on what comes later in

4:22the keynote you're going to see why I think

4:24that was appropriate.

4:25But nonetheless they did it in the

4:26following context.

4:27They were talking to consumers, developers,

4:29IT professionals and of course analysts.

4:32Those were sort of the four headline groups

4:33they were talking to.

4:35And so if you're any one of those users

4:36this keynote is for you and so let's get

4:38stuck into some of what was announced

4:40throughout the rest of this keynote.

4:43Now I have to preface this next section by

4:45saying Tableau had to talk about AI at this

4:47conference.

4:48AI has just mothballed the entire

4:49technology industry and whilst they'll all

4:51admit they're still scrambling to figure

4:53out where it's useful, how it works and

4:55everyone's basically just throwing

4:57everything at the kitchen sink.

4:59They still had to talk about it and so this

5:01next section was about how Tableau is using

5:03AI and it was split into two sections.

5:06First and foremost the way Tableau is

5:08already using AI.

5:09Let me pull up my notes here because I can

5:10't remember all of this off the top of my

5:12head.

5:13Firstly there's obviously been investments

5:15into AI over the last few years.

5:18We've had our data, data stories, explained

5:20data and of course Salesforce had Einstein.

5:23So AI has been powering all of this

5:24technology and so I think it's important

5:26for everyone to remember that because it

5:29might seem that AI is just something in the

5:31last few months but actually chat GPT did

5:34something.

5:36It made it sort of a colloquial term and it

5:37's made it a term that everyone's now paying

5:40attention to.

5:41So Tableau when they sort of preface this

5:42stuff they have to kind of explain that

5:44look we're not just jumping on the

5:46bandwagon.

5:47We've been part of this train for some time

5:49.

5:50And then they moved on to their

5:51announcement called Tableau GPT.

5:54Now this is a super sort of hard thing to

5:57explain.

5:58I sort of want you to bear with me here and

6:00the reason this is is because when you've

6:02heard of chat GPT.

6:04People have often thought that GPT is a

6:06technology that chat GPT and only chat GPT

6:09has actually innovated.

6:11GPT is just a type of technology that works

6:13specifically around large language models.

6:17And any company can essentially build

6:19technology using GPT technology.

6:22And so in this case Tableau themselves have

6:24been investing time and effort into

6:26building their own large language model

6:28that works within its products.

6:30And so it's colloquially named Tableau GPT.

6:33But when you see this name you might think

6:35they're just deleting chat in front of it

6:37and putting Tableau and calling it a

6:39product.

6:40Well it can seem that way but that's

6:41actually not the case.

6:42And so you're probably wondering well what

6:43is it then?

6:44Well the thing is that this is an

6:45underlying technology.

6:47The way you have to think of it is you have

6:49to think of it a bit like the way Hyper

6:51works.

6:52Hyper is a data source engine if that makes

6:54sense.

6:55It's sort of a way of moving data from one

6:57place to the next in a fast and performant

6:59way.

7:00And it works in Tableau Prep, it works in

7:02Tableau Server, it works on Tableau Mobile.

7:05All of these technologies use Hyper and

7:07Tableau GPT is very similar.

7:09It's not a feature, it's not one thing, it

7:11's going to be an underlying technology that

7:14turns up in lots of different capabilities.

7:16And actually throughout this keynote we saw

7:18it in Tableau Prep, we saw it in the embed

7:20ding platform coming up with different ideas

7:23.

7:23And then we also saw it here inside of the

7:26Tableau Pulse which was the next thing they

7:27announced.

7:28Tableau Pulse being a new feature building

7:29off the work they've already done on Metric

7:31.

7:32So let's hop into Tableau Pulse.

7:34Now to tee up Tableau Pulse they

7:35essentially brought a customer on board.

7:37They had a short promo video talking about

7:38John Lewis and Partners and how they're

7:40using Tableau to push the business forward.

7:43It was sort of put to use as if John Lewis

7:45and Partners are already using this new

7:47capability with Tableau Pulse.

7:50But I think that was just maybe a little

7:52bit of staging and storytelling.

7:54It might actually be the case because this

7:55feature is probably still in development

7:57and it's probably still being sort of

7:59worked on.

8:00But they needed to bring it to life and I

8:01think John Lewis and Partners said,

8:03"Yeah, look we are going to be using this

8:04feature in the way that you've described it

8:06."

8:07So it kind of makes sense.

8:08Now they moved on to basically what feels

8:11to me like Tableau Metrics Unleashed.

8:14So Tableau Pulse is essentially this, think

8:16of it as a place, think of it as a landing

8:18stage.

8:19Think of it as a new tab on the left hand

8:21side of Tableau Online or Tableau Server

8:24that allows you to view certain bits of

8:26information.

8:27And previously I would have called these

8:29just metrics.

8:30Metrics are essentially the capability

8:31where you can go to a dashboard, pick a

8:33data point and then create a sort of a

8:35number or a chart that is driven by that

8:37data point throughout time or throughout a

8:39certain context.

8:41This interface, this Tableau Pulse setup is

8:44another level.

8:45Essentially it allows authors to build

8:46these metrics without having to use a

8:48dashboard or visualization and then present

8:51them in a place which is Tableau Pulse.

8:54And then once they use it in Tableau Pulse,

8:56what is happening is that Tableau GPT

8:58technology is sort of being enhanced into

9:00the whole product.

9:02So you can get prompts for example, what

9:03questions to ask.

9:05You can type in questions yourself and

9:07Tableau will go off and analyze it.

9:09But you don't have to type those questions

9:10in a way that's prescriptive.

9:12You can ask for example, they showed a demo

9:14of asking Tableau what else should I know

9:16about and it came up with an answer because

9:18it understood the sort of points of the

9:20question and it had a good understanding of

9:22the data source and the metadata.

9:23To go and do that analysis for you.

9:25And of course the whole time you can always

9:27step out of this experience straight into

9:29something familiar for an analyst that

9:31allows you just to build and play with a

9:32chart.

9:33So that is a very brief description of

9:35Tableau Pulse.

9:36I'll do a separate video breaking this down

9:37.

9:38But I think it's just important to call out

9:39that staging and that orchestration of how

9:41this all comes together.

9:43It's a whole range of features.

9:45Tableau Pulse being the new thing.

9:46Tableau GPT being the underlying technology

9:48and metrics what we used to know as metrics

9:50being unleashed into this new capability

9:53with a dedicated home.

9:55Let's move on to the next part of the

9:56keynote.

9:57Now Francois then moved us on to Pedro who

9:59talked a bit about embedding.

10:01Now the thing about embedding is embedding

10:03has been sort of this momentum that's been

10:05growing over the last two years.

10:07And the most important thing that I think

10:09has changed in embedding which I'm going to

10:11cover in a separate series of videos has

10:13been the embedding API version 3.

10:15That seems to have opened up a whole world

10:17of possibilities and it feeds nicely into

10:19some of the announcements here at the

10:21keynote.

10:22What Tableau did is they started off by

10:23reminding everyone what's been released

10:26because there's been a lot.

10:27There's been something in every release for

10:28the last year.

10:29And then Tableau dropped I think the most

10:31important thing at this whole conference.

10:35When I first saw this and I have to say I

10:36didn't first see this in the keynote.

10:39I saw it a while ago.

10:40When I first saw this and I heard about it,

10:42my mind was blown.

10:44Tableau announced something called the

10:46Tableau VizQL Data Service.

10:48And what this fundamentally means is a dec

10:51oupling of the way Tableau works.

10:54And by that what do I mean?

10:56Well let me use an image to explain this.

10:58I'll put something up on screen.

11:00When you build a visualization and you then

11:02put it on Tableau server and Tableau

11:04renders the visualization, there's a series

11:06of steps in the pipeline.

11:08The first one is obviously getting

11:09information from the data source, feeding

11:11it, running all the queries and then

11:12generating your viz.

11:14What this announcement is basically talking

11:16about is decoupling that.

11:18So now you can have an application query

11:20data from Tableau without having to render

11:23anything.

11:25So it allows developers to build

11:27applications and technology on top of VizQL

11:30.

11:30Tableau's proprietary data engine that

11:32requests data from data sources and

11:34requests data from lots of different

11:36services.

11:37Now there wasn't too much detail in this

11:39keynote.

11:40We didn't go into all the weird demos and

11:41quirks that you might have around things

11:43like virtual connections and all of that.

11:46But if Tableau really implemented this

11:48really well, this could be absolutely mind-

11:50blowing.

11:51Because it could open up a world of

11:53applications that sit alongside and on top

11:56of Tableau.

11:57So super excited to see how this works and

11:59super excited to see how Tableau enables

12:00this throughout the whole extensions

12:02universe.

12:03I'm hoping that's going to be something

12:04that's possible, have an extension inside

12:06of a dashboard that's able to call this

12:08service and do incredible things.

12:10That would be amazing. Anyway, that's a

12:11very brief rundown.

12:13Again, I'll break this down in more detail

12:14in a separate video.

12:16Tableau then finished this section by

12:17talking and giving a demo of how this works

12:19, essentially showing an embedded platform

12:21solution from a company called Cumulus Bank

12:24.

12:24And they walk through all the different

12:25things you can do, embedding the

12:27visualization building interface that's now

12:30possible.

12:31As of a year and a half ago.

12:33And then of course, showing how this visQL

12:35data service can query data just using a

12:37few lines of code.

12:39And the very last thing, they talked about

12:40the Tableau embedding playground.

12:42Now, if you've heard about this before, you

12:44might know Andre.

12:45Andre built an embedding playground

12:47previously in his previous role at the

12:49Information Lab.

12:50And of course, Tableau has basically taken

12:51that idea and taken it to 11 by building it

12:53out and adding a whole load more of drag

12:55and drop capabilities to it.

12:57So super exciting to see.

12:59Let's move on to the next session about

13:00Salesforce Data Cloud.

13:02Now, this section about the Salesforce Data

13:03Cloud was super interesting.

13:05Essentially, it hinges around Salesforce

13:07and it's a capability that I'll try and

13:09explain in as short detail as possible.

13:12First of all, you have to ingest all your

13:14data into Salesforce Data Cloud.

13:16Once the data is there, you can start to

13:18sort of draw these relationships, build out

13:20a data model essentially across all your

13:22data sources.

13:23And then once you've done all that, once

13:24you've invested all that time and effort,

13:27it's very easy to just take all of that

13:28work and use it straight away in something

13:31like Tableau, whether it's in the browser

13:33or in Tableau desktop.

13:34They showed a demo of downloading a single

13:35data source that already had relationships

13:37built.

13:38And you have to be honest, if you're an

13:39analyst, this is sometimes the bread and

13:41butter of your work, finding the right data

13:43source and making sure it has the right

13:45things in it to do the work.

13:47And so if you can build this sort of

13:48collective hive mind of information that

13:50keeps this place, that keeps this

13:52information centrally, then I think it's a

13:55super compelling reason to consider Sales

13:57force if you don't already have something

13:59that's doing this.

14:00And so, you know, Salesforce is always a

14:01bugbear in the room sometimes when you're

14:02in a Tableau community.

14:04I think at the beginning of the conference,

14:05Francois said something about Salesforce,

14:07one person cheers and, you know, Francois

14:08had to say, yeah, we can celebrate Sales

14:10force.

14:11You have to be honest, like, you know, this

14:13is going to happen. This is the way forward

14:16. This is always going to be sort of the

14:17unique selling point that Tableau has

14:18working really well with Salesforce data.

14:20The other thing that I thought was super

14:22interesting here is it's very evident that

14:24Salesforce is now using hyper technology to

14:26make the Salesforce platform work.

14:29That's sort of the other way of looking at

14:30this. If you're sort of a Salesforce user,

14:32you'll look at this and you'll go, oh, okay

14:34, this kind of innovation really wasn't

14:36possible because Salesforce didn't have the

14:37sort of analytical power in the backend.

14:40To make these things work and with hyper

14:41and some of the innovation from Tableau,

14:44these things are starting to move again.

14:46And I think that's a good thing to

14:47celebrate for your Salesforce users. So

14:49that was good to see. Right.

14:52Let's move on to the thing that I love most

14:54, devs on stage. Now for devs on stage, I

14:57counted 17 features showcase.

15:00I think this is the most number of features

15:01I've ever seen on a dev on stage. And it

15:03was kind of hard to sort of realize this

15:05because a lot of it was sort of joined

15:07together.

15:08So the key things they showed were Tableau

15:10prep, Tableau public, Tableau desktop and

15:12Tableau server. Of course, those are the

15:14core products.

15:15So that's how they sort of packed them

15:16together. But it wasn't sort of as clear

15:18cut as that. They sort of flowed through

15:20the different demos in a really interesting

15:21way.

15:22Again, I've said this already. I'll do a

15:23separate video in each of these 17 features

15:25. I think they all deserve it.

15:28So over the next week, be sure to subscribe

15:30and follow the playlist that this video is

15:31in to find out more about the features.

15:34But let's call out some of the highlights.

15:35So the first one I'll talk about is Tableau

15:37GPT. We actually saw some practical use

15:40cases of Tableau GPT working both in Table

15:42au Pulse.

15:43But now we also had it solve a problem in

15:45Tableau prep. There was Tableau GPT in the

15:48calculation window that wrote a piece of

15:50regex to go and get the code to extract

15:52emails.

15:53That's the kind of stuff that I think is

15:54mind blowing. That's what we were all going

15:56to do in chat GPT.

15:58But to be able to do it inside a Tableau

15:59prep without having to do all of that other

16:01stuff, that's simply sensational.

16:03Now, I hope this doesn't require things

16:05like Tableau Cloud and I hope this kind of

16:07doesn't get limited to people who are using

16:09cloud technology.

16:11I'd love to see it in Tableau server as

16:12well. So let's wait and see the details of

16:14this before we get too excited.

16:16The other area that got a whole load of

16:17love is Tableau Public. Tableau Public got

16:20a whole suite of UI changes, the ability to

16:22sort and filter, the ability to add

16:24pronouns.

16:25And then moving on from Tableau Public, we

16:27also got some capabilities in Tableau

16:29Desktop, for example, shared modeling in a

16:32data model.

16:33So instead of just having single

16:34relationships going out, you can now have a

16:36data model that looks a bit more familiar

16:39with something like Power BI.

16:41Essentially allows you to build that

16:42multiple relationships without having to

16:43constantly bring the same thing in.

16:46And then obviously, as you're building your

16:47visualization and you're giving the context

16:49to Tableau, it's able to switch between

16:51these models dynamically.

16:53For accessibility, we have the ability to

16:55edit all text. So that's the ability to

16:56essentially add the alternative text that

16:58you normally get with visualizations.

17:00In previous releases, they actually added

17:02this capability using the data stories

17:04capability to sort of enhance that.

17:07But now it's a separate thing that you can

17:08edit as part of the interface.

17:10We got geocoding natively inside of Tableau

17:12, so you don't have to go off to third party

17:14geocoding services.

17:16The desktop products and I assume the web

17:17product will have this built in.

17:19Again, I really hope this doesn't hit the

17:21cloud, but we need to wait for the detail

17:22to find that out.

17:24We also got some nice features around

17:25charts and we saw a demo of a Sankey chart

17:27being built really, really quickly.

17:29And that was super cool because I've

17:30actually shown a separate video of how

17:31quick and easy that is.

17:33But to me, that's a hint. That's a hint of

17:34more chart types coming.

17:36And I actually think Francois called this

17:38out because he said we've got the core

17:40audience that we're going to focus on.

17:42And that's what we think as analysts,

17:43people who use the desktop product day in,

17:45day out.

17:46Now, there were a couple of other small

17:47things, but again, I'll go through these in

17:49more detail.

17:50Go ahead and watch the keynote yourself.

17:52What I'll try and do is I'll try and put

17:53timestamps of the keynote so you can go to

17:56Salesforce Plus and just look at the

17:57individual timestamps.

17:59But I have to do that in a way that doesn't

18:01mess up my own timestamps here.

18:03So be sure to go ahead and check those out.

18:05But that's pretty much it. That's the

18:06keynote in a nutshell.

18:08Those are all the things I wanted to call

18:08out. I really wanted to try and keep this

18:10brief.

18:11It's always impossible, but hopefully this

18:13has been a helpful roundup of everything in

18:14the Tableau keynote.

18:16As I said, we've got more videos coming up

18:18this week.

18:19Stay tuned and I'll catch you in the next

18:20one.

18:21[BLANK_AUDIO]

18:31[ Silence ]