Team DataFam Tableau User Group 29 Sep 2026

0:01 Chris McClellan: So welcome to Team DataFam Tableau User Group Meeting number 93.

0:05 This TUG aims to capture all the in-person and unrecorded presentations and give them a wider audience. So I didn't r- don't really care where you've presented before, actually don't care if you've never presented before. Um, everyone's gotta start somewhere. Uh, please get in touch. We'll find a time slot for you. My email is team.DataFam@Gmail.com.

0:27 But today we've got Matthew Miller, who's the Vice President of Product Management at Tableau, presenting Data Demos and the Art of Storytelling in the Age of AI.

0:39 And Matthew, I'll hand it over to you, and you can take it from here.

0:43 Matthew Miller: Thank you, Chris.

0:45 About 30 years ago, I was in high school and an aspiring leader and not very good storyteller, demo-er, or public speaker.

0:57 And I thought, "You know what?

0:59 I wanna practice leadership," and so I decided to sign up and to run for student government. Now, we've got a wide range of folks on here, including many who are not from the US. I'm not sure what the norms are in high schools everywhere, but let me paint this picture.

1:15 I'm at an all-boys Catholic prep high school, the kinda place where you had to wear a jacket and a tie and a button-down shirt and ill-fitting trousers and ugly shoes you bought from a secondhand store, at least that's what we did.

1:31 And student government is a, uh, the idea that each class elects some officers, a president, a vice president, a secretary, and they have some authority.

1:41 I'm here to tell you in retrospect that was all a myth. We had no authority whatsoever. But it was fun to play act like we were leaders, and I thought it really meant something.

1:50 I didn't think, I didn't have the hubris to think I could be class president, but I thought I could be class vice president, so I ran for class vice president, and that meant I had to give a speech on election day.

2:05 So election day comes. Uh, picture your typical American gymnasium. If you're not from America, just picture every high school movie you've ever seen. It's got the basketball goals at one end or the other, kind of pulled up. It's got that lacquered floor and those very, very uncomfortable bleachers that can i- i- retract and extend. And my whole class is sitting along the entire length of one side of the g- gymnasium, and right in the center at half court was this lone microphone on a lone microphone stand, a terrifying place to begin.

2:42 But I was prepared. My mother had helped me make my posters, and she had helped me with my speech. And we had my speech on index cards. So I had a set of index cards in front of me like this, and, and I had my points very well organized and quite logical. I was quite thoughtful, and, and I'd done what research I could about the needs of my classmates, and I began.

3:06 "Uh, hi, my name is Matthew. I'm running for, uh, uh, junior class but-" "Louder. We can't hear you." Well, the problem is I was at that time about a meter 93, about 6'4", we'd say, scrawny little bean pole of a kid and much taller than the previous speaker, who had been boring and terrible, I might add. And, and so I'm, I'm having to adjust the microphone.

3:30 So now I'm flushed. I'm nervous. I'm fumbling. I'm trying to not drop my index cards while I adjust this microphone stand, and it was unfamiliar with its operation anyway. And then, a- and this, this is where it went from bad to worse. Mrs. Cross, the algebra teacher, had to get up from her seat at the edge of the, of the gymnasium and walk all the way across and help me adjust it, which is really embarrassing.

3:57 And once it got to the right height, now I resume my squeaking speech.

4:02 If it had been possible to turn sideways and to become invisible, hiding behind that narrow mic stand, I absolutely would have.

4:10 I was so embarrassed, but I, I pressed through with my terrible speech with my points, things like… The very inspiring points, mind you, like, "If I am elected your class vice president, I will lobby the administration to get us Mountain Dew and Jolt Cola." That tells you how old I am precisely because there was this cola with absurd amounts of caffeine that was sold for, like, three years before parents banned it everywhere.

4:31 "I'll get Jolt Cola in the cafeteria." I was powerless to do so, mind you, but I was a politician, right? Big promise.

4:39 It was terrible. I hated it. They hated it. People were awkwardly fumbling with their notebooks, anything to avoid the, the awkward, painful, uncomfortable experience of listening to me speak. I got through my points. I slinked back, and I sat down.

4:54 And the next candidate, a young man named Peter Nonas, curly-headed kid, sat behind me in, uh, English class. He, he approached the microphone in a way I can only say he stalked up to it. He swaggered. Is swaggered a verb? He swaggered to the microphone, and he grabbed it out of the mic stand. I don't think I have a microphone in here. We're gonna use my water bottle as an illustration.

5:17 He, he grabbed it like this. He took command of the microphone. He coiled the microphone cable around his arm like, like, like he was in Limp Bizkit or something or, or a Beastie Boy, and he broke the fourth wall.

5:31 The thing I was trying to hide behind, that narrow microphone stand, he stepped right in front of it, and he brought this energy. I mean, he looked at all of us, and he said, "Who's the best junior class ever? We are.

5:45 We're gonna have a great year. You're gonna vote for me, and we're gonna have the best class ever and blah, blah, blah, blah, blah." Now, I say blah, blah, blah, because I don't remember a word he said, and none of us did, and it didn't matter.

5:56 Because immediately we all dropped our homework we were trying to cram or the exam we were trying to quickly study for, and we began paying attention to Peter Nonas. And he paced back and forth like a panther, owning the room- With his charisma, with his energy.

6:11 It wasn't about the words so much as the way he presented it. He told a story that was inspiring. He cast a beautiful future for the Bishop Hendricken junior class of 1998.

6:24 And then he did something I, to this day I think I'd be afraid to do. It just takes confidence, but he did this. He said, "My name's Peter Nonas. Now, who are you gonna vote for? Peter Nonas. When I say Peter Nonas, I say Peter, you say…" And he held the microphone out in front of him.

6:44 Now, if he had approached that moment, we're gonna pause, record scratch, freeze in time. If he had approached that moment with the level of confidence and charisma that I, hiding behind my microphone, terrified, shuffling through my index cards, I would've said, "When I say Matthew, you say Matthew Miller." Boo.

7:06 But, but he said… I'm gonna un-pause here, because he said, "When I say Peter, you say Nonas. Peter." Everyone goes, "Nonas." "Peter." "Nonas." "Who are you gonna vote for?" "Peter Nonas." "Peter Nonas." "Who has the best class art?" "We are." "Yeah, vote for me." Slammed the microphone back in the stand, walked back to the stands to thunderous applause.

7:26 If my name hadn't been on the ballot, I would have voted for Peter Nonas.

7:31 Now, I just turned 45. That story took place almost exactly 30 years ago, and I've not talked to Peter since graduation day, so 29 years.

7:40 I will never forget his name.

7:42 I'll never forget his name.

7:45 Covertly I was going, "Nonas? Nonas?" Because he taught me something really important about communication. He taught me something really important that's applicable to the work that all of us do in data. 'Cause if data lives in a vacuum, right, if a, if a viz falls in the forest and no one tells its story. Someone said recently that stories are data with a soul.

8:08 One of our customers said that to me at Dreamforce.

8:11 These three things here, data, demos and the art of storytelling, they weave all together because they are the way that we really translate the things that we find in data. We bring it to life by showing it to people, and they understand it when they hear it in the context of a story. So, uh, it's such a delight for me to get to spend time with you today to share stories of my worst mistakes, uh, show you some demos that I've done.

8:34 I get to show you, I think, one of the first vizs I ever built in Tableau. Please don't boo me. I won't look at the chat so my ego remains, uh, protected. There's some really talented viz designers here, so go easy.

8:45 And share the things I've learned, and then at the end we'll have some time for some Q&A. And I'm very grateful for the long line of mentors and coaches as I share my screen here who've invested in me over the years. I hope, if nothing else, I can share back with you what they've taught me about these three topics. And this is a Gemini-generated image, uh, uh, attempting to capture the energy of Peter Nonas.

9:05 But you'll notice one really important distinction between reality here,

9:11 we didn't have girls in our school. That would've changed everything. I would've been even more terrified.

9:16 I won't say her name, but if the girl that I, I really liked when I was a junior had been in the audience, oh my goodness, I would've never even done it.

9:23 So we're gonna talk about data, demos and the art of storytelling, analytics with impact in the AI era.

9:28 Uh, my name's Matthew. If you wanna get to know more about me, you can scan that QR code or grab me at Dreamforce or TC. I'll probably be swapping stories with someone here. I work in product at Tableau. I've been doing that a really long time, and, uh, you can tell that by the gray hair here and the lack of hair here.

9:46 But we're not here to talk so much about me. All you probably need to know is I've got an enormous family. That's a really important data point. We have between four and seven children, because with large numbers, you have to approximate. If you were to visualize my children in Tableau, it'd be a bunch of hash marks because the text would get cropped.

10:01 That's an OG joke. I know some of you get that. Give me your emojis in the chat if you know what I mean.

10:06 We lived in Europe for about nine years. Two of those kids were born in a bathtub there. Um, I would say two and a half of the kids are native Dutch speakers. The half is the 14-year-old, who's kind of forgotten about half of it. Uh, we live now in Chicago, where my wife is doing a PhD. That's her up here in the top right.

10:23 I've been playing a card game since I was a little kid. The game is called Cribbage. Cribbage is one of the few ancient card games where they actually know the name of the person who invented and when. A guy named John Suckling in the UK in, I believe, 1590-something. Someone will Wikipedia that and correct me.

10:39 This is a picture I took myself. I've been playing this card game since I was a little kid. My grandparents played, my parents play. In fact, I play with my siblings all the time on a online app.

10:50 And what I love about Cribbage, and this is not a story about Cribbage per se, trust me, it's about data. But what I love about Cribbage is that it's pretty easy to play. I like to say you can, if you can count to 31, you can play Cribbage, and it's a really good mix of skill and strategy, but also blind luck. So a brand-new player isn't gonna get blown out of the water.

11:10 Like, if I sat down with Magnus Carlsen, I would win zero games out of 100 if we played chess. But if Magnus Carlsen learned to play Cribbage, I would wipe the floor with that guy, but he'd probably win about a third of the time.

11:23 Cribbage is a very simple game, and your goal is to get to 121 first. That's this part here. This is a non-standard board. Uh, I won't explain the details behind that, but you, uh, get a lot of little points in the game. So throughout the game you, you score hands in a bunch of different ways. The details aren't important. You get a point here, two points there, four points there, sometimes 12 or 16.

11:43 This board is, is in the shape of a 29 because that is the mythical best hand you can possibly get. It's never happened. I've never gotten one.

11:50 So when I first discovered Tableau, and I've told this story a few times. We're gonna go over to Tableau. I opened a product that looked awfully like this. It was version 6.1, so it was a little grayer. It was on Windows, but it looked a little bit like this. So, uh, my first viz ever was an Excel file for work. I can't show that to you, of course, even though who cares about customer data from 15 years ago?

12:11 But I dragged in this thing and I saw a time series, and I was like, "This product's really cool." The second or maybe third viz I ever built, though, is I took a data set I had- On the probability of winning in cribbage. I mentioned that you get this, uh, board like this, so if you're playing a two-player game, your score's going back and forth. And the way the game works is that, uh, the deal goes back and forth.

12:30 So y- the two players take turns dealing, and the dealer gets a bonus hand. So the dealer tends to outscore their opponent, but then they hand the deck of cards over, and their opponent deals, and then they tend. So there's this, like, leapfrog effect throughout the game.

12:44 Well, I had this really interesting data set that I tried to explore about the likelihood of winning. Now, I'm gonna show you the raw data. Just look at all the insight leaping off the page. Just simply… I, I think I say that same thing anytime I show a table of data. I think I said this at DataFam Europe. I mean, tables of data are great, but, like, there's no insight to be found here.

13:08 I c- can't see and understand a darn thing. But I took the dealer score and the non-dealer score, and I put them on the X and Y axes, and I had this little kind of quasi scatter plot. You can see for each combination. So this is saying, let's say that the dealer's current score is 50, and their opponent is 62, and they're about to shuffle and deal for that hand.

13:28 The dealer has a 28% chance of winning on average.

13:32 Okay, that's kinda cool. So I brought in your chance of winning, and, uh… Let's see. I think it's this one. There we go.

13:40 And immediately, whoa, I saw some things in the data that I'd never seen before.

13:46 Uh, I call this the flag of cribbage. I gotta credit Ben Bausili, some of you may know from InterWorks, for calling that. He and I have played thousands of cribbage games.

13:54 I'll point out a couple of things that just jumped right out to me. First of all, when you play the game, someone's gotta deal first. I mentioned the dealer tends to score more than the opponent, and that goes back and forth. I didn't think it mattered who dealt first, so when I would play with my little sister, Joanna, she'd be like, "I'll deal first." I'm like, "That's fine." Well, check this out.

14:13 The, uh… Even here, like, at 0-0, I'll zoom in so you can see that, even at 0-0, she had five percentage points above the, above the middle. She had a, a, a, actually, I guess you say 10 percentage points greater likelihood of winning at 0-0, just because she so graciously agreed to be the one to shuffle the cards. I should ask her if she knew that.

14:36 She's very smart. I wonder if she knew that. So she was getting an advantage just by generously offering to deal. Now I'll just say, "Hey, let's cut for deal. Whoever gets the lowest card goes first." The other thing I learned immediately looking at this, do you see how this, this kind of, like, sine wave approach that gets more intense toward the end of the game?

14:53 Because of that leapfrogging effect, I observed this really interesting pattern right here, where there's a run of positions where you can be anywhere from, like, 11 or 10 points behind your opponent to tied, and you have about the same likelihood of winning.

15:12 Fascinating. You might be at the end, end of the game, going, "Wow, they're five points from winning. There's no way I'm pulling this one off." And actually, you've got a 50/50 shot of still pulling it off. Things I'd never known about a game I'd played a million times, from a data set I'd had sitting in an Excel file from one of my Brown University math nerd friends I played cribbage with and didn't see and understand.

15:33 This is what I love about Tableau, is it takes the data, and it turns it into a story I can see and understand, and I fell in love with it, and here we are 15 years later.

15:45 Couple of other stories that changed my life and impacted the way I think, uh, a- about the, the power of data and storytelling. Uh, when I was a consultant living in Europe, I got sent to work with a pharmaceutical company, and I said, "Okay, what's our product?" They said, "Well, uh, what we're doing is we're doing, uh, early-stage cancer research with ELISA kits." And I quickly googled what that was,

16:07 because I took one high school biology class and did okay. That's about all I got. Well, they said, "We're doing antigen-antibody combinations." And then I googled under the table to remember what those were, and they put them together in these little pipettes.

16:20 A- and so think of these, as they explained it to me, like a little grid of test tubes, where they can run hundreds or thousands of experiments all at once. The way that these experiments work is they set up an experiment, and they do thousands and thousands of these, by the way. They'll have dozens of these trays running at a time. So they'll arrange ahead of time which experiment will be in which of these little cells, right?

16:40 From top, what you'd call A1. In fact, you can actually see faintly the lettering here, A1, all the way through, uh, down here, it looks like H16 or something.

16:49 So they, they set that up, and they have a database where they track that. Then they put these little experiments in this very s- uh, light-sealed compartment that has very sensitive light sensors, so it's blacked out.

17:01 And what they're looking for is some kind of a light reaction that's, like, not visible to the human eye. They collect those with these really, really sensitive sensors. They store that data in a database, and they try to understand which of these is an interesting candidate to move on to the next stage of research. And I may have gotten the details a little fuzzy there, but this is my, you know, l- l- layman's attempt to understand that.

17:23 Here's what that data looked like in their Postgres database. Once again, no insight just flying off the page here. But as soon as we visualized it in Tableau, and I'm sure the database ex- experts here will have words about my choice of colors. In retrospect, I probably ha- should have the brighter ones be brighter.

17:41 Uh, I'll, I'll… You can send me a book on database, and I will definitely put it on that shelf. Uh, so a few things jump out. One is over here there's this cluster of four of them together, and what they explained is sometimes a reaction is so intense that it causes, like, light pollution. And so they can't tell which of this cluster of four was actually the one that was meaningful.

18:01 So they have to find these clusters of bright lights, of bright reactions, and then they want to randomize them the next time, and put them in different places on the next tray. Run the same thing again to see if which of them was actually the one that mattered.

18:14 And then you can see over here they have a control column on the far right. So if there's a reading, which you can see if I zoom here, the little reading here, that should tell you actually something might be wrong with this sensor array altogether. Maybe it's out of calibration. There, this should be a null or a zero instead of some reaction. You can see how visualizing this made it so much easier for them to look at this and go click, click, click, click, click, click, click, click, click, and, and to optimize for the next round of experiments.

18:40 So there it was. You know, I- when I was a consultant, my kids would go, "What do you do for a living?" Uh, you know, and I'd ask them, "What do you think I do?" They're like, "We think you make pretty pictures." I'm like, "That is a little part of it. That is true." Um, be like, "Well, you think we think you live in airplanes," but that is actually part of it too.

18:53 We think you buy fancy shoes for a living. It does feel that way. Uh, but at this time, I got to come back and say, "You know what Dad did for work this week? I helped cure cancer," like a little tiny, very indirect way. But it was really cool to see how that was done, and how Tableau took them from gathering data to seeing and understanding and acting on it, increasing the rate of, like, compounding interest for scientific research.

19:15 It's one of my favorite experiences I ever had as a consultant, and one of the coolest, like, examples of how Tableau unlocks people's ability to see and understand data. One more example. This is famous. Um, this slide taken with, with full credit to Andy Cotgreave.

19:30 But, uh, you've got Iraq's bloody toll. This is a great example of brilliant data viz. Accurate data, but, but very intentionally inverting, literally inverting or subverting the traditional rules of data visualization to tell a more compelling story.

19:45 And when you flip it upside down, well, not quite as powerful. It's actually tremendously powerful when it's inverted, and of course, the use of color added so much more impact to the message.

19:57 So this was… This is the scientific basis of what makes Tableau special, birthed by scientists, backed by research to tap into the GPU of the mind.

20:07 When we're presenting, when we're storytelling, there are a few things you can take from some of the things you just heard, and I'll… I've got three practical points for you that I'll share in the next section, uh, before we get to Q&A, of things I've learned that each of these and many others have informed. Sometimes you gotta break the rules.

20:24 You can see just the power side by side. On the right is what I would've been taught was best practice data visualization.

20:30 The left, that tells a story.

20:34 And that leads me to my first point, which is hook the audience. Break through the noise. What did Peter Nunes do? He hooked the audience. He hooked the audience before he even said anything. He hooked the audience by his, by his stage presence, by the way he carried himself.

20:51 The thing I've learned about storytelling and public speaking is how important it is to get everyone's attention. It's a quest for attention. It's a fight for attention. Attention's harder to come by than ever. It's really hard to hold the audience's attention, and you don't have any right to it. Even in an internal meeting where I've, you know, people are, uh, I guess you could say, paid to attend, even in that situation, you still don't have a right to their attention.

21:16 You gotta earn it. So hook your audience and break through the noise.

21:20 I'm gonna do something. I know this is recorded, and it'll be on YouTube, and so I apologize, but I don't apologize that much. We're gonna show an example from a competitor of what not to do, and if you're a big fan of Power BI, um, more power to you, but this is not about Power BI. Um, and I, I know this guy, and he's great, but oh, boy. A few years ago, I got to do what's called the Gartner BI Bake-Off, which is the most intense demo I've ever done.

21:45 It's three competitors. It was Tableau, Microsoft, and I think Oracle, and one representative from each had to have a 20-minute long entirely scripted and memorized demo that covered all these things with a data set they gave you. Think, like, kind of Iron Viz, but in front of 4,000 chief data officers who have roughly no sense of humor. It was the most terrifying and difficult demo I've ever done, and thanks to Francois, who some of you know, Francois Ajenstat, the former CPO of Tableau, who coached me a ton on this, we were able to hook people.

22:15 But I'm not gonna talk yet about what we did to hook. Let me share with you how they began their riveting presentation.

22:22 Video clip: For this year's Gartner Bake-Off, we were asked to investigate the likelihood of more frequent flooding events worldwide, and to see if there's any mitigation strategies governments could take to reduce the risk of flooding disasters.

22:33 First, let's take a look at the data we pulled together to analyze this topic.

22:38 We started by pulling in data related to the OECD river flood, uh, risk indicators, which looks at how many, uh, people live in, uh, flood prone areas.

22:47 Matthew Miller: I'm sorry, I can't take it anymore. I am so bored. What am I looking at? Why should I care about this? You have boxes. You have lines. Who doesn't love boxes and lines, but what is this?

22:58 Wh- why should I be excited about this? If this is… How does this impact me or the world?

23:02 I'm, I'm, I'm so bored. Please hook me with something interesting.

23:07 And what, uh, I was coached to do, and I gotta credit, uh, Francois for this, was we came out and instead we, uh, we, we came out and we said, um, let's see if I can do it from memory. It was like, uh, "When we heard that this year's Gartner Bake-Off data set was about flooding risk, it hit really close to home, because home for me for almost a decade was the Netherlands." I put up those pictures of my kids walking around with an umbrella.

23:35 "A country which, as you'll see, is one of the worst in the world in terms of risk of flooding. Most of it's below sea level.

23:42 Hi, I'm Matthew, but this isn't about me. This is about you and the world and the threat that we're under from rising sea levels, climate change, and increases in flooding. As you'll see, with the power of Tableau, you can," ba, ba, ba, ba, ba, "unlock data," et cetera. I mean, we, one-minute intro, everybody's hooked.

23:57 Uh, it helped. Some of you know Christian Felix, Iron Viz winner. He built this incredible viz. I should've pulled it up. Maybe during the Q&A I will. He built this incredible viz that showed, like, using the pages shelf, or maybe it was some parameter thing, but it showed this, like, animation of the flood of the water across the Netherlands. People were…

24:17 And these are, like, cynical chief data officers were pulling out their phones and taking photos of Christian Felix's viz like it was a Taylor Swift concert, and I was like, "I, I think we're gonna win." We did.

24:27 Hook your audience. Again, sorry for throwing shade at Microsoft, but it, the contrast was so obvious. Like, w- again, with, with the help of my coaches and, and others, we didn't make it about me, the presenter. We didn't make it even about Tableau and how great our platform is. Microsoft put up an architecture slide, God forbid, of Azure and Power BI.

24:46 We talked about the world. We talked about them. We talked about, about the impact that this change in the world has on us.

24:53 One more example of this that, uh, to me is, like, just gold standard.

24:57 I did a, an executive briefing in California for a company that does robotic surgical equipment, and I'll confess, like, I didn't phone it in, but I didn't, I didn't put my maximum effort in to, like, coming up with, you know, custom slides and picking just the right story. I thought I did a good job, you know, but I, I didn't bring my A game, and I finished my section of the briefing.

25:20 And you could tell the customer was like, "Uh-huh, okay, cool. AI features, great. Cloud, server, good." And then I sat down. And there's an SE who some of you know, a Tableau SE named Vikram Benderapudi, and he got up, and he put up a slide, looked something like this, and he said, "Surgery.

25:38 The word alone fills you with anxiety." By the way, just that, that opener, every one of us was like, "Mic drop. We're listening. I'm listening. The customer's listening." He said, "A few years ago, my dad required surgery." When the doctor said that, we were all like, "What does this mean?" We were afraid. Like, is this a risk of death? Like, is he gonna be okay?

25:58 Surgery, it's terrifying. "But then the doctor said, 'Don't worry. We're using this new robotic surgical equipment. It's laparoscopic. It's blah, blah, blah, blah, blah.' " I don't remember the words.

26:09 "'He'll be fine.' It put it all us, uh, put us all at ease.

26:13 That's why it's such an honor to speak with you today, 'cause the work that you're doing is making… Is like the reason my dad is still alive, and puts people like me and my family at peace." I mean, you don't always have that story, and like, I- I… My dad has not had robotic surgery. I don't know what I would've done, but wow, I was sitting there going, "That is incredible." That's how you connect.

26:36 And he went on, by the way, to talk about the IoT devices they've got in their robotic surgical, how they collect data. He Googled this, and they have some blah, blah, blah, blah, blah. Now I can't wait to tell you how Tableau can help you unlock the value of that data, so that you can save more lives and put more families at ease.

26:52 Oh my gosh, gold standard stuff, and I introduced myself afterward. He's like… I said, "How did you become that good?" He said, "I took a course on storytelling at Stanford." So there are a bunch of techniques you can use to learn how to do this kind of storytelling, public speaking, um, and hooks. So definitely a shout-out to Toastmasters if you've done that.

27:10 There's a book I recommend called Do You Talk Funny? by Dave Nihill, and there's many other places you can go to learn and practice public speaking. But hook your audience, number one.

27:19 So, so important. Number two, establish trust.

27:23 Why should I care? It's my pinned tweet. I think it's been my pinned tweet for about five years, uh, that I have this assumption.

27:30 No matter what audience, whether it's a Tableau audience or not, whether it's a customer I know really well or one I don't, or a TUG, or internal or external, I don't assume people want to hear me speak or care.

27:44 I have to establish trust. I have to establish a reason for them to listen,

27:50 and hooking the audience is only part of it.

27:52 How many Super Bowl ads have you seen where you're like, "Well, that hooked my attention. It was very memorable. What was that an ad for again?" Right?

28:01 I guess I can't quote one of my favorites, 'cause it's for an alcoholic beverage, but, uh, one of my favorite ads had some frogs in it. You may recall. We quoted it for, y- I still quote it, who am I kidding?

28:12 Memorable, but attached to the brand, so you couldn't forget it. Or maybe a safer, uh, close-to-home example, the, uh, "The man your man could smell like" ads, if you remember that, the Old Spice ads.

28:22 Memorable, hooks your attention, but then brings you along the path to believing, oh, you can actually deliver on what you said.

28:30 This right here is the most impactful viz I've ever created while at Tableau, and it taught me a lot. And, and maybe the thing to take away from this section, it, about establishing trust is it goes beyond just data visualization. I had been taught that a good data viz requires no explanation. Have you ever heard that? Good viz explains itself, tells a story.

28:50 And look, we should strive for that. You think of, like, the Iraq's bloody toll. You don't need a, a, a companion YouTube video or a 20-minute presentation to understand that viz. You look at it, and you're like, "Wow, that's awful." You get it.

29:04 But you can't always do that.

29:06 So I'll tell you what this viz is about, uh, without, well, mentioning anything confidential or internal. Now, if you've ever used Tableau c- uh, Tableau Cloud, if you've ever administrated Tableau Cloud, you've used a feature called Tableau Cloud Manager.

29:19 Uh, give me some emojis in the chat if you have ever administrated a Tableau Server or if you've ever administrated a Tableau Cloud cluster, if you've ever used TCM. Anybody? I know, uh, Chris for sure. Christian, I, I'm sure you have. I'm sure there's others.

29:34 John definitely has.

29:36 Well, if you are a little older, you may remember there was a time where we didn't have that Tableau Cloud Manager equivalent for Tableau Cloud. We had something like that for Server. We didn't have it for Cloud.

29:46 So when I joined the Tableau product team six and a half years ago or so, gosh, almost seven years ago, my first project my boss brought me, she said, "I want you to go fix multi-site in Tableau Cloud." I said, "Okay.

29:59 I know the problem. Sure." And we talked about it, and it was, like, like, the easiest product thing to do in the world, because there was not a lot of original research necessary. We knew there was a problem. Customers were buying Cloud more and more, having more and more complex deployments with more and more users and more and more sites, and they just needed a thing to manage them that worked the way it did on Server.

30:18 As customers migrated from Server to Cloud, they would go, "Hey, can we get in Cloud that thing we've got in Server?" I mean, you know, I don't want to overdo it here, but this one was just pretty much a copy/paste, right? No need for, like- New innovation and research. It was like take the screens over here, put them over there, make them do the same thing.

30:36 Seemed like a no-brainer.

30:38 So I put together a strategy doc for this that, I mean, reading between the lines now, it said… It basically just said, "Duh, we need to do this." And I want to be clear, the, the thing I was asking leadership for as I put together this six-page proposal, was a lot of money. Because I was asking for a lot of engineers to do a big, difficult thing in Tableau Cloud that we already had in Tableau Server.

30:59 It would require, even if it looked the same and behaved the same, required a lot of under-the-hood, very different plumbing. It was a big investment. A multi-multi-million dollar investment I was asking for, so it, you know, I didn't take it lightly. But I, I think there was a little bit in there of like, like if you could've boiled down my six-page document to, "We should do this, duh." And probably some of that came through.

31:21 Side note, I'm really proud. The title of that document, if you like dad jokes and puns, the title of that first strategy document I ever wrote was called Delivering Data Insights.

31:30 Get it? Delivering Data Insights. I know, I was proud of that. No one shouted out my pun.

31:35 Well, I wrote this strategy doc, I deliver it to leadership.

31:38 It's completely rejected.

31:41 I, I remember it distinctly. It was, it was a, a severe severe, uh, embarrassment. I c- I didn't get it. Their set of information and knowledge and expectations about that area are… This was our product and engineering, and, uh, executive leadership at Tableau at the time, was totally different than mine.

31:59 They had a different angle on this. We're asking for a huge investment. We're just b- about to be acquired by Salesforce. We'd already announced we were gonna, you know, our, our, what will our new, our new owner think of such a big spend? Is this really that important? Uh, do customers care about this? And I was like, "Yeah, they do. They all do.

32:16 They all need this." And I made this scatter plot, and I'll explain it real quick.

32:20 Because I thought, oh, well, if you don't get it, I remember, like, I remember sort of, uh, flapping about in the wind in this meeting, like exasperated and flustered, off my game. Like, you know, like, like, like back on that basketball court going, "Well, but, but, but obviously we do this." "Well, it's not obvious to us. Explain it." Uh, w- but everybody needs this.

32:43 But, but thankfully though, I had data viz, right?

32:46 Yeah.

32:47 I, I pull up this chart. Let me explain it. It's a log-log scale, meaning that each i- uh, inch, each index, uh, mark, tick mark on the axis is, like, an order of magnitude more.

32:56 Uh, so you have the number of total active users on a site on this axis, and you have the number of total active sites on the Y-axis. And you can see, like, there's… This was for the top 200 customers we had at the time, and the average number of sites they had was 73. This is in Tableau Server.

33:15 Uh, the outlier at the time had 2,000. I think the outlier now has 5 to 10,000, and I think the customer with the most server clusters has over 1,000 server clusters, which is just ridiculous.

33:26 So customers have used the heck out of this site feature in server. They're trying to move to cloud, and at the time, cloud would let you have one site at a time. If you bought another one, it was like a totally separate thing. They didn't talk to each other. You couldn't manage them.

33:40 So I thought this would do it. So I pull this up and I go, "Well, see?" I mean, "See? Don't you see?" And they're like, "We don't get it." And I remember the CEO of Tableau at the time, a real smart guy named Adam Selipsky, who'd come to Tableau from AWS, he, he asked me, "Matthew, I thought we settled this. I thought every customer had only one site." And the, like, let's think of it from his perspective.

34:02 He's a sales and marketing leader. Uh, he'd been hired to really turn around our, how we did sales and marketing. He's the one who introduced data management and advanced management. He did a lot of great things for Tableau, but he wasn't a nerdy product expert. He'd never fired up TSM and administrated a site. He'd never installed a Tableau Server.

34:20 Uh, by choice, I'm wearing my certifiably Tableau shirt I received when I got my Server Certified Professional certification. He had never spent seven hours, you know, waiting through VMs, trying to reset a backgrounder. He w- he, he had a different set of context. All he thought about mostly was Tableau Cloud, and he thought in Tableau Cloud people just had one site.

34:40 Why on earth would anyone need this? This is all news to me. You can't just waltz in here and ask for, you know, tens of millions of dollars without helping us understand.

34:49 I tell you that story because it was the first time I think I really had to run into, like, the reality of I can build a great viz, an accurate viz. A viz, to me, just totally explains it.

35:02 But if I don't understand the rest of what's going on in my communication, my storytelling, if I don't understand the context, if I don't know my audience, it's all for naught. So I licked my wounds. Uh, I, I went away from that meeting rejected and sad. And I got a, a ping from Francois, who, if you can't tell by now, has, uh, been one of the biggest influences in my life as a mentor and coach and manager, the only boss to tell me, uh, STFU, but not the acronym, at twice in a meeting.

35:32 Um, like the, just a, an absolute hero.

35:36 And he pinged me and said, "Look, let me help you." He said, "You don't understand the politics in the room, which is that, like, you're asking for something very, very technical that's out of the CEO's depth, and you're asking him to make a funding decision that he has to make, and not giving him enough context.

35:51 So he's gonna look to his team in the room, and you didn't prepare them." So he looked to them, and they said, "I don't know. We haven't really dug into this. Yeah, some, but we don't know if it's the thing we should focus on." So I went and took my deck and my document and my viz here, which to me was self-explanatory, and I walked the room, so to speak.

36:11 I met with every single one of those leaders one on one.

36:13 I brought the same thing back, I don't know, three months into my, four months into my Tableau journey. Presented basically the same thing again, same image, and Adam goes, "All right. Is this what we're doing? Should we prioritize this?" Everyone in the room was like, "Yep," and it sailed through, and we built TCM.

36:28 Wow. I learned a lot that day about data and the impact. 'Cause I thought you build a viz like a rat's bloody toll, and everyone goes, "Oh yeah, obviously, we should, we should not do that." You build a viz like this, and they go, "Oh, yeah, obviously we should do that." Not always the case.

36:43 The way that you win hearts and minds, about knowing your audience, it's about establishing trust.

36:49 Many of you have heard my story about the, uh, the pine tree. I'm gonna shorten this a little bit. It's one of my favorite things to do, but I'm… Just in the interest of time so we get the Q&A, I'm gonna show you the short version of this.

37:00 Um, give me a shout-out in the chat if you've seen the pine tree story. I've done this a bunch of times now, but it's still one of my favorite things.

37:08 All right. We'll g- I'll give you the, the, the CliffsNotes version of the pine tree saga. Uh, this is an actual photo of my pine tree as it was, I think, two years ago. It's right there outside of the window where you can't see it on the other side of this webcam. And, uh, my pine tree was dying, so I called an arborist, who came over and said, "Well, y- I know what your problem is.

37:26 The soil's too dry." I said, "Great." And he said, "I bet I should water it. Yeah, yeah, yeah. Keep the soil moisture between 40 and 60% relative humidity. You'll be fine. Your tree will be fine. That'll be 400 bucks." And he left.

37:39 And then I was left standing there looking at the tree going, "How the heck do I do that? I don't know. Do I, like, uh, I'm not calibrated. You put your thumb in the dirt and guess." So I bought this little IoT sensor. I plumbed it into BigQuery, and it looked something like this. There you go. Don't you see the insights flying off the page?

37:57 So I've visualized it.

37:59 Let's drill in. Let's look, uh, the backyard pine tree. This is accurate, so it's on an extract, but it's… This is live essentially. Uh, can anybody guess what the spikes are in the data?

38:10 Shout-outs in the chat if you got spikes. Anybody?

38:14 There you go. Precipitation. It's rain.

38:16 Uh, if I had more time, I'd show you when you bring the weather in, uh, I think I have it here somewhere, with the little rain, my little rain symbols. Yeah.

38:24 Look, I know this is bad data viz. Sue me. Uh, it's my viz, and I'm gonna put little icons, and they're gonna be sized by how much rain there was.

38:33 I know, I know. You can, you can claim hypocrisy, and you're, you're probably not wrong. And so I made my little color thing. This is great. Uh, but the problem was that I had to go to this dashboard to look at it.

38:44 And so, uh, what I found was that I wasn't looking at it as much, and so I finally built it in Pulse. And I had, um, I've signed into Pulse, uh, on my other monitor, which I'm gonna go find and drag over and show you here in about five, four, three, two. All right. You can see Pulse or a window that is loading Pulse, Chris? Yeah, I, I can see it on the preview.

39:05 So once I built this in Pulse, Pulse went ahead and started monitoring this for me and sending me my daily digest of how my pine tree was doing, and visualizing it automatically, and showing me when it was outside of the relative range. I had a link back to my, my original dashboard. It would show me my trends. And, and i- this is where it gets really fun.

39:23 I was looking at this on my phone, and my wife was like, "What's that?" And I was like, "Oh, that's Tableau Pulse. That's our backyard pine tree." And she's like, "Oh. Oh, that's cool. Can I get that on my phone?" So I went out and bought another of these stupid little sensors, and I bought one for the fiddle leaf fig. This is the fiddle leaf fig in the dining room, and I stuck a sensor in that thing, and I built a metric.

39:42 And I think I'm signed into the wrong cloud site, so, okay.

39:46 Uh, that's my bad. But so I got little metrics. I'll show you how they look in Pulse. So now she can see how her fiddle leaf fig is doing. Well, flash forward, now I think we have, I don't know, 15, 20, 30 of these things around our house. I have one in the front lawn. I got one in the plant in the dining room. It's gotten really out of control. But what that's done is it's taken the power of the data visualization, and it's surfaced it in a place that made sense and could drive action, so Becky can just go, uh, "What the heck is going on with our, uh, humidity in our backyard pine tree?" Um, but explain it in, let's see, um, Fi's here.

40:26 An Australian slang.

40:29 Fi, I know you're… I think you're actually Kiwi, but we're gonna, uh, put New Zealand references. There we go. And so now what we, what, what's happened is I've taken the, the power of Tableau to correctly pull together that, that unusable data as it flowed out of the IoT sensors. I've taken the insight that I've validated and verified in my visualization, but it's no longer just about the viz.

40:54 It's that truth in the trusted data. And so, uh, here we go. I, I have no idea if this is offensive or funny. It should be pretty good. "It's not a drama. Humidity's eased off a touch compared to last quarter. Right now it's cruising along steady. It's still within the sweet as seasonal range." Yes. I've been to Auckland once, and I still remember sweet as, and having someone explain to me that it wasn't how I heard it.

41:17 That helped a lot.

41:19 This is the power now of bringing together the, like, the trusted data that I could see and understand with these AI tools to help me not only, like, disambiguate things that are as vague as, "What the heck is going on?" But also translate that into new ways to take action.

41:34 So this is some of the cool things I'm excited about now. If you've not used Tableau MCP, for example, you totally should have. Uh, I will say the way I predominantly interact with the data around my house now is I build data sources. I build dashboards and vizes that I'll use for certain places. I have the Pulse metrics actually on my home page. If you've not used this, it's super cool.

41:52 Relatively recent feature. Uh, you probably can't see this, but I've got the Pulse widgets on my home page to track the soil humidity of that tree, and my pushups, which is a depressing stat.

42:02 Exercise to date. And you can put the metrics that you care about right there, and the prodo- predominant way I, like, interact with it conversationally is through, uh, Pulse Conversational or with Claude through MCP, which is really cool.

42:16 So now, uh, in the process of… Yes, that was kind of a demo of Tableau, wasn't it? But it was also a bit of a story. Let me take you through what I did there, or attempted to at least model, in the course of that story.

42:30 And it was inspiration, casting a vision for a beautiful future. So this is a framework I learned from Francois. Problem, solution, achievement. So I had a problem, and I tried. Hopefully I did a good job there with that little meta demo, right? A demo of a demo about a product.

42:48 Knowing my… Well, this is a smart technical data viz audience, so I try to work in some data viz slangs. I, I try to address a problem that I know we all have, which is I got data around me, I'm trying to understand it and surface it in the places that I am, and, and try to make it our problem.

43:04 I never wanna stand and look at an audience and point at them. I wanna, like, metaphorically stand next to them looking at our problems. Like, oh, my goodness, isn't it such a pain? You know, my tree's dying, or my, you know, my sales are down, or failure rates are up.

43:19 Then I, on to solution, and here you wanna be clear, you wanna be succinct, you wanna be passionate. It is no fun to hear somebody describe amazing technology like they're describing sorting their sock drawer or weeding their garden. Be clear and succinct and be passionate about what you're doing, and then highlight the achievement, the beautiful future.

43:39 Highlight the way that things will be better once they have this solution. This is a framework we use with the Devs on Stage crew. Every demo, every feature we talk about or highlight needs to have an easier, better, faster, or more, and ideally more than one of those.

43:55 Now it's easier to get your pulse metrics right on the homepage of your phone, right? Now you get better insights thanks to the pulse insight engine. Now it's faster to visualize 'cause we build the charts for you. Now you get more understanding. You have more types of charts. You have more types of insights. You have more places that you can access your data.

44:17 Easier, better, faster, more. Problem, solution, achievement.

44:21 So that framework has helped me with, uh, craft a demo. Uh, there's so much more we could do, uh, talking about demo craft. Uh, one shout-out, you can go listen to Fi and Sarah's podcast, where we spent a whole deep dive just talking about demos and crafting demos. But that right there, the problem solution achievement framework, is been absolutely clutch for, and it's how I start with crafting, uh, every, every demo or presentation.

44:44 As a bonus, I wanna share some tips with you.

44:47 Uh, these are, uh, from a doc I wrote a while ago, and I shared it internally. I was unable to, thanks to our security, which, like, said it was confidential or something, I was unable to, like, make it wide open to the internet. But if you scan that QR code and request access with your Google account, I'll approve all of them. Uh, there's nothing actually confidential in there.

45:06 And this is Matthew's top 10 tips for, uh, doing a great demo or a great presentation.

45:13 Uh, that is a, that's me with hair in that grainy YouTube video presenting a really terrible piece of software, uh, not Tableau, pre-Tableau, at some conference. But, uh, I'll, I'll call out maybe one thing that I, I didn't mention so far, which is using physical metaphors, and kinda similar to that point eight about humor.

45:35 When I was a Tableau Server trainer a long time ago back in the pre-TSM era, I would explain to people, I'd be training them on how a cluster is designed and how the worker nodes, as they were called, and the, the main node or the primary node interacted with each other. And I would use this metaphor. I'd go, "Imagine you got a mommy bird, you know, these little baby birds, and the baby birds are like, 'Cheep, cheep, cheep, cheep.

45:56 Mommy, we're hungry.' And the mommy bird's like, 'You're gonna eat food,' and they're like… And then the baby birds will eat whatever the mommy drops into their mouths. That's it. They have no choice whatsoever. There it is." And Tableau Server's primary node is like the mommy bird, and these worker nodes are like the baby birds. And, you know, when I trained a, s- a, a German, picture, like, a German audience of Tableau Server administrators, in fact, I can picture one of them, uh, in Frankfurt, you know, looking at the weird American with the weird shoes cheeping like a bird, but I bet you they didn't forget that metaphor.

46:30 So file that one away for the next time you train somebody on Tableau Server, baby bird and mommy bird. Top 10 tips. Feel free to grab that and, and use it. There's a bunch more detail than just that, those bullet points in that doc.

46:43 That one's a freebie. All right. We're gonna come in, uh, ooh, we're gonna come in quick on Q&A really quick. Let me just say o- one example where I got to use this was my favorite demo of all time, and I'll, I'll drop the video in the chat or somebody can Google it. It's the Devs on Stage from TC'22, I think. It's the Tableau Gestures demo. Uh, e- easily my favorite demo ever, and, and all I did, I was just the talking head doing the demo.

47:07 It was our research team who did all the actual amazing work and our design team who made this look… It already looked good. They made it look amazing.

47:14 Uh, what we did is we took that demo that they had, and, and we'll kind of workshop this together. I, I, I can probably find quickly, uh, an example of what it looked like before, and I'll put that in the chat.

47:26 Um, the original version of this, it's like, "Hello," looked like this. First of all, it was called, I'm not making this up, uh, Augmented Chironomia for Presenting Data to Remote Audiences. Talk about a, just a hook of a title. If you scroll down and look at what they did there, I guess you have to click through, there's, like, a video and other audiences.

47:47 Uh, uh, if you click on the video there, you'll see it. What they built was a thing where you could pull up the webcam, and then you could drag some completely finished charts into view, and it would sort of sit on top of your face.

48:01 So we took a look at this, and we applied that problem solution achievement framework, and we applied that easier, better, faster, and more to it. This is Francois, me, and a few others. But, hmm,

48:12 why would we ever want this? Well, the problem we have is that back in the day when we were in person in offices and we presented, I'm 6'4", meter 93. I'm st- standing next to a screen about yea big. I'm about the size of the screen. Like, the ratio of Matthew to screen is, like, more reasonable. I'm gesticulating with my big hands. You know, I'm Italian-American-ing the whole thing.

48:34 In the webcam world of the pandemic, like right now, I can't even see your faces.

48:41 My content is huge. I don't even know if you can see my face when I'm presenting.

48:45 And I'll tell you, my, my facial expressions are a big part. The gesticulations are a big part of the communication.

48:52 So that's the problem, is that we lose a huge percentage of how people receive information from you in the webcam world.

49:01 Further, when we tell a story, we don't want to overwhelm people with the total picture and obscure our faces with it. We want to be somewhere where our face is still visible, but we can indicate with it and build it gradually. So we inverted the whole thing to work like the way the rest of Tableau works, where you drag the data out and the vizs build themselves.

49:18 With somewhat… Scripted's not the word, 'cause you could go kind of in a nonlinear way, but you had to define these scenes ahead of time. You'd say, "This is the finished viz.

49:27 In this zone, there's gonna be a map. In this zone, there's gonna be a bar chart." And then you could build it in whatever order you wanted to tell your story.

49:35 And we did this at the Gartner, and we, we won in a landslide, something like 80% of the vote. We did this at Devs on Stage. My favorite demo ever, because we did exactly that, like, three beats sort of approach. The good, better, best. The ba, ba, ba sort of thing, where the first thing we did was the bar chart, and the second thing we did was a scatterplot, and we showed them linked.

49:53 That was kinda cool. And we showed this map that was linked. All right. And then the third beat was this globe.

49:59 And if you were there, you'll remember it. It's easily one of the coolest things I think we at Tableau have ever built. Uh, it was certainly my favorite demo ever to do. I knew it was coming, so I'm watching the audience, not even looking at the screen, when I got to reach up and spin that globe.

50:14 Just an amazing, really fun moment. Uh, gasps across the crowd.

50:18 And, and we made it, we made it really clear. Like, the problem is we lose connection to people. The solution is something that s- brings the data into the webcam. The achievement is now easier. Now better? Yes. Faster? No, more. So we anchored on better and more. Better communication to more ways of visual storytelling.

50:40 So hope that workshopping that through with you, uh, if you look at that video and see where they started. Again, no shade to Research. They built something amazing, and it's when we applied those principles of visual communication and storytelling that we took it from, you know, a seven to an 11.

50:57 All right.

50:58 And with that, we're coming in. So I'll give you just a bonus as, as we close and we'll have time for Q&A. I can stay over if, Chris, if necessary, uh, but I did, I promised I'd wrap up at about an hour. Um, I always say three points and a poem, right? I was raised in a, uh, a faith environment. Well, that was basically every Sunday. You could almost sort of predict it by the rhythm.

51:16 22-minute-long message by the pastor, and you could kinda go, "Three points. Ooh, he's coming up on the poem here." Three points and a poem.

51:23 Uh, the poem today is Embrace the Change. You got our three points about hooking the audience, building trust, inspiring them. Your bonus point here is embrace the change. Uh, theirs is a time of immense change.

51:35 Change technologically.

51:37 People are, uh, changing the way that we work. The old things don't work anymore. We gotta embrace new things. The things that wowed people 10 years ago don't wow them anymore, so embracing the change. And that gives me the chance to share with you perhaps my favorite photo I've ever taken of one of my kids. Uh, years ago, th- that little girl now is a high school track star, but she was tiny then, and we went to this place in the Netherlands called Pony Park City, where everybody spends a weekend in…

52:04 Well, I think the name kind of explains it, doesn't it?

52:07 Trapped in this, like, weird town where, with ponies free-raming, free-roaming everywhere. A nightmare for parents, a delight for kids. And so Estelle got this pony. Was trying to get the pony to do something. I think you can tell from the energy here how the pony felt about it.

52:22 So as we look at this world of change, changing the way we've done things before, maybe changing, maybe you're coming to this, and maybe, uh, you're, maybe you're Estelle, my daughter here on the right, and your company's data culture is the pony. In which case, like, we're here to support you with things like, uh, um, the Tableau, uh, blueprint methodology and others to help you change that culture.

52:45 Maybe you're the pony, and this, new ways of doing things. Maybe you're like, "I built the chart. It should be good enough. I built the dashboard. Why don't they use it?" And maybe over here, Estelle is, uh, MCP and, and AI and all new ways of thinking, right?

53:00 Uh, regardless of whether you're Estelle or whether you're the pony, em- my encouragement to you today is embrace the change. Lean in.

53:07 Continue to hone your craft, not just for data, but for showing it and for telling a story with impact. So thank you all for the chance to spend time with you, and, uh, we'll open it up for questions. I can stay on for probably another half an hour at most.

53:23 Chris McClellan: There was a lot of comments as we went through. There's not a lot of questions.

53:28 There's some-

53:28 Matthew Miller: Do you wanna summarize if there are any questions there, or do you wanna raise the, uh, raise… We can do raised hands. That's fine.

53:34 I don't know if you want people to come off mute, Chris, or how you run it here, but we can start with, uh, raised hands if you want.

53:39 Chris McClellan: Oh, and, and the same

53:39 Matthew Miller: Any questions

53:40 Chris McClellan: If, everyone can come off mute when they need to and, and say something as well.

53:44 Matthew Miller: Fi, my question's for you. How bad, on a scale of one to 10, was my half-Australian, half-Kiwi accent?

53:51 Fi Crocker: Uh, yeah, the I think Pulse did a good response. Maybe your accent could use some work.

53:58 Matthew Miller: Thank you. You're very kind and diplomatic, and you'll go a long way in the world of storytelling with that kind of diplomacy.

54:04 Fi Crocker: Doesn't happen often, let me tell you.

54:07 Matthew Miller: Thanks. I'm gonna bring the, uh, I'll bring the q- comments over here so I can look right into the questions.

54:11 Fi Crocker: I, I do have a question, just while, um, we're, we're waiting for everything to come up. I think, um, you've been blessed in having some amazing leaders throughout your career, and we heard you talk about Francois and maybe, uh, that he gave you a bit of a shock during a couple of meetings. Uh, what are the things that you wish that your team or people that you've come across could also learn from these amazing, inspirational people that you believe will help to take them forward within their careers?

54:54 Matthew Miller: The thing I've been learning recently, I credit my wife for this, Becky, who's, uh, r- really brilliant. She's an expert on emotions. And our worlds have kind of collided because emotions are about our brain's predictions of what will happen, and, and AI is about predictions of what text it thinks we want to hear. So our worlds are colliding now.

55:14 The thing that she's been showing me is, is that people are, are fundamentally compassionate and curious. If you've, uh, had any exposure to the therapy modality called internal family systems, they talk about compassion, curiosity, et cetera.

55:28 And, and so when I run into someone who I'm trying to influence, and influence is basically everything I do. Trying to get somebody to spend money, trying to get somebody to, you know, do something, change a decision. Trying to get… Trying to not have it called a m- Trying to talk people out of calling it Tableau Plus when we have a Tableau Pulse.

55:45 Led to one of our French customers saying, "You have Tableau Plus and Tableau Pulse, and I cannot tell them apart." And, you know.

55:51 Right? I f- I lost that battle. I lost that battle. But it's all about influence.

55:56 Very often I run into people who have a different view from me, and I have to convince them.

56:03 And when I approach it from the perspective of, "You're wrong.

56:07 You're wrong and you're bad. You're wrong and you're bad because you have bad ideas or a different value system." And they can s- even if I don't say that out loud, they can sense it in my demeanor, right? That I'm, uh, that I'm impatient. Ugh, get out of the way. Right? Young Matthew was… Oh, just ask people I worked with at InterWorks.

56:27 Young Matthew was a hothead. I'm still a little bit of a hothead. And so I would go at it like, ugh, so and so's blocking the thing I need.

56:34 I'll set up a meeting with them, right, and I'll plow through their objections. And I know, like I've been told, like you, you know, catch more flies with honey than with vinegar. I know all this, but I still would just go at it, like this combative view that presumed they were wrong, they were bad.

56:50 Instead of saying, "All right, we have very different perspectives on this." But I want to be compassionate, I want to be curious. I want to assume you have good intent. I want to assume that you're standing on the same side as me, looking at it and going, "Yep, our baby is ugly," but you have a very different opinion of what we should do about it.

57:06 And so let me understand why. Let me ask, and, and really inquire and understand it in the process. I had to do this today on a call, by the way, uh, where a colleague and I disagreed about something, and we both got a little worked up initially.

57:21 And I just stepped, stepped back and said, "Listen, I have enormous respect for you. The reason I wanted you on this project is because I've seen your previous work, and I…" And this was not lying. This was not manipulation, by the way. I've, this was all 100% true, and this person could sense that I was very genuine. I said, "Now, we clearly see this very differently right now.

57:39 I want to figure out why that is." And we managed to find our way through it. It was really good. It meant we had to go half an hour longer than the meeting was planned for, but we got there. And, and so the thing I'm learning now, at 45 and older, is it takes a lot more time to do that. Like if you have the authority to just, you know, hammer down, "You'll do it my way or the highway," it feels more effective.

58:03 But in, in, it's tremendously shortsighted, and it misses out on the humanity of the work that we do. So I could preach about that for another hour. That's been a huge sea shift for me.

58:14 Fi Crocker: Yeah, I love that. It really resonates with me. I was listening to someone else yesterday actually, and they said exactly the same thing. Assume that people come with good intent.

58:24 And, and it's, it's just as easy to assume that they come with good intent as assuming they come with bad intent, but it's just a mindset shift.

58:31 Matthew Miller: Yeah. It's been a big change. Thank you for the question.

58:34 Anybody else? Throw a hand up, or come off mute.

58:39 I'm gonna thumb through to see if there's anybody with cameras on that I know.

58:43 I see Andy Holt. Hi, Andy.

58:46 Anybody else?

58:47 Fi Crocker: I do have more questions, but, like I, I can wait. Um-

58:51 Matthew Miller: Go ahead Fi.

58:53 Fi Crocker: Uh, so there's a lot of change at the m- Like, things are swiftly changing at Tableau/Salesforce, um, at the moment, and obviously we saw a lot that came out at, at Dreamforce as well. Um, M- Mark's come along this year.

59:10 Uh, w- what's it like at the moment being caught in amongst that change? And do you think that there's the potential of risk where you're getting things are, people are going off in many different directions at once? How will it come together and have a cohesive roadmap?

59:33 Matthew Miller: Uh, we, we have the luxury challenge at Tableau of having a passionate and experienced power user base, a group of practitioners. I've compared us a lot to, uh, Photoshop, because if you were at Adobe and you took over being a product leader for Photoshop, you could change one keyboard shortcut and have millions of angry photographers who've been like, "For 20 years it's been Control Alt Z F4, stand on one leg and say this special word," and it did the thing, and you broke it, right?

1:00:08 So we gotta respect that. But also, we have a whole new, uh, wave of users who don't know the keyboard shortcuts, and don't even understand iron biz, and are just like, "Why is this so complicated? Do I have to drag and drop everything?" Uh, yeah, my, my kids are used to multi-touch devices by default, and they're used to, they use Gemini at school with pr- school permission.

1:00:31 They're, they have an expectation. It's just very different. So, uh, how we bring the brilliance of Tableau forward while leaving what every product has, which is some accumulated baggage, is the biggest challenge. Uh, and now that I can talk about it, which I couldn't do, uh, when I presented on my birthday, um, this same talk to the Austin Group, uh, I can now talk about Tableau Studio, which is, uh, while I'm tremendously proud of what we've done.

1:00:59 Uh, in that we've said, "Look, take a step back." Uh, agents need tools.

1:01:05 We, we got a 3,700-page manual. I said I pulled up the PDF, but the, the Tableau Desktop user manual, which I've not looked at in a while, it's got a little longer. It's 3,700 pages.

1:01:14 Chris McClellan: Right.

1:01:14 Matthew Miller: Nobody, not even Anya A'Hearn or Jonathan Drummey knows every page. Agents need tools.

1:01:20 What if we took a step back and we said, "What if we could just give agents that tool surface, they could do anything, and give people still that full control?" So we, we said, "Okay, our product principle is gonna be no feature left behind. You want custom background images? We still got you. You want a weird custom TMS file?" I'm gonna try to pick some random stuff that OGs will know.

1:01:39 Uh, what's the thing we do most people don't know? You want raw SQL? Oh, yeah. We got raw SQL. We got explain data. Like, all the, all the s- all… No feature left behind. Not to say later we may finally kill off some features that nobody actually uses, but for now, we're gonna go from a s- strategy of let's serve all the power users, but just make it way, way easier by enabling agents to do it.

1:01:59 Um, you know, time will tell if we got it right, and I'm sure you all will tell us if we didn't. But, uh, I, I think that that has resonated really well with people. Uh, the hard part has been making sure the agents find their way through that 3,700-page manual to the right thing. I think, Andy, you've… I've written about this, and you've seen it.

1:02:16 Like, that's been hard, but, but we're getting there little bit by little bit. Um, I actually neglected to show this earlier, so I will. But you saw the pine tree story, and the, the real, like, juice of that, and the reason I show that story is it's about how Tableau enables you to build once, like, a basically a data product, and then use it a million times, right?

1:02:36 You can use it in all different ways and places. Yes, you can drag and drop, but you can do AI, you can do Pulse, or you can do MCP. And so one good data product like my IoT data set here that I have as a published data source on Tableau Cloud, I can use it in many experiences, but it's one source of truth. So that's the whole point there, but I was using it with Agent and, uh, showing this in a keynote, and just because I like to live dangerously, I said, "And by the way, with the new agent, you know, you, you hear about, uh, uh, vibe coding.

1:03:05 Well, I'm gonna vibe code a new chart I didn't have before," and I asked it to make me, like, a truly chart junk viz.

1:03:13 I said, "I want a pine tree, an actual pine tree that's smiling when its humidity is good and frowning when it's bad," and it built me this. Look at this.

1:03:23 Isn't this the dumbest chart ever? Like, but, but the fact is I can take my stupid pine tree mood and, like, slap it in the middle of a dashboard now, and it works. So, like, this is, this is, I think, bringing together the best of the vibe coded world, where like, look, I was able to come up with a random chart, and I'm over here going, honestly, the chart was never the hard part, right?

1:03:45 The, the viz was never the hard part. The, uh, the dashboard was never the hard part of Tableau. It was always actually getting the data in the right shape. But if you want, if what you really want is a, a pointless chart junk viz of a pine tree frowning at you or smiling, well, we can do that too.

1:04:03 Um, that's one of the things we… I think it's been kind of fun, and people love this.

1:04:06 So anyway, there's the, uh… That's one thing I, I neglected to show when I was showing the pine tree is, is like we, we're, we're can bring back the feeling that people had when they first used Tableau, what it did versus what the previous world before Tableau. Like, the freedom I had to just whip these vizs together, it felt like magic the first time I used Tableau.

1:04:27 Well, now, like, agents feel like a similar kind of magic. Like, it'll do anything, but it can do a lot of dumb things. It doesn't understand my data. So, oh, can we bring that kind of flexibility of, like, you can do anything, but bring that into the trusted Tableau context?

1:04:41 Um, I think that's k- that's pretty killer. So, um, I don't know if the marketing around this has been super clear. I don't know how much it's shown through in the Dreamforce demo. You will see a heck of a lot more of this at TC.

1:04:51 But basically, I've, I've been calling this, like, agentic, uh, analytical vibe coding that scales. Like, it's a trusted platform we all know and love. You still got the reins. If you wanna drag a drop by hand, you can, but you just want the agent to, like, you know, create a cr- complete custom UI over the trusted data, it's got you.

1:05:07 So that's been really fun.

1:05:10 Chris McClellan: And that's a-

1:05:10 Matthew Miller: Totally

1:05:10 Chris McClellan: … hundred, right? That's complete chart junk, but-

1:05:13 Matthew Miller: Oh, yeah. This is, this is-

1:05:14 Chris McClellan: … it, it-

1:05:14 Matthew Miller: … awesome

1:05:14 Chris McClellan: … but it answers the question brilliantly, right?

1:05:18 Matthew Miller: Yeah. It's, it's-

1:05:18 Chris McClellan: So it's just smooth to, okay, what does a tree feel like? Bang, got it.

1:05:23 Matthew Miller: Right. And, uh, you know, I love it. I, I, like, it's… I, I know, I know that I'm sure, um, Tufte is like, you know grimacing in, in his heart right now, but, but it makes me smile or frown.

1:05:36 Chris McClellan: And then it just puts more power back onto the analyst to say, what should you build? Because when you can build anything and everything, um, where are those limits and guardrails?

1:05:47 Matthew Miller: Yeah. Well, I think we're still figuring that out, right? But the-

1:05:49 Chris McClellan: Yeah

1:05:49 Matthew Miller: … at least we'll be able to say, "Look, you want the YOLO whatever you want, crazy viz spinning globe thing." I think we're gonna see an explosion of chart junk, but then again, we kind of saw that in the early days of Tableau. Do you remember the… And this is not sh- throwing any shade at Noah. It was impressive. But do you remember the, um, Mandelbrot viz?

1:06:06 It's probably on public still.

1:06:09 Um, just while we're, you know, OG Tableau things.

1:06:12 Chris McClellan: I'm gonna say the thing was… Yeah.

1:06:13 Matthew Miller: Uh, Noah Salvaterra, just one of these things you look at and you're like, "Why did you build that in Tableau?" 'Cause you could. It was super cool.

1:06:22 Um, not, not… Doesn't, didn't tell a story in any particular way. Um, uh, I think it was on Drawing with Numbers at the time. I think it's on Tableau Public. If I can find it, I will. Um, he was a Zen Master, like, a long time ago.

1:06:36 Um, this blog post is not that, but it's, um, it's… It talks about it, and I think it links to it.

1:06:42 There you go.

1:06:43 Just, you know, again, kind of chart junk, but, like, it was cool. And I think what we've seen is, you know, for the most part, this is where Pulse came. The Pulse mentality was, look, Tableau let you do anything, and we did an awful lot of anything, didn't we? But most of the time, a bar chart and a line chart is what you need. So for the 80/20, why don't we just do that automatically?

1:07:01 Uh, which is a different mentality from I build every viz by hand, but I think that's okay. But I think what's gonna happen is that the people who are like, "Look at this weird custom dashboard I made in, in, uh, in Claude. I can, like, spin a globe and see all my charts flowing out of it," and I think before long they're gonna be like, "Can I just see the number?" Like on my phone.

1:07:20 Like please. So, yeah, I'm, I'm… I c- I think there's gonna be this like people get it out of their system again, just like we did 15 years ago, and we're back to building sensible things that brands can… Um, while I'm ranting on this, when we did the Apple Vision Pro project, uh, we, we did a ton of exploration of what it enabled 'cause, 'cause the Vision Pro is an augmented reality device, we did a ton of exploration of what it enabled you to do in terms of three-dimensional visualizations.

1:07:45 There's been a lot of work done there, and Tableau Research went nuts and built all these cool things. And then when we actually talked to customers, you know what they said?

1:07:51 They're like, "Can we just get like a number?" Like the nu- the… One example was, I wanna walk through a building, and when I look at like the elevator, I wanna see a metric about days of failure, or days since maintenance, et cetera. So actually, when, in, when we got into like the VR/AR space, what people wanted was OG simple one-dimensional or at most two-dimensional vizs in that space.

1:08:12 They actually didn't want crazy new viz types that were harder to understand, they wanted simple to understand things that were just right at their fingertips.

1:08:20 Which is a, was a big aha for me after we did that, uh, that project. So we went back to like, uh… Unfortunately, the whole thing got shelved. The Vision Pro didn't sell quite as much as everybody wanted. I was real excited about it, but I, I was in charge of it at the time, and it got defunded, which made sense. I really liked it though. But I had this vision for like walking around my house chucking pulse metrics onto the wall.

1:08:41 And they do that now. They allow you to pin things in the space of the Vision Pro, so it like, uh, will just stick. So imagine like I walk outside, not that I walk outside wearing my like Apple Vision Pro, but when they come out with-… a lighter weight one, like I would literally look over at the tree and see the frowny thing going, "Water this stupid thing," right?

1:09:00 Like, that's the Holy Grail. I don't actually most of the time care about it unless I… You know, if I'm walking by though, I wanna see it.

1:09:07 Um, so you know, you pull up to your house, you see your lawn is dry, you see exactly the trend. That's when you wanna see it. Walk through the factory floor and see the, the metrics. You, you know, people do that with like Tableau TV wall boards. You see those all the time at customers. In fact, we have a… This is something you may not know, we have a special SKU for like Tableau display licenses, because those people tend to hammer.

1:09:29 They've one- they'll use one user account, and they'll hammer us every five seconds in Tableau Cloud to the point we had to build a bunch of optimizations just for that use case. We, we charge, we monetize that separately. And so the, the display license it's called. And you'll see a wall board. Well, that's cool. But like what if it sort of followed you around and highlighted on each of the devices?

1:09:47 I thought it was just purely neat. So anyway. And so what we learned there, again, was that the fancy cool new stuff that people are doing, like, "Look at this 3D flipping race chart. I can spin it around." Like, you probably don't really need that, you know? If you ever looked at like a Formula One race car driver's like complex screens, there are lots of simple charts.

1:10:07 It's dense information, but simple.

1:10:10 Anyway. Grumble, grumble. That's my, my my cynical old man view now is that the s- the fancy stuff is less, far less useful most of the time.

1:10:19 Chris McClellan: But it's the same thing.

1:10:20 Matthew Miller: That's it.

1:10:20 Chris McClellan: I remember a software long before Tableau, and they had w- you know, the new feature was 3D bar charts.

1:10:27 And we could-

1:10:28 Matthew Miller: Yeah

1:10:28 Chris McClellan: … we used to teach clients how you could fly through, you know, fly through the bar chart and everything.

1:10:34 And, and just… I think it got ditched 'cause no one used it, because it's just not… You know, it's cool, but it's just not practical.

1:10:41 Matthew Miller: That's it. Lines, and bars, and BANs.

1:10:43 Chris McClellan: Yeah.

1:10:43 Matthew Miller: All right. I see a few people left on. Han, John, Melissa, Frasier, Anne, Sarah, I know. Hi, Sarah. Evelyn, Kate, Raul, Kenwal, Jacob, I know, and Ilya, I know.

1:10:54 Um, maybe you just walked away and left your laptop open, but if you're still here-… and you have a question, wanna say hi.

1:11:01 Just to, to… You know, it's a rare that I actually have like this much time to hang out, so it's really nice to just, just, uh, hang out. If there's anything I… Stories I can tell you, questions you've had about Tableau, uh, so long as it doesn't violate my employment contract, I got you. Hey, Jacob.

1:11:15 Jacob Rothemund: Finally say hello. Hello.

1:11:17 Matthew Miller: Hey, good to see you.

1:11:18 Jacob Rothemund: Hiding away in the office.

1:11:20 Sarah Burnett: Hey, Matthew. Um, love all your stories. Heard a few, a few of them before. And, um, I think one of my favorite ones though is your near disaster, I think, on Devs on Stage this year.

1:11:33 Um-

1:11:33 Matthew Miller: Oh my gosh.

1:11:35 Yeah.

1:11:36 I'm, I'm happy to tell that if you, for those who haven't heard it. I don't think it's been told like widely.

1:11:42 Oh my goodness.

1:11:44 Uh, in sh- you want me to tell that?

1:11:47 Sarah Burnett: Yeah.

1:11:47 Matthew Miller: Uh, in, in short, uh, uh, when I teach people about demos and presentations, craft and so on, we talk about the blast radius. Uh, the blast radius is like the, the risk of failure and the size of the audience. So in a situation like this, there's 100 people on. You know, I have fallbacks for things. I have my personal laptop ready in case something went down, but I don't have a backup demoer.

1:12:12 If I, you know, lost my voice or gotten sick, we would've just rescheduled, you know? And also, I could somewhat read the room, and so if I make a mistake, I can go back and do it again.

1:12:21 Also, I consider you a friendly audience, right?

1:12:24 But, uh, TC, Devs on Stage, you're in front of 8,000 to 10,000 people. There's 15,000 watching live or more. I think this year it was like 25,000, and it's gonna be recorded in 4K forever on the internet. So the blast radius is enormous. So what we do is we make sure that we have redundancy all the way through the stack. And I say this tongue in cheek, even though it's a little dark.

1:12:42 I say, like, "We've built such redundancy that I could have a heart attack and drop dead, my laptop could catch fire, right? An entire AWS region could go down, and they would drag my corpse off the stage, extinguish the laptop, and the show would go on because there's another moral who knows every word for word. There's another backup laptop. We have backups for everything." And, and I get flack for it every year.

1:13:07 They're like, "Do we really need two servers? Do we really need another Snowflake instance? Do we really need another… " Yes.

1:13:13 They're like, "An AWS region isn't gonna go down like it's happened." Uh, so we, we make sure we've got resiliency all the way through the stack.

1:13:20 So, it's time for Devs on Stage, and I'm going to be presenting Tableau Solve, which was another just one of my favorite demos to do. I love, love, love what the design team did. Shout out to Bailong Zhou and Esther Cheneaux, whom some of you know, the work they've done on that is just incredible.

1:13:37 We're, we're just completely inverting the paradigm, very Tableau style, where instead of I tweak the numbers in a table and then the chart updates, the chart is always read only. You can actually interact with the chart, and everything in the viz becomes a parameter for further analysis. Brilliant stuff. And I knew it was gonna be a wow moment, by the way, and as a demo-er, I just love those, so I was super pumped.

1:13:55 And a good percentage of the demo was AI stuff. It didn't start that way. It started with some sort of build the mental model, right? The problem, solution, achievement. I had to start with things we knew, and then there was a point where I was gonna go into the agent, and that was, like, better than 70% of the demo. But we had resiliency all the way through the stack, right?

1:14:12 We had someone else who knew my script and could walk out there and kill the demo. We had backup laptops, backup servers, backup everything, all the way through. And even for AI we did. So we had a direct connection to Anthropic, and we also had a connection through a Salesforce-hosted gateway that went a different, completely different path.

1:14:28 So we were covered.

1:14:30 Uh, Devs on Stage starts at 9:00. Uh, uh, the way the schedule worked then, I was gonna come out about halfway through.

1:14:36 And about 8:48, I'm running it through backstage just like over and over to get it up in my head and make sure it's thoroughly memorized, and Anthropic goes down.

1:14:45 Like, totally down. Claude's down globally.

1:14:48 So I'm like, "Ooh, that's not good. Okay." So I switch over to the Salesforce one, and a few minutes later it went down. Totally unrelated. One in a billion coincidence. Like, just… And the thing is, I had even, in some meeting prior to, I was like, "You know, just in case they both went down, we should probably have like a local server backstage running an LLM.

1:15:06 It would be slow, but it would work." And they all looked at me like I had eight heads. 'Cause I was like, "But you know, like that's a complete dependency and da, da, da, you know, Anthropic's uptime." No, we're not gonna do that. It's too much money and time, and the old man wasn't wrong.

1:15:21 The Salesforce gateway went down.

1:15:24 And I was like, "Oh no.

1:15:26 Oh no." I realized I had no safety net. Like, and I'd built net, after net, after net, and everything had fallen apart. And I was like, "Did, did… But this, this, this demo, you can't work around that." 'Cause most of it was the agent. Like, I knew I'd get through the first few words, I knew I'd get through the first drag and drop, and then I would go, "And now we turn to the agent." So I'm backstage, and the tech team is like, "Give me your phone.

1:15:46 You got to put your phone away. You can't go out there on your phone." Like, "I know what I'm doing." I'm… And I, I've got my phone, and I'm refreshing the status pages for both of them over and over, and they're both down, solid red. I, I was to the point of texting people I knew working at Anthropic, and s- literally going, "Please bring it back up." Um, I, I actually said, I have a screenshot of this, I said, "If you unplugged it, could you plug it back in?" Uh, um, and they so they're like, "We're doing all we can." Not that I'm that important, but I literally went as far as I could up the organization, you know, I…

1:16:15 Please tell me you're gonna fix this. If not, I don't know.

1:16:18 So I'm backstage.

1:16:20 Uh, Lauren and Sophia are doing a great job hosting, and the moment's gonna come where they're gonna invite me up. Tech team's like, "We have to have your phone." Like, "Do you really need it?" I refresh one last time, Anthropic's down. I check, the gateway's down. I'm like, "Oh, no.

1:16:33 I'm gonna bomb in front of our brand-new GM.

1:16:37 8,000 of my closest friends, 25,000 watching online, it'll be on YouTube for all eternity. Well, it's been a good run here." I handed my phone to the tech person, and they're like, "Matthew Marri, like ah." I walk inside, and you can see it in my face, like the color is drained. Like, this is it. This is the end. It's been a great 15 years. I was just…

1:16:55 I was terrified. And it's, it's a little of the… You know, I don't know, I don't think it was Francois who taught me this, but like the me- the image of like punching fear in the face when you're public speaking. You know, there's a, the, the pit in your stomach, and I, I try to do that. There's a, a meditation, like breathing routine I walk the folks through before we go on stage, and I had done that to myself, but my heart rate wasn't coming down.

1:17:15 I mean, oh boy. I walked out there, and uh, they're like, "All right." You know, there was like a little intro of the three demo-ers for Labs, and I was the first to go. So then Lauren and Sophia are like, "And now Matthew." And I walked over to that desk and put my hands on the desk, looked at it with all my friends and my new boss and all my peers.

1:17:34 And I was like, "Oh man, it's been such a… I really like working here. It's… I wonder, I don't want to go work at Microsoft. They give boring presentations." Just I was, I was like this, I was really, I was really torn.

1:17:46 Um, so, you know, I go, "Hello, TC." I, I, I've watched it a few times. You could see the terror. I was just like, "Oh, this is it. I'm toast. It's down. It's, it's gonna bork." So I, I do my little drag bit. And um, it, it's funny, I accidentally typed the wrong number. You may remember the number, like I typed the wrong… There's a typo out, and the bar jumped too high, and I made some quip, and I was like so nervous.

1:18:08 My hands are shaking. You lose like fine motor control.

1:18:12 Uh, I'm just fumbling with the keys. And then the moment came where I had to ask the agent the first question. I said, "You know, this is more than just a whole new way of… It's a, it's also an agent.

1:18:22 Or it might be, if any of them are up. Let's find out." Now again, the team back- backstage is so good. They've got multiple machines with the same demo on it, so they can swap between them, and you wouldn't know the difference. Um, I think it was last year, two years ago? A few years ago, we had one fail. We went to the backup, and thankfully the team was smart enough, they brought the primary backup, and then the backup failed, and they switched to the primary.

1:18:40 The person on stage had no idea. 'Cause on stage we don't… We just have a keyboard and mouse. Like it's like a KVM. So, so they, they will switch what machine you're demoing from, and you don't even know. It's brilliant. So I don't even know what device I'm on. I have no idea if any of them have come back up. I don't know what's going on.

1:18:56 And there's that moment where I ask it, "Hey, you think you can forecast my sales for me?" And I hit enter, and it says, "Thinking." And when it would fail, it would go, "Thinking," and that big ugly red like, you know, 401, you know, service is down, 500 error. It's thinking, and then it goes, "Oh, I can see a forecast. I've looked for this." It was…

1:19:17 The heart attack was over. And the rest of the demo I was just in euphoria, because it was, it was like, all right, everything, nothing, nothing can top this. Uh, but I think, I think in my entire career I have never been, uh, more terrified on a stage, because I knew- I knew what was coming, and it worked. I don't know which… I actually don't know i- which, whether it was Anthropic that came back up or the gateway that came back up.

1:19:38 The team would know.

1:19:40 Uh, but it, it, it worked. And, and what made that so much fun was the, that we were showing, like, how this, the Tableau team that came up with this had approached it and said, "Hey, what if we could take the primitives that we all know and just use them in new ways?" I just love that. It's, like, quintessentially Tableau. I'm ho- I'm honored that I got to be the one who was the mouthpiece for it, and I'm very grateful that the, uh, gateway came back up and we got to show it in its full, full color.

1:20:08 Anyway, thanks for prompting that story. It was a…

1:20:11 I will never forget that. Oh, my gosh.

1:20:13 Terrified.

1:20:15 Sarah Burnett: It could have been quite a different end to Devs on Stage if that hadn't all aligned for you.

1:20:21 Matthew Miller: Right. And a, and a reasonably storied career too. Like, I I mean, I c- I, I, I could, I could drive Uber with my minivan. I think I'd be a good Uber driver. I'd be the one that, like, talked too much.

1:20:32 All right.

1:20:32 Chris McClellan: Anyway.

1:20:32 Matthew Miller: Um.

1:20:33 Chris McClellan: That's, that's good. We've got a, we've got a few other things to do. I, I don't know if you were trying to wind up.

1:20:36 Matthew Miller: Go for it.

1:20:37 Chris McClellan: Okay.

1:20:37 Matthew Miller: We're totally fine. Thank you all so much.

1:20:39 Chris McClellan: That's cool. No problem.

1:20:39 Matthew Miller: Uh, great. And good to see you. Matthew, it was nice to see you again. And, um- Cheers. Thanks… great. You're a great storyteller.

1:20:46 Thanks for the chance to be here. Thank you.

1:20:48 Chris McClellan: Always. Um, if you don't know how to contact Matthew, here's all Matthew's contact details.

1:20:54 Pretty easy to find somewhere on the internet.

1:20:58 Um, I think everyone's done the survey. I will not start the clock. I'm just gonna go to the results, because I hope everyone's done that by now.

1:21:05 Uh, just to quickly go through the results. As to the questions for today, the standard questions, where do you live, so state and country. We can see where people are from around the world. Whole heap of different places, which is good. Um, I like someone's from America. Not the US or, or America, but anyway, that's good.

1:21:26 Uh, what industry do you work in? Just to prove that Tableau can be used across a wide range of industries as well.

1:21:33 And how many years have you been using Tableau, just to see that people have been using Tableau for a long time like Matthew, or a shorter time. That's okay. But there was once a time when no one knew Tableau, and everyone started off from scratch.

1:21:48 And the quiz questions for today.

1:21:50 If a view- if a user with a viewer license attempts to edit a published workbook on Tableau Server, what will happen?

1:21:58 I think the correct answer is they receive a permission error and can't edit because they're a viewer.

1:22:04 Um, which statement about Tableau's viz in tooltip feature is correct? It can display a s- a separate worksheet in, as a tooltip.

1:22:15 And what file extension is used for Tableau extracts created with the Hyper engine? It's the hyper file, because a TWB is a workbook, and a tds is a Tableau data source.

1:22:28 And when using a calculated field that uses the ZN function, or ZN for some people here, um, what does that do? It replaces nulls with zeros.

1:22:40 And is there one more?

1:22:42 In Tableau, which of the following statements about data extracts is true?

1:22:46 An extract can be refreshed on a schedule to keep the data up to date.

1:22:51 And what topic should be presented at future TUG meetings? It gives me some things to go on, so there's more Matthew Miller comments here.

1:22:59 And also, who would you like to present at a future TUG?

1:23:02 I thought there was Matthew Miller comments there as well. Um, basically Matthew is going to be presenting at a future TUG meeting in about a month's time for a Ask Me Anything on Tableau Studio.

1:23:17 But you'll see more about that within the, um, the TUG meetings as well. Um, so Anne just beat Andy by seven seconds.

1:23:26 And Jojo and IS and Fi, and it goes down from there.

1:23:32 It doesn't matter. Even if you didn't get the questions right, you've probably learnt something as well, which is the whole point of the thing.

1:23:40 Um, if you want to submit questions for those quizzes, you're quite welcome to. The QR code or use the link or just get in touch with me as well.

1:23:48 And our next TUG meeting is next week.

1:23:52 Fortnight away, sorry, not next week because I'm having a week off.

1:23:56 Um, is David presenting his first in a series of AI MCP all the new stuff for Tableau.

1:24:05 So he was showing this off at, um, the Sydney TUG a few weeks ago, and, um, he's made it into, like, a three-series thing to all go over October, November, and December, just to give more clarity on, and, um, presentations on that as well.

1:24:22 But everything we do, this recording, a whole heap of other TUGs, is all online at TeamDataFam. Uh, there's a Spotify podcast, there's an Apple podcast if you'd wanna listen rather than watch the things, depend on, on what you've got time for. Or just email me at teamdatafam@Gmail.com.

1:24:40 But thank you, Matthew, for your time. Uh, we're right on time as well, but, um, thank you for obviously that presentation, for all the Q&A at the end, and, uh, for the AMA that will happen in a few weeks' time as well.

1:24:53 Matthew Miller: Brilliant, thank you so much.

1:24:54 Great to see you all.

1:24:55 Chris McClellan: Awesome. Okay. Thank you for your time.

1:24:57 Sarah Burnett: Thanks, Matthew.

1:24:58 Chris McClellan: Hope to see you.

1:24:59 Matthew Miller: Cheers.

1:25:00 Thank you.

1:25:00 Sarah Burnett: Bye.

1:25:00 Chris McClellan: Bye.

1:25:01 Matthew Miller: Thank you.

1:25:01 Sarah Burnett: Thank you.