Austin Tableau User Group 10 Sep 2026

0:01 Celia Fryar: Hey, thank you, Jane, for kicking us off. You guys, welcome for… I'm so glad you guys are, have made it here the, on this, um, beautiful afternoon to help us celebrate Matthew's birthday. No, I'm just kidding.

0:13 But it actually is his birthday, and I'm, like, really excited for this program today. If you would kindly, like, pull up your chat, tell us where you're dialing in from. We'd like to… This is always fun to see, uh, where folks are connecting from. I love that about the virtual world. So, I'm Celia Fryar. I'm gonna just kind of run us through what we're gonna do today a little bit.

0:33 Thank you for that. That's so good, um, in the channel there. Hope you saw that, Matthew. That was really creative. Was that you, Julia?

0:40 I think probably so. But in any case, um, we have a really rock solid agenda for today. We're gonna…

0:48 Little bit of news and announcements, and then we're gonna hand the baton to Cameron Brown. She is, uh, part of our, our advisory team here in Austin, but she also works for Tableau, and she… We'll, we'll introduce her in just a minute. But we've implemented this new segment this fall about stepping out of the comfort zone. What are we doing with AI?

1:05 What are the practical things? What are the elaborate things? And, like, we're gonna take a little turn to a lot of different types of things that we could do with AI. And then, um, Matthew Miller's here to talk to us about data demo, uh, demos and storytelling. In, in this age of AI, the value of that is, especially the storytelling part, is really, uh, uh, there's a premium there.

1:27 So with that, a l- little bit of announcements.

1:31 Wanna start with the… This is our advisory board here, and we, um, really enjoy working together as a team.

1:38 Glad for that to be, um, the way we work here in Austin, and just encourage you that if you have some fresh ideas, wanna jump in and be part of the team, you know, reach out to one of us. We would be, uh, grateful to have, uh, new folks, new ideas, and all that.

1:52 Um, this meeting is clearly virtual TUG. We do get together a couple times a year in person, and that's always a lot of fun. We did that late July and had a big data biz competition. It was so much fun.

2:03 And, um, but here we're gonna be using chat, so let's please do chat. Let's talk and ask questions and keep things c- lively in the chat. Uh, the recording will be available in a few days, um, thanks to Chris McClellan. He does a fabulous job of turning that around for everybody in the DataFam. And then it'll be posted on a couple different places where you can access it.

2:25 Speaking of conversations, if you have not joined the Tableau community channel yet, please do. You're missing out. That's the place where a lot of information gets passed. You learn about events. You can comment on and give feedback on other people's visualizations. You can give feedback on your own visualizations. You can learn about betas, be involved in betas.

2:45 Uh, there's a lot going on there. So I did update this number last night. It's just shy of 24,000 people in that channel. So, um, you can use the QR code. That will get you a link to go, um, to sign up. And if you miss it, I'll be happy to put it in the channel in just a few minutes as well.

3:03 So, um, what's on deck for us this fall?

3:07 We have Ann Pregler coming to talk to us in October about her Iron Viz experience, and then Andy Cotgreave, formerly the Iron Viz host and also technical evangelist for Tableau, well known in the Tableau community, will be our h- our guest speaker in November.

3:22 And then we'll ha- we're gonna… I think we're gonna aim at a regional Christmas party, y'all. It's gonna be a good holiday party in December. Um, the Makeover Monday team and the retail team, we were, like, looking at a South, South Texas kind of get together, may try to recruit some folks out of the Dallas area as well. So, um, but that, look forward to that, and we'll have lots of news and, and get that word to you sooner than later, so.

3:45 And then keep an eye to this new segment that we're trying out this fall. Is, and we did interview techniques this spring. This fall, the, the consensus was we wanna talk about practically what we're doing with, with AI, and that's where we're gonna start. So, um, Cameron Brown from at Tableau, she works on the enterprise accounts ex- e- team, and she is, um, focused on public sector, especially K12, and has a heart for people doing things practically and really solves a lot of challenging problems that come with being in the K12 public sector.

4:20 I have great adm- admiration for what Cameron does. So with that, my friend, I'm gonna pass the baton to you and, and have you share from your side. All right?

4:29 Cameron Brown: Perfect. Thanks so much, Celia.

4:31 Celia Fryar: Welcome.

4:31 Cameron Brown: All right.

4:33 I love that you added the robot to the peer as well.

4:36 Celia Fryar: I couldn't, I couldn't resist, you know. It's just too cute. Did a great job.

4:40 Cameron Brown: Awesome. Well, hi, everyone. I'm so excited to chat with you guys today for the community portion of the user group.

4:47 Actually, first I do wanna pause because I want to make sure you all know what this portion of the TUG is about. So Celia, I know we said no pop quizzes, but to put you on the spot, do you wanna talk a little bit about what this component of the TUG is?

5:01 Celia Fryar: Sure. We have rotated through different types of, um, sections. I've learned this from some of the other Tableau user group leaders, and this particular one was, um, voted in by a lot of folks who are participating over the summertime. And it's just meant to be practical ways that we're using AI. So there's a lot of meetings that we are hosting these days about MCP and REST APIs and how we're gonna directly address the data.

5:28 Well, I think that the getting started part of that and making it more successful is like, what else can we do in our re- regular worlds with it? So a- as I've, um, interviewed a few people about, like, jumping in and participating, that's what I say to them. So with that, um, Cameron, does that set the stage a little bit for you?

5:46 Cameron Brown: It does. It does.

5:47 Celia Fryar: Okay. Good deal.

5:47 Cameron Brown: And I just, I just wanted to encourage the audience while I'm speaking today, this could easily be one of you guys in the future. So please engage. Don't be scared to raise your hand. Um, but a little bit about myself. As Celia mentioned, I'm currently a Tableau specialist here at Salesforce. I have been supporting Tableau for the last five years in a few different ways, but primarily have supported what we classify as highly regulated industries here at Salesforce, meaning that I work with folks who deal with very sensitive data, though I am pretty sure everyone can argue that all data is sensitive at this point.

6:22 But I have seen everything from a Tableau dashboard mock-up of a fighter jet depicting what parts need replacing, all the way to working with cancer research programs, and now currently support our K12 community members across the East Coast, where they leverage Tableau to ensure student success for this next generation, which is pretty neat.

6:42 I have also been supporting this amazing advisory council for the Austin Tableau User Group for a little over a year now, and they are just an amazing group of people. I don't have enough good things to say about them. Or, they're fabulous, and they put so much time and energy into this TUG. So if you have not connected with them already, please make sure to reach out.

7:03 Uh, as Celia mentioned, we've got four events per year. And for the folks that I recognize on the call today, thank you guys for joining us again. And for the folks who maybe this is your first time attending, welcome to the fun.

7:17 So for the community spotlight today, I'll be walking you all through my approach to leveraging AI to be more efficient. I'm a firm believer that AI is not here to replace me or you, but truly to make our jobs easier and be more effective.

7:32 So that said, of course, I was initially hesitant of how AI could be used, um, but I do feel really fortunate to work at a company that gives me exposure to innovation, because we get to experience a little bit of the future in real time, which is very neat. So as part of that, I wanted to share a little as to what I've learned and how I've leaned into AI in the last year or so, and how it's helping me in my day-to-day to not replace the things that I already did, but to do more with the tools at hand.

8:03 Before I talk a little on that, I do want to start with a story on what made the importance of the human element finally click for me.

8:13 Now, when ChatGPT first launched, I remember I did what most people do, and I took a photo of my fridge and my pantry, and I asked the AI how I could make the best and healthiest chocolate chip cookies with the ingredients at my disposal.

8:28 Sure enough, it gave me a response back, first acknowledging how impressed it was by my selection and recognized the work I evidently put into my grocery shopping. But second, it then built out a recipe that conceptually looked pretty promising.

8:44 I'm sure you all know where this is probably going, but after baking my cookies, I learned that I did not in fact, make the best and healthiest chocolate chip cookies. They were very gnarly. Um, borderline radioactive. I had to throw them away, take them out to the dumpster. I really didn't know that sugar could taste that bad. But it then became my recipe paradox, and I'm going to ask you guys to be very creative and apply this logic to the business sense.

9:17 Conceptually, the AI generated a very thorough recipe that made sense to me as a baker, but practically, the execution showed the flaws.

9:29 The cookies needed a taste tester. When I build a deck for my customers, each one is different. Each one is tailored to what I know about them and what I know will resonate with them. Some people like details, some people just like me to share my screen and demo.

9:45 My relationships and my human context helps me know what's best.

9:51 AI can be a great accelerator, but it is surely not a replacement.

9:58 So I want to move into a little bit as to how it accelerates my day.

10:03 First, it is with partnership and metric monitoring.

10:07 My customers' problems are my problems, so leveraging our technology with Tableau, I actually get a Pulse digest of case submissions from my customers, so I can ensure I'm aware of any problems with their deployment, um, and I can be a best, uh, partner to them.

10:28 Now, if everyone is not aware of what Pulse is, I believe Matthew might be showing a little bit here in his portion, so stay tuned.

10:36 Um, secondly, though, in my day-to-day, I leverage AI for trend visibility. Um, as I mentioned, I do support our K12 community, and it's across twenty-five states, and their business needs are directly impacted by legislation changes, and AI has helped me collect the information and synthesize it into something that's digestible and that I can articulate easily back to my customers.

11:01 But also, technology, education are not my only interests, so I also use AI to help m- maintain personal development.

11:10 I used it this spring to help arrange my plants in my planter so that the ones that need more water are all next to each other. There are so many examples of how you can leverage AI to accelerate your day-to-day. Now, this is a Tableau user group, so I think it's always important to know that offloading tasks to save time means that you can repurpose that time.

11:35 Now, does repurposing your time mean testing the effectiveness of the dashboards you guys have built to your end users, or is it adjusting the dashboards to be more accessible in line with federal mandates? There are so many ways that you guys can use it to your advantage. So as I wrap up here, I do want to encourage you guys to think of ways you can leverage it in your own life, whether that is personal, whether that is professional.

12:03 And if the starting point is with Tableau, you have a lot of folks willing to share their stories that are on this call today. But also feel free to connect with me on LinkedIn. I love to grow the community, and I'm also happy to get you a resource from our broad- broader internal team. Um, so with that, I do appreciate the time today. I hope it was somewhat inspiring, but also excited for the second half of the call with Matthew's presentation.

12:28 So I'll pause and, uh, happy to answer any questions. Uh, Celia, did I miss any in the chat?

12:34 Celia Fryar: Nope. The, we had some comments about things that people had tried. Um, sounds like Kevin had a success with a, a dinner, and, uh, Louis used it to install a sink generator, uh, aerator.

12:47 Cameron Brown: Oh, wow.

12:47 Celia Fryar: I guess.

12:48 Cameron Brown: Okay.

12:48 Celia Fryar: Yeah. Right on. That's-

12:50 Cameron Brown: Oh, I, I love-

12:51 Celia Fryar: … super helpful

12:51 Cameron Brown: … I love the inspiration. Yes, Matthew, um, I do not encourage you to use AI to give you a birthday cake recipe. Uh, steer clear of that. But thank you again for the time.

13:03 Celia Fryar: Um, yeah.

13:03 Cameron Brown: And hopefully I'll see one of you guys, uh, presenting for the community segment for our next TUG.

13:10 Celia Fryar: Yeah, please do take us seriously. I would, would very much like to have folks, uh, volunteer and, but I'm also up for recruiting, so let's do that. We're grateful that on, seriously, on his birthday that Matthew would come and share with us today about, um, a- and, and just a little context, uh, I didn't meet Matthew until TC this year, but I have been someone that I, I came across his LinkedIn publish- publishing that he'd done, and just was impressed with the storytelling and the communication techniques that he uses, and was very much challenged to do better in my own realm as an instructor and trainer, um, at listening to what, how he presented and his use of language and how he'd, like, set the stage for unpacking things for folks.

13:57 So if you get a chance, check out his LinkedIn profile. I'm sure that there's a few of them. I think I linked to one of them when I was sending out messages on LinkedIn myself. Um, so if that, that's incredible strength that he has. And then to know- to then meet him at D- TTC this year and learned that he was the man behind the, all that with Devs on Stage, which has been such a favorite of our, everyone in my realm anyway.

14:20 That was just amazing. So Matthew, thank you for being here today. I'm gonna pass the virtual baton to you and ask you to take it away.

14:30 Matthew Miller: Thank you. It's great to be here.

14:32 Celia Fryar: And Louis, yeah, Louis is actually recommending that we sing Happy Birthday. I think I have heard him sing before. That might actually work, but let me save that for later though.

14:41 Matthew Miller: There you go. Um, I'm 15 years old, and I, at an all-boys Catholic high school, the kind of place where everyone wears a jacket and tie every day.

14:52 And I was an aspiring leader, but not much of a communicator, so I thought I would run for student government, and that meant there were positions like student junior class president and junior class vice president and junior class secretary. In retrospect, none of those roles meant anything at all. We were powerless. If we wanted something and the faculty didn't want to give it to us, they didn't.

15:15 But gosh, when I was 15 years old, student government seemed like such a great opportunity to practice leadership. I didn't think I could be president, but I thought I could be vice president.

15:25 So I put my name in and I put up posters. My mom helped me make them.

15:29 And the day came of the vote, and ahead of the vote everyone had to address the entire student body and give a speech.

15:37 So it was my turn.

15:39 And picture this. I'm in a high school gymnasium, your kind of classic gymnasium, bleachers on both sides. They pulled out all the bleachers and they put my whole class along one wall. Right in the middle of the basketball court, right at the halfway point, they put this one lone microphone stand and microphone.

15:59 And the speeches were all terrible, but I was ready. My mom had helped me present and prepare. She's a great speaker herself, and my dad's a great speaker and storyteller. I was ready. I had note cards in hand, little index cards. Each one had my points and subpoints on it, and I had rehearsed around the piano the night before and I felt ready to win over this audience and win the junior class student vice president role.

16:25 So I step up to the microphone.

16:28 I'm from Rhode Island, probably the only Rhode Islander you'll ever meet, and most Rhode Islanders are short, and I was 6'3", 6'4". The microphone was too low, so I start kind of squeaking, "Hi, everyone. I'm Matthew Miller. I'm a junior." "Louder. We can't hear you." Now I'm thrown off, and now I'm fumbling with the microphone stand, but I was unfamiliar with its operation.

16:50 People are starting to laugh. It, it was already falling apart, but it kind of went from bad to worse when Mrs. Cross, the algebra teacher, had to come all the way across the basketball court and help me adjust the microphone stand.

17:04 I was so embarrassed.

17:06 I stumbled through my points. "If you vote for me, I will, uh, lobby the administration to get us more drinks in the cafeteria, and sodas that we like." The microphone stand was like this wide, but if I could have turned sideways and hidden behind it, I would have.

17:25 I was terrified. My knees were shaking. It was one of the worst speeches ever. It was awful. I wouldn't have voted for myself.

17:34 Eventually, I got through it and I sat down.

17:38 The next kid, a friend of mine named Peter with curly hair, sat behind me in English class, got up, and, and I can only say he stalked to the microphone. He walked to the center of that room, of that gymnasium, like he owned the place.

17:54 He walked up. I have a microphone somewhere. I should have grabbed it to illustrate. But he walked in. Here, I'll use a flashlight.

18:00 And he grabbed the microphone out of the stand like this, and he coiled it, the cable, around his arm, and he broke the fourth wall. He stepped in front of the microphone stand, the invisible wall I'd been trying to hide behind.

18:14 And he stepped forward to where we were all sitting in the bleachers. He wasn't afraid of us.

18:21 He, he was in control of that room. Instantly, everyone who had been awkwardly kind of fumbling, this is before we all had cell phones, so there were no cell phones then, but we were fumbling with whatever we had, just bored at each progressively worse speech. Everyone was paying attention. I was hooked. He stepped forward. And the thing is, I can't tell you everything he said.

18:42 I don't even know what his positions were. Did he have a stance on Mountain Dew in the cafeteria? I did. It was well-reasoned, and I had a plan to convince the administration to let us have Mountain Dew.

18:54 He stepped forward and said, "We're the best junior class ever. Who's the best junior class ever?" "We are!" "Yeah, it's gonna be great. We're having a great year. I'm gonna be your vice president, and it's gonna be great. So, you're gonna vote for me. When I say Peter, you say Nonas. Peter…" And he held up the microphone. It had a little of that energy of, like, if you've seen the play Peter Pan, where they go, "Clap.

19:16 Will you clap?" And the audience is supposed to clap.

19:18 Apparently, J.M. Barrie said that when Peter Pan was first performed, they were afraid people might not clap, and they did clap. Well, he says… I mean, what a, what a terrifying thing. What if people just said, "Boo!" But he said, "When I say Peter, you say Nonas. Peter." Everyone, "Nonas." "Peter." "Nonas." "Who you gonna vote for?" "Peter." "Nonas." "Peter." "Nonas." "Yeah, vote for me.

19:41 Who's the best class? Yeah." Put his mic back in the microphone stand, stalked back to his seat, and sat down.

19:47 Standing ovation, everybody cheering.

19:51 I turned 45 today.

19:53 That was 30 years ago.

19:55 I haven't talked to Peter Nonas in 29 years.

19:58 I have never forgotten his name, and I have never forgotten the lesson that he taught me then.

20:06 Unforgettable. So, it is a delight to be with you today to talk about three of my favorite things. I'll share my screen now. Let me know when you can see it.

20:18 Data, demos, and the art of storytelling.

20:21 Because it turns out, and this is maybe the cynical 45-year-old speaking, it turns out that actually I don't remember anything that Peter said, and I don't know that many of us remembered what Peter said. If my name had not been on the ballot, I would have voted for him.

20:40 But there was something about what he brought to that experience that made us all go, yes. He changed hearts and minds. Whoever you came in, your, your cliques, and your buddies, and who was cool, everybody walked out of here wanting a landslide.

20:56 So today, we're gonna talk about data, demos, and the art of storytelling. This is obviously not a photo of me in the 90s. This is, uh, AI generated. But it looked kinda like that, except we were an all boys school, so it was even worse.

21:07 Absolutely. Peter Nonas, I will never forget his name.

21:11 So much. My kids would say he had rizz, he had aura.

21:15 So today, we're gonna talk about data, demos, and the art of storytelling. What I hope I can share with you and install in you is the things I have learned in the 30 years since I absolutely bombed that absolutely dreadful presentation. I changed no hearts and minds.

21:31 And years later, I've been very grateful to have many great mentors and teachers who've taught me some things, and so I hope to share th- those with you in the balance of our time together, and we'll have some time for discussion at the end. So, I'll feel like I've gotten the birthday gift I want if all of you walk out of here with something, just something, maybe it's Peter Nonas, that levels up your ability to do analytics with real impact in this new i- world, new world.

21:57 My name is Matthew. Uh, appreciate the introduction. That is my LinkedIn there, so you can find me by scanning that QR code or, uh, Google me. I've been in product at Tableau for a while now, and been doing Tableau for a really long time. And in, uh, since this is gonna come up, since I've rolled up my sleeves here, it's kind of warm here today, yes, that is a Tableau tattoo.

22:19 Uh, I got that at TC this year. And no, I was not inebriated. Someone was like, "Whoa, what hour of the morning was it?" I said, "It was noon, lunchtime." Lindsey Betzendahl is a bad influence. What can I say? Maybe a good influence.

22:31 A bunch of us got them.

22:33 So, I've been in the Tableau world a long time. This, uh, presentation, though, isn't about my history, but just the little bits you might need to know. I lived in the Netherlands for a long time, about a decade, with my family. We have five kids, or as I like to say, it's between four and seven. With large numbers, you have to approximate. They are amazing children.

22:51 They are bilingual. But we now live in the Chicago area.

22:54 Uh, my wife Becky's doing a PhD.

22:57 And, uh, there's never a dull moment in the Miller house.

23:00 Uh, some of them got the, uh, the gift of gab as well, and there's days where I'm like, "Is this what I was like as a kid?" It must be.

23:08 There's a card game that I really enjoy playing called Cribbage, so we'll try to make this a little interactive. Has anybody ever played Cribbage before? If so, give me some shouts in the chat or some emojis. I'm gonna look over at Zoom.

23:21 All right, there we go.

23:23 Cribbage, you know, Cribbage has been around, it's one of the few card games that you can… they actually know exactly who invented it. It was invented in the 1500s in, uh, in Great Britain. And it's kind of an old person's game. My grandparents played it. My parents played it. I played it. I played it since I was a little kid. I play all the time.

23:39 I was playing just today in an online game with my sister. It's one of my favorite games because it's a nice mix of strategy and skill. I think that's probably why it's had such long-term appeal, 'cause a, uh, new player can win roughly a third of the time, but an experienced player will win more often, but not so much that it blows people out of the water.

23:58 So, it's a really fun game. I've taught it many times, and I play all the time.

24:02 In late 2011 or so, I was actually cycling and I get a call from a friend, uh, uh, my friend from college, Ben, Ben Balsille. And he said, "Have you ever heard of Tableau?" And I said, "No, but I, uh, I'm taking French, so I can spell it." So I Googled, and I downloaded this little piece of software called Tableau Desktop.

24:22 I didn't watch any training videos. I didn't… There were no, there were no TUGs. Actually, I think the Atlanta TUG was the first, and it hadn't been founded yet, uh, or maybe had just started.

24:32 There– I didn't watch any videos, didn't Google anything. I just opened it.

24:35 And I was working as a solutions engineer at the time, and I dragged a piece of data from my work into this Tableau, and immediately saw things in my data that I'd never seen before, and I was absolutely hooked.

24:46 I got a job as, uh, the first European Tableau consultant for a company called InterWorks, a Tableau partner, back when nobody had heard of Tableau. I mean, nobody, even the French were like, "Huh? Come on?" It was, uh, it was, it was tough. All… When people ask like, okay, you know, Tableau's this huge company now, it's a big brand, and how big was Tableau then?

25:07 All of mainland Europe was covered by one Tableau salesperson at that time. It just a very different, different… It was, uh, a radically different time.

25:17 So I get this new piece of software, what can I do with it? Well, I, I, I can see and understand data, that's what they said. With this tool, we help people see, under- understand data. What data do I have around me? Well, I have data about cribbage. So I'm gonna show you, and I'll rebuild for you from scratch, one of the first vizzes I ever built in Tableau.

25:35 It's always a little embarrassing, because I know there's some very talented viz designers on here, and anyone who knows me, I'm a paint-by-number dashboard builder. I, like I know the product really well, and just have no design skill whatsoever. I envy all of you who are such a bill- brilliant infographic designers.

25:51 So I get this tool, and I find some data about cribbage. Now, this is not a demo about cribbage, it's about what Tableau blew my mind with what it enabled me to see in data. I'll just have to tell you just enough about the game though, so you know how it works.

26:06 Like many card games where you're playing with two players, you shuffle the cards, you deal people some cards, and the deal goes back and forth. So you take turns who's the dealer. The person who's the dealer counts an extra, like, bonus hand, which means the dealer tends to score more points when it's their turn. So the two players tend to kind of leapfrog like this throughout the board.

26:26 So I get this data set that has the odds of winning for the dealer and their opponent throughout the game.

26:34 The game's played on a board that goes to 120, as you saw in my slide, and the first player to 121 wins. It's all you gotta know for this to make sense.

26:42 So this is a very simple little scatter plot. Cartesian plot, you got 120 points for the dealer, 120 points for the non-dealer, and each of these marks represents a spot on the board. So I'll go back to my, uh, slides, wherever they are, for a second. You can see this board is kind of in a funny shape, but it goes to 120 points and the first person here wins.

27:00 So imagine that the dealer's at 75 and their opponent's at 65, this is their odds of winning. Well, I had this data set already. It was an Excel file I'd gotten, but there was no insight there. I didn't really understand it. I would sometimes look it up and try to see what the odds were based on where I was in a game.

27:18 It was when I brought out the chance of winning and put that on color… Whoops.

27:24 There we go. This one. I put it on color.

27:26 Immediately I saw something I had never seen before. There's this like sine wave effect that goes through the score set, that as you start out, the person who deals first actually has a 5 percentage points chance of winning, 5 percentage points better chance of winning. So I play with my sister, and she'd be like, "I'll deal first." I'm like, "Yeah, it's fine.

27:49 Doesn't really matter." Oh, it totally matters. She, she was getting a, a competitive edge every time she graciously volunteered to be the first person to shuffle and deal.

27:58 The other thing that jumped out at me is at the end of the game, that back and forth nature actually means you can be a full 10 points behind your opponent, still have the exact same likelihood of winning.

28:06 Again, this is not a demo about cribbage.

28:08 It's a story about how I had data, but I didn't have insight, because I couldn't see and understand it. And literally, I just dragged those three things in, and immediately I saw and understood things about a game I have played a million times since I was a little kid.

28:25 That's my cribbage story.

28:27 So let's continue.

28:29 Uh, a few years later, many years later actually, I was living in Europe and working as a Tableau consultant, and I was invited to go, uh, spend some time with a customer who was doing early-stage cancer research. And in my time as a Tableau consultant, I worked on all kinds of analysis for customers, uh, from soup sales in the Netherlands, insurance, ice cream, financial services, lots of things.

28:54 And, you know, when my kids are like, "Dad, what do you do at work?" And I try to explain it, they're like, "You make pictures?" I'm like, "Well, yeah, kinda." Or they'll be like, "You do data, right?" Uh, one of my kids thinks that all I do for a living is buy shoes, which sometimes feels like it fits. But this time I got to come back and say, "I got to help cure cancer." Well, a little bit.

29:11 I go to this customer in Germany and they said, "We have this problem. We have the data that we collect on these early-stage cancer research experiments that go into a database, and we are trying to understand them and see them so that we can do better experiments faster." Uh, to give you a little bit of explanation, these little things you see on the screen are called ELISA kits.

29:34 Each of those little circles is like a pipette. Think of it like a little tiny test tube. Into each one they put a specific experiment, and they keep track of these, and they run these by the thousands.

29:44 And each of these little pipettes holds an antigen and antibody, uh, the different combination. What they're looking for is infinitesimally small flickers of light.

29:55 If there's a little bit of light given off by the reaction between the antigen or the interaction between the antigen and the antibody, then there might be something there. This is way up at the fu- top of the funnel of how they find drugs that cure diseases like cancer.

30:10 They put them in these boxes, these little sensitive cases that are completely blacked out with very, very s- uh, sensitive light sensors, and they measure the light that emanates from these experiments. That data goes in a database, they rinse, lather, repeat, they do tons of these things.

30:24 This is what the data looked like when they showed it to me.

30:27 You can just see the insights flying off of the page, right? Well, look at it. They, uh, experiment number 24547, and the combination of this and that. They were… That's, uh… There's no insight to be seen here.

30:40 And so what that meant was that after they would run an experiment, they had to do a lot of collation to come up with which of those experiments were worth keeping, which ones were worth throwing out and not repeating, and which ones they wanted to run again and again to see if there was something repeatable.

30:56 And when I brought it into Tableau, immediately there were insights apparent that we hadn't seen before.

31:04 And just with the little bit I've explained to you there about my, like, layman's understanding of what they were doing, I'm sure some biologist in the chat will tell me, uh, I did it wrong. Oh, no, don't skip ahead a slide. Spoiler.

31:15 There are a couple things that jumped out. One is sometimes there's a reaction so intense, look over here in the… I don't know if you can see my mouse moving, but I'm circling this little section here, where there's the four marks clustered together that are darker green. The darker green indicated a brighter reaction.

31:32 Probably a poor choice on my part, actually. Someone in the chat tell me what I did wrong with the colors here. I was new.

31:37 Anyway, darker green meant more intense reaction.

31:40 So these four clustered together, sometimes there was such an intense reaction in one of the experiments that actually you couldn't tell whether it was this one, this one, this one or this one that had the reaction.

31:50 And so what they explained is when we have a cluster like that, we actually, we'll run the experiment again. We'll run all four of them again, but we'll split them up. We'll rearrange them somewhere else on the next set of pipettes that we run through, the next ELISA kit. So if you go back to this, this is really hard to see in a table of data.

32:05 It's so easy to see here, and what they wanted was the ability to basically just lasso these and go, "Yep, those marks, run them again." We didn't have dashboard extensions yet, so that was harder. We had to do some hacks, but very cool. One other thing I'll draw your attention to, this column 24, it's a control column.

32:21 They don't put anything in column 24.

32:24 I said, "Why is that?" They said, "Well, some– the sensors have these, like, arrays, and sometimes a sensor will get broken, and we want to see if that gives a, if that gives any reading at all, then something was wrong, and all those readings are suspect." You can see here, there's a kind of a, I don't, I don't know if screen zoom works on… When I zoom with my screen, do you see a zoomed-in screen or just the same, Celia?

32:42 Celia Fryar: We're seeing just the same, but we are seeing where your cursor's pointing.

32:46 Matthew Miller: Okay. Yeah. Z- zoom doesn't pick up Mac screen zooms. Uh, this, uh, this little mark here tells you something's wrong. Like, there's a m- th- this should be zero. So if this is wrong, then probably every reading along here is suspect. Run them all again.

33:01 And so with that process, like, Tableau massively improved the pace at which they could run their experiments and run them more effectively, so they could get through these combinations, and they could find cures for cancer, which is just really powerful. One more example, and many of you know this one. I, this is, uh, borrowed from a deck by, uh, our good friend Andy Cotgreave, Iraq's bloody toll.

33:22 This is an absolutely famous visualization because it, it breaks some of the rules that we're used to for visualization to tell a really compelling story.

33:34 And in, Andy, I give, give Andy full credit for this, uh, in his version, if you, you know, kind of flip it over, you don't get the same impact. When you look at it this way, it's the same information. It's probably what we would say is the best practice.

33:48 But sometimes you got to break the rules a little bit to be able to communicate with impact. Sometimes you're told, "Stand behind the microphone. Read your speech." But you got to grab the microphone out of the mic stand and coil the cable around your arm like a DJ and break the fourth wall.

34:07 And I love the comparison that Andy has done, like showing what this would look like if you'd done it in sort of traditional best practice viz. You know, shades of one color.

34:16 You don't get the same impact at all in this sort of more traditional approach, even though it's the same data, same charts.

34:25 The power of visualization. And, and that, uh, I'm sure we have a wide range of Tableau experience and history here on the call, some of you who may be new to Tableau.

34:34 I, I love every opportunity I get. I'll, you know, go off track for a second to say, I just love the Tableau user group so much. It's one of the cool things about the Tableau community, these self-organizing groups of people who get together to share best practice. I really hope, like, I, I, I value the opportunity to spend time with you, and so I really hope that I can give you something of high value since you're investing this time in being here.

34:55 We love the, the Tableau user groups. Um, Mark Recker, the GM of Tableau, and I are talking about doing a competition to see who can speak at more TUGs this year. So thank you for the chance to get a, a head start on that competition as well.

35:09 The thing about makes Tableau special, if you're new to Tableau, you're not using Tableau, or you're just learning, and maybe even if you've been around a while, is that there's this fundamental philosophical difference from every other BI tool. And I've been doing data stuff actually long before Tableau. I'm old. I started using business objects in 1998.

35:27 And what every other tool does is it goes, "What chart do you want?" "Uh, I don't know. A pie chart." "Great. Where's your data?" "Okay." And it's there. There's no opinion. It just says, "What do you want?" And even in the AI world, even if you've used AI agents, they'll just do what you tell them, "Give me a pie chart. Give me a line chart." Whereas Tableau is opinionated and says, "Actually, hold on a second.

35:51 What data do you care about? What story are you trying to tell? Let me make it easy for you to explore different perspectives to find the one that's the best." Because Tableau was built by scientists and backed by research to tap into the GPU of the mind, not built by, you know, MBAs in suits to flip their startup to the next big company.

36:09 Part of our DNA. So the, the, those, those stories hopefully paint a picture of, uh, or give you a way to understand my background, but without boring you and reading you my LinkedIn.

36:18 Hopefully gives you a sense of my Tableau journey and what, what blew me away the first time I experienced Tableau and what still is so powerful today. Hopefully gives you some, uh, awareness of some of the ways that Tableau can have a lot of impact, and data visualization can have an impact.

36:34 So the remainder of the talk I want to share with you, uh, I've got three kind of key points and a bonus point as well that you can use as you are– for those who are the note-taker, and I know you're there. You're, you know, write, writing things out, or maybe you're having Gemini transcribe it.

36:47 Uh, three principles and points with some illustrations to help you use data and storytelling and demos to have greater impact.

36:56 Because what I hope you took away from those opening stories is actually- Data and demonstrations of dashboards or demonstrations of applications, and storytelling, those are all means to an end.

37:08 They're not means in and of themselves.

37:11 The purpose of all this is to see and understand your data and take action. Why?

37:17 So that we can cure cancer, so we can improve healthcare outcomes, so that we can sell more soup perhaps, so we can reduce failure rates, so that we can understand how much YouTube our kids are watching on their school Chromebooks. That's something I've done. Whatever the thing is that you're trying to drive toward, all this stuff we're talking about here are just tools in your toolbox to get toward that impact.

37:37 So I hope we don't lose tra- track of that as we add that stuff. All right, principle number 1, hook your audience. Break through the noise. That's why I opened with the story about Peter Nonas. That's why I kind of cold opened with that story because, because it's so attention-grabbing. It grabbed my attention.

37:53 One boring speech after another, uh, forgettable. No one from high school remembers me. But Peter Nonas, it's been 29 years, what a legend. Right? He broke through the noise.

38:06 I think sometimes the mistake we make in our world is we assume that the person who brings the most data is the one most likely to influence. Have you ever seen a dashboard where the designer clearly thought, and maybe you've been guilty of this, I know I have, where the designer clearly thought, "You know what will maximize the impact of this dashboard?

38:26 I'm gonna cram every possible insight and filter into it, and maybe if it's got 16 charts and 30 filters, it's gonna change people's minds." Not true. Turns out that actually there's this inverted U curve, this diminishing marginal returns.

38:40 I'm gonna be a little bit cheeky.

38:41 I'm gonna share with you an example from one of our competitors of what not to do.

38:46 I was part of a demo showcase. Unfortunately, mine is, uh, not recorded, but I've got some bits of it to show you. Uh, I was part of a demo showcase in Florida a few years ago, and on the stage was me representing Tableau, and a guy from Microsoft, lovely gentleman named Patrick, and someone from, I think Oracle.

39:04 Uh, and we're in a room in front of 4,000 chief data officers at the Gartner Conference. Kind of terrifying, maybe the most terrifying demo I've ever had, 20-minute-long demo, 70 screens, all memorized. This is how Patrick from Microsoft opened. We're gonna use this as an example of what not to do. Tell me when he hooks your attention.

39:26 For this year's Gartner Bakeoff, we are asked to investigate the likelihood of more frequent flooding events worldwide to see if there's any mitigation strategies governments could take to reduce the risk of flooding disasters.

39:37 First, let's take a look at the data we pulled together to analyze this topic. 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 plo- prone areas, and how much, uh, built-up area there is that is at risk.

39:54 Matthew Miller: I can't take it anymore. It's so bad. Make it stop.

39:57 Like, why should I care about this? What is going on in this screen? Lots of boxes and lines. CDOs like boxes and lines. But what's happening here?

40:07 Why should I care?

40:09 That principle, in fact, it's my pinned tweet. It's been my pinned tweet for, like, five years, is that I believe no one cares. By default, no one cares about me or what I have to say. I, I keep the introductory slides as light as possible. I have in another deck where I talk about this, an example from a customer where, uh, their, one of their executives gave a talk, and then I was gonna do a keynote, and he opened with this slide about himself and spent probably five minutes on it.

40:35 I just remember at the bottom, after, like, I was like, "Dude, no one cares where you went to high school. You're killing us." At the bottom he had, "Fun fact, I have a dog named, like, Fluffy." And I had to cover my mouth because my f- my brain said, "I don't think even Fluffy finds that fun." Um, so it gets me in trouble. It just, nobody cares.

40:55 You have to earn the right, even at work. Even with colleagues who have to, you know, are paid to listen to you, you gotta earn the right to have people's attention. You gotta hook your audience with something that is memorable. Let's see if we can…

41:07 We brought in information about major flooding disaster.

41:09 Matthew Miller: Try to stop the… My, uh…

41:11 My, my sheets are, like, stuck. Let me see if I can move to the next slide from another monitor. It won't let me go.

41:16 For this year's Gartner Bakeoff-

41:18 Matthew Miller: Oh

41:18 … we were asked to investigate the likelihood of more fre-

41:21 Matthew Miller: It's, it's killing me. We- it's, we're, like, permanently stuck on Microsoft. Wow, I was close. Woo. All right. So you've seen what not to do to hook an audience.

41:30 Uh, let me tell you what to do. And here, I gotta give a shout-out to one of my colleagues named Vikram.

41:36 Vikram is an SC at Tableau. He's absolutely brilliant. I'll do my best to relate what he did. Uh, I was, uh, in San Francisco to meet with this customer for an executive briefing, and I will confess, I kind of phoned it in. I didn't go above and beyond. I delivered, I had about an hour, I delivered the same slides I often do about our roadmap. And I thought I was engaging, but I didn't really…

41:56 You know, I didn't put my best foot forward. I kind of phoned it in. And then Vikram went, and I'm glad he went second.

42:04 Vikram gets up there. Now, this customer, uh, has, uh, does robotic surgical equipment similar to what you see on the screen.

42:12 So Vikram doesn't pull up his slides quite yet. He doesn't introduce himself. He doesn't stand behind the microphone.

42:18 He said, "We found out last year that my dad needs surgery." The whole room just all focused on him.

42:28 He just had, like, the…

42:30 He had, like, the air of a, of a preacher, in a good way. And he said, "We found out last year that my dad needs surgery.

42:40 Surgery is such a scary word, a lot of fear. We don't know what that's gonna mean.

42:44 Like, is he gonna be okay?

42:47 We didn't know. But then we went and met with the doctor, and the doctor said, 'Actually, this is really low risk. It's laparoscopic, meaning it's not… It's minimally invasive.' Now we're using this new robotic surgical technology from," the customer we were meeting with- And as a result, we felt at peace.

43:07 The surgery was a non-event, and my dad's fully recovered.

43:11 So for me, it's a particular delight to get to be with you, with the people who helped save my dad's life. Now, you don't always have that story for every audience. I wish I did for every audience, but I was sitting there going, "Holy crap. That's incredible," right?

43:24 In that moment, he had absolutely connected with them in a way that was just so powerful. All the previous presenters, myself included, were just womp, womp, womp, womp, womp, womp, technology, Salesforce, data, semantics.

43:37 And Vikram connected on the human level first, and they just ate out of his hand for the rest of that hour.

43:43 And I, I went over afterwards and said, "That was a master class. Teach me." It turns out he'd taken a class on storytelling at Stanford, and, and so that, that principle of hook the audience is principle number one. It turns out that nothing else matters. If you've lost your audience, whether it's a story at work or it's a data presentation or it's a demo, if you've lost your audience's attention, why are we still talking?

44:06 Hook the audience and retain their attention. Point number two, establish trust.

44:12 Why should I care?

44:14 I have two examples. I was debating which one to share first. I'll go with this one.

44:21 Uh, when I joined Tableau as a product leader after years working with Tableau, uh, the first thing I was asked to work on is what's now called Tableau Cloud Manager. Uh, give me some shouts in the chat if anybody has used, uh, TCM or knows what TCM is. If you've administrated any Tableau Cloud site recently, you have had to use Tableau Cloud Manager.

44:42 It's what everybody uses. Anybody? It's so hard to just talk into the webcam. There we are. We got some… Yeah, Christian's here. Hey, good to see you, man.

44:51 Um, thank you for the happy birthday note this morning, my friend. Uh, yeah, so you know what TCM is. TCM is for managing your Tableau Cloud sites. Well, I was the first PM for TCM, and it began with my boss saying, "All right, uh, Matthew, I'd like you to come up with a strategy and a plan for dealing with the multi-site problem in Tableau Cloud." I was familiar with the multi-site problem.

45:11 I had worked with many customers before. I had sold and installed Tableau servers and helped people migrate to Tableau Cloud for ages, so I was very familiar with the challenges of managing multiple sites.

45:21 I- in short, customers who came from Tableau Server were accustomed to a certain amount of flexibility in how they managed the buckets that they organized content into, and users in Tableau Cloud didn't have that same flexibility. We actually had the capability under the hood to some degree, but we didn't have the same APIs and UI, and so admins who used Cloud or who migrated to Cloud said, "Ugh, we lost some of the flexibility and power we, uh, we enjoyed about Server.

45:43 We want that same thing on Cloud." So to me, as a product person, like this is the easiest feature ever. There's no real innovation to be done. How should it work in Cloud? How about the same as Server? What expectations do users have? The same as Server. What UI should we have? About the same as Server. Like, this was just kind of let's take this stuff over here and just make it work the same over there.

46:06 To me, this is a no-brainer. That said, it was a lot of work, which meant we had to get approval for a big investment. It was gonna be millions and millions of dollars. Uh, I guess I shouldn't say how much it cost, but it cost a lot. Building Tableau Cloud Manager was gonna be very expensive. It was a lot of engineering. We were asking for a lot of support from leadership, which meant we had to get an approval.

46:26 So I wrote up a strategy document. It didn't seem very hard. I wrote it up pretty quickly, and most of the tone of my document was kind of, duh, we should do this. How much will it cost? A lot. Should we do it? Yes, obviously. It was so obvious to me. I was fully bought into my own idea.

46:45 And I went and built this visualization. It's probably the most impactful visualization I've ever built, which is why I show it to you here. I've had to anonymize it.

46:52 I'll explain it. On this axis, uh, uh, sorry, the X-axis is a log-log scale of the active users in a customer's deployment. This is a survey of enterprise customers who are using Tableau Server at the time, and this is five, six years ago, and this is their number of sites, also a log-log scale.

47:10 So you can see there's quite a distribution here, and, uh, it's not a perfect correlation, and I knew this intuitively, and the data backed it up, that there's not really a correlation between the number of users you have and the number of sites you have for reasons that are not important for this conversation. And, and some customers had thousands.

47:28 By the way, I think the maximum I've seen since then is, like, 5,000 sites. We have one customer who has, uh, over 500 Tableau Server clusters.

47:37 That's incredible. So I knew there was a lot. I knew people had lots of these, and so I go into this meeting, and I present my document with probably some arrogance of like, "Yeah, obviously this is a no-brainer. Just give me your thumbs up and your… You know, write me a check, and we'll go build it." Completely shot down, rejected. Tableau's leadership at the time, uh, we had a CEO named Adam Selipsky who came to Tableau from AWS, and he was like, "I don't get it.

48:03 Why should we do this?" I said, "Well…" I was taken aback. Like, I mean, why not? It's, like, really obvious. And he said, "I thought every customer had one site and only one site." And I said, "Uh, what?" And it turns out that he had a misunderstanding based on… You know, he's a sales and marketing kind of leader. I want to speak positively of Adam.

48:21 He was a great CEO of Tableau. Uh, but he had a understanding of the business that was not anchored in, uh, technical or product reality. He was a sales and marketing leader. Very good.

48:32 So he said, "I thought every si- every customer had one and only one site." I said, "No, no, no." So then I show him this viz.

48:37 And here's the mistake I made, and maybe I'm overselling the point, but I'm telling you, I, I came into Tableau after almost a decade as a user and consultant just totally convinced that I would walk into this nirvana where, where I would go, "But see the data," and everyone would go, "Wow, we should just do what the data says." I can't tell you how wrong I was.

49:02 Just smack into this wall of the reality of humanity.

49:07 But he said, "What?

49:09 Huh?" In fact, I have a screenshot, a comment he made on my strategy document, which was just a series of about 50 question marks. He did not get it. My proposal was rejected. I was crushed, but I had the data, I built the viz. I mean, this viz is so obvious, right? How could I… I was proud of it. And my boss was very helpful, and several other people, some of you remember Francois, who was our CPO, uh, pulled me aside afterward virtually and said, "Look, um, there's some politics in the room.

49:35 There's some misunderstandings.

49:36 Adam is not really a product leader. He's a sales and marketing leader. He's gonna look to his lieutenants in the room to see what they all think. You have to convince them first." I said, "That's exhausting. Why don't they all just get it?" "Matthew, come on." And so I did. I went on a tour. I met individually, one-on-one, with every single leader who had been in that meeting, and I walked them through my document one-on-one.

49:57 It was a lot of work. To me, as a data person, I was like, "Why didn't the viz just speak for itself?" That's what they promised, right?

50:05 See and understand your data. Iraq's bloody toll. I look at it, I get the message. Uh, things I've heard over the years, things I've said in trainings, like a good data viz shouldn't require any documentation. You'll just get it.

50:16 It ran smack into the reality of humanity.

50:19 Once I had done that process though of walking the room virtually, so to speak, and met with e- each person and overcome their objections and answered their questions and walked them through the process and the reality, I presented the same strategy again, the same ask, with basically the same document, uh, three or four months later to the same leadership team, and it sailed through.

50:40 And we built TCM, and here it is.

50:42 And it was such an eye-opener for me. Maybe I spent longer on this story than I should have, but if you take anything away from this actually, it's probably Peter Nonas. And it's the sort of, uh, skepticism earned in my 45 years that actually decisions made by humans are complex.

51:01 People didn't judge the speeches for student class vice president on their merits. If they had, I probably would've won. I had a good plan, dang it. My mom wrote it. It was great.

51:13 They were judging it on all kinds of other factors. That's the, like, harsh reality for me as a data nerd. But, but, but the data, right?

51:22 The, the data is only one ingredient in the transformation that we're trying to drive. This is the most impactful viz I've ever made, but it still got flatly rejected until I understood the full 360 degrees of how data fits into the kind of impact. And that's where storytelling, by the way, was very useful. Um, I'll, I'll go into this a little bit more later, but I'll…

51:45 I, I wanna move on from this one, so let's go onto the next thing. But, uh, uh, there's a just a kind of concept here about how to surround an idea with all the things, I, I coach PMs on this, that anyone might want to look at, and one of the most powerful things is story. Uh, not a story I planned to tell, but I'll tell it very short. The next project I took on was our action integration, like a MuleSoft integration and Salesforce Flow and Slack workflow integration.

52:09 Big, thorny, you know, lots of people with lots of opinions. I had to kind of herd all the cats and come up with a plan. So this time I, uh, I understood what I had to do, and I made sure that we led with a story. And I did a lot of customer interviews to understand what people wanted. I was talking with a customer in Houston, actually, a environmental management company, and I said what we were gonna do, what our plan was, wanted his feedback.

52:35 I'm gonna have to, pardon my French here, but I'm gonna quote this guy directly. He said, "Wait, wait, wait, hold on. So you're telling me I can have like a dashboard that's got all my customer accounts on it, and I can see them like visually which ones are likely to attrit, and then I can like select some of the ones that I wanted to focus on. I can click a button, and it would go like update Salesforce and kick off a campaign or, like, send some Slack messages?" "Uh, yeah.

52:58 No, that's exactly what we're building." And he said, "Holy shit! That would be incredible." And so over the, uh, objections of my manager, the headline, the, the beginning of that doc, I should've put a screenshot in here. The, my strategy doc was like Matthew Miller, Tableau's act flow- action workflow strategy.

53:16 Big words. Pull quote, "Holy shit."

53:21 And then the rest of my, uh, my story.

53:24 And it– that one sailed through, got funded right away.

53:27 Because I led with an anecdote.

53:29 Helped that it had a big customer name, but led with an anecdote. They got it. And I think I should've done it with Adam. I didn't start with an anecdote. I should've started with, like, talking to, uh, you could pick any, any friend who's a Tableau Server admin or Tableau Cloud admin and, and go, you know, Bob. Bob struggles with managing multiple sites.

53:47 Bob's life sucks. He complains that insert competitor does this easier. Loved what he had on Server, misses it on Cloud. He's considering switching back to Server, right? Let's help Bob.

54:00 Like, it sounds tongue in cheek, but it turns out the anecdotes are really powerful. There you go.

54:05 The best viz I've ever made, and also, uh, most impactful. All right. We're gonna move on. I think we're doing okay on time. Nope. Gotta go a little faster.

54:15 So we're, we're, uh I, I've kind of went way down that path. It's, it's, it's… The thing is, I mulled this over yesterday, Celia, and I was like, "What do I wanna leave people with tomorrow?" I'd put together the deck and, and had some principles and stories to tell, but I thought, gosh, actually if I do anything, it's to almost like disavow us of the notion that data is the only thing that changes hearts and lives- Yeah…

54:33 hearts and minds. Yeah. And data viz is everything. So I, um, if you take anything away, take that.

54:39 Um- That's great… we're gonna… I'm gonna skip this. I have a video of this. It's the pine tree story.

54:44 But instead, I'll just go. I told you I would show you Pulse and, and how this fits into, like, changing hearts and minds. So, uh, I'll do this really quick. This is my backyard.

54:54 Uh, some of you w- maybe, uh, have seen this. This was, this was online. I, I've done this in a couple of keynotes. I did this at DataFam Europe, so forgive me if you've heard this story before.

55:02 Uh, um, this is a true story and a data story. So this is my backyard, which is right there out that window, and this is my spruce tree. And you can see this is from a couple years ago. It looked a little thin. So I called a tree doctor, and he came over, an arborist.

55:15 Uh, called himself a tree doctor, charged 400 bucks, and he said, "Yeah, your soil humidity is too low." I said, "Okay.

55:23 Do I look like I know… Like I could just, you know, put my finger in the dirt and go, 'Oh, yeah, it does feel dry.' I don't know." And he… I said, "Well, what should it be?" He said, "Between 40 and 60% relative humidity." "Okay.

55:35 How do I fix that?" He's like, "Well, just water it more." Great. And he charged me a bunch of money and he left.

55:40 Okay, now I gotta figure this out. So I, I, I bought an IoT sensor, this little green thing there with the flags, put the flags there so my kids wouldn't kick it over, and I hooked it up to a little home automation thing and piped the data into BigQuery and brought it into Tableau.

55:55 So this is, and this is real, you can see this is as of, I think it's on, uh… Yeah, there you go, just, uh, an hour or so ago. This is the actual real time humidity of my… the soil around my backyard pine tree.

56:07 And you see we're getting low here. I need to go out and water this thing.

56:10 So I started measuring it, and measuring it made it easy to me- uh, manage. Uh, just for the data nerds here, um, you see some spikes here in the data. Anyone have an idea what those spikes are?

56:21 Somebody in the chat?

56:23 Just wanna see if you're paying attention, or did I put you to sleep like my, uh, class VP story?

56:30 Rain. Yes, precipitation. There you go.

56:33 Uh, exactly. So you can see those spikes, and it's okay for it to be a little, uh, damp during the summer. You see it's getting dry. We're going into that dry scene, and I gotta water it. So I start tracking this, but of course, the challenge is I have to go look at this dashboard. I have to go look at this viz.

56:48 So I start doing that, and, and then I re- realized, hey, actually, we have a tool for, for actually managing this kind of stuff. So I, I go into Pulse, and I create a Pulse metric for this. So now I get this daily summary that's sent to my inbox. I… Actually, if you try this out, it's really cool. I just set it up a few weeks ago. We now have Pulse widgets, uh, if you have iOS.

57:08 So I don't know if you can see that, but I have my pine tree right there at the top, the Pulse widget right on my homepage of my phone, so I can tap into this. And what, what we did with Pulse, wi- without preaching about Pulse too much, and we said, okay, what do most people do? Well, what they do is they sign back in again in the middle of a demo.

57:25 What do most people do? Oh, thanks, Tableau.

57:29 The irony here that my demo is interrupted by Tableau Cloud Manager is not lost on me. So if you're laughing, good. All right, switch back into our site again.

57:39 There we are in… There's my, my viz. And let's go back into Tableau Pulse.

57:43 What we did with Pulse is we realized, hey, how can we help people see and understand data proactively?

57:48 The power of the drag and drop experience that I showed you, like my cribbage viz and those cancer vizzes are, are very powerful. But most people build line charts and bar charts.

57:58 So can we do that automatically?

58:01 Well, we can. So that's what we did with Pulse. And then there's a lot more that Pulse does. This isn't a Pulse demo per se, but I built this in Pulse, and what it enabled me to do is to get this now in all the places that I wanted it. So I've got it pinned on the homepage of my phone. I've got it sending me emails every day with summary about how this thing is doing.

58:19 It's unusually low. You can see it actually warned me today. The front house, uh, I was just thinking earlier today I need to go out and water it this evening when our local suburban regulations allow me to.

58:29 So I start using this to track my backyard pine tree, and it starts to get healthier. Not long after, my wife Becky said, "Ooh, what's that?" I said, "Oh, that's our backyard pine tree." She said, "Oh, can I use that with indoor plants?" I said, "Absolutely." We have this fiddle leaf fig in the dining room, and it's our third fiddle leaf fig because we killed the prior two because we forgot to water them, because she's a busy PhD candidate, and I'm on the road every week.

58:51 And so I went and bought another IoT sensor, and I stuck it in the dirt. And so Becky, who is very smart and very analytical but is not a data nerd, crucially here, started using Tableau for the first time. And she knows, she's heard a ton about Tableau. Sh- Tableau has dominated my life and career for fifteen years, so she knows a lot about Tableau.

59:10 I have tried, at least in the early days, naively, to be like, "Hey, honey, don't you want to drag and drop with me? We could sit on the couch together on the weekend and work on public." Look, you know, you can laugh if you want. I tried. That's just not her jam. She's analytical, but she's not a data nerd. She doesn't like to sit there and fiddle with, you know, columns and rows.

59:26 But with Pulse, she gets what she wants. So she gets the actual insights. Like, she doesn't really care about Sankey charts versus Marimekko. She's not in that… She's not in the zone of that stuff. But she cares about keeping her pine– her, uh, fiddle leaf fig al- alive, and so this has been so transformative for her. The problem is that once I did this, she was like, "So how much are those sensors?" I'm like, "They're ten bucks." She's like, "Cool." And so now we are tracking all kinds of stuff.

59:50 I filtered some of them out, actually. But we're tr- we're tracking a whole bunch of plants and other bits of our house with these IoT sensors, and we're running it all through Tableau. And the primary way we can consume it now is through Pulse and through MCP.

1:00:03 So as a side note, like, I'll use this in Claude, and I'll, "Hey, how's, how's my pine tree doing?" And it reaches out to Tableau, and because it goes down through the Pulse layer, it's not going straight to this raw data, which doesn't really have insight. We talk about semantics, but what Pulse adds is all this additional insight. All of these, this trend information, all those things that it's keeping track of, what we call the insight engine of Pulse, are all fed to the model in addition to the raw data, so that it knows much more about, uh, what actually healthy means.

1:00:32 For example, if you look at this viz, I encoded into it with this reference range that forty to sixty percent is where we wanna be. So I have to know that. I have to encode that as an analyst, and I have to look at it visually to see, yep, we're getting close. Like I, I'm the one who encoded color here as the, uh, you know, orange means, like, it's getting dry.

1:00:51 But with Pulse, I just set it up. That's your expected range. And you can see we've been within the range, uh, for the fiddle leaf fig, but it's getting a little bit low.

1:00:59 So, uh, I wanted to show you that because one of the ways that we could, like, drive change is th- through that concept of what gets measured gets managed.

1:01:09 And so part of our journey and where AI fits into this, and this is, this is… I say AI, but it's kind of a broad… This is not generative AI. This is more like traditional AI and machine learning and, and also base statistics. Is about how can we go from a world that we started in 23 years ago where I have to open Tableau, I gotta connect to some data, I gotta know how to drag and drop the right things, I gotta know how to do a nested table calc, I gotta write the LOD.

1:01:36 Ooh, now I got insight. To a world where Tableau knows enough about us, and knows enough about the problems around us, and has enough context and access to the data that it can say, "Yeah, I can just build this for you." I said this at DataFam Europe, I said, "I didn't have to drag and drop a damn thing," which is like what? Drag and drop is what we do.

1:01:55 But that's actually, I think, part of the future, is how much can we use the, the, the new types of technology that we're adopting to continue to drive toward that mission of help everyone see, understand, and act on data. And that's not always gonna mean drag and drop anymore.

1:02:10 All right.

1:02:11 Uh, there we go. Now, we're, we're gonna keep, uh, we're gonna keep moving. We're okay on time. I do wanna leave time for Q&A, though. I'll pause, though, for I've, I've been just talking straight into the webcam. You know how much I love being with a live audience, because I can tell if you're paying attention. I have no idea if this is landing. Um, so far anyone have any comments, thoughts?

1:02:30 Is this helpful, useful? Maybe we'll take one question now.

1:02:35 Anybody at all.

1:02:37 Celia Fryar: You guys are free to come off mic and just ask your question.

1:02:40 Matthew Miller: Yeah, or through the chat. Either one's fine.

1:02:44 It's a quiet bunch.

1:02:46 Celia Fryar: Mm-hmm.

1:02:47 Matthew Miller: All right. Well, then. So we've hooked our audience. We've established trust. Now what do we do with that? So we've gotten their attention. Uh, have you ever seen, like, a Super Bowl ad? I realize this is a very US-centric kind of message. I saw Peter Johnson was on from Sweden, but Peter knows what the Super Bowl is. Uh, so if you've seen a Super Bowl commercial and been like, "That was so funny.

1:03:10 What was that for?" Right?

1:03:12 Like, you can have, uh, an ad that's brilliant, it stands out, memorable, but doesn't actually achieve its goal of inspiring the kind of action that you want. I think that happens sometimes in the data world. We're like, "Look at the cool viz." Awesome. That was a very cool viz. What did that radar chart tell me? Like, what should I do differently?

1:03:34 So cast a vision for a beautiful future.

1:03:36 Here we get into the science of demonstrations.

1:03:41 So I'm gonna share with you a framework that I learned from our good friend, uh, Tableau's former chief product officer, Francois.

1:03:49 Problem, solution, achievement. And this is useful for demos. It's actually useful for almost everything. But specifically, this is what I teach the teams that do Devs on Stage when you approach a demo, and I see the opposite of this all the time. Think of that really, really awful Microsoft, uh, video I showed earlier. Trust me, it got worse. It got more and more boring.

1:04:10 Um, and th-this is, this is how you make sure that what you're talking about and what you're demoing lands. So these principles are useful whether you're showing somebody a dashboard, you're demoing some software. First, the problem. Be crisp about the problem.

1:04:26 Know your audience. I had a mentor early on in my career when I was a s- a solutions engineer. He was a salesperson.

1:04:33 And he said, uh, "Matthew, here's how you approach sales. First, you find out what wounds and pains your audience has.

1:04:42 Then you rub all the salt you can in the wounds." My eyes are getting big like this. I'm like, "Isn't this a little bit masochistic?" He's, "No. Rub all the salt you can in the wound until they're crying and begging you for relief. And then you stand back and go, 'Only I can heal you from your pain. Sign here.'" While that's a bit extreme, that's basically what you want to do.

1:05:03 Figure out what it is that they care about. And by pain or salt in their wounds, I mean, let's say I was talking about TCM. I'm gonna demo TCM, Tableau Cloud Manager, to an audience of admins who've never seen this before.

1:05:16 They're administrators of Tableau Cloud. So I'm doing a talk, let's say, at Tableau Conference. I know it's all admins in the audience, so I can be technical.

1:05:24 Uh, and I, I know what kind of problems they have. They have governance issues. I'm thinking about what their bosses care about.

1:05:30 Oh, man, how many sites do you guys have on average? Oh, man, we have 10 sites, 5 sites, 20 sites. Yeah. Managing multiple sites is a pain, isn't it, right? But, but, but you can't, uh, you can't, uh, you double count users. I'm trying to remember the problems we were addressing. Double counting users. You can't allocate resources across them. You can't this, you can't that.

1:05:48 Oh, man, that sucks. Wouldn't it be amazing if that were easy?

1:05:54 Introducing Tableau Cloud Manager.

1:05:57 Then you talk about your solution. Here be clear, be succinct, and be passionate about what you care about. I have, uh, yeah, witnessed examples where this isn't done. I won't tell that story. But, uh, yes, I've seen examples where people have really exciting things to talk about, and they describe them like they're describing, uh, the death of a pet or something.

1:06:15 It's just, you know, hey, if, if you actually care about what you do, you, it should come through, right?

1:06:23 You should have some passion for the things you do. So explain your solution and what it, what it is.

1:06:28 Be clear and succinct. Know your value prop. And then, and I think people forget this far too often, show me what the achievement is.

1:06:36 That is, show me how the world is better thanks to this thing.

1:06:41 And, uh, we, over time, we've codified this with the Devs on Stage crew i-into these four things, which is easier, better, faster or more.

1:06:50 And if the thing that you're telling me about, the feature we ought to build, a demo we ought to do, a project we ought to run, if you can't tell me how me or people I care about or a problem I have or a project I'm working on will be easier, better, faster, or more, ideally more than one of those, I won't be interested.

1:07:10 Always find a way to show what is the easier, better, faster or more.

1:07:14 And I hope that you saw some of that. I tried to actually, my sort of goal here, I don't know if I achieved it, was to open with some examples that sort of told my career story, but also gave you some examples or at least ways that it worked for me to do that.

1:07:27 Uh, I've got some more examples for you. But first, a little gift. Uh, I wrote this, uh, Matthew's Top 10 Demo Tips. That is a… I think I was still in my 20s, uh, at a conference demoing some of the worst buggiest software I've ever worked on. This was not at Tableau, my pre-Tableau days. It was, it was awful, and, um, but really buggy. And I had to demo this at a conference in front of, like, 450 people, including a bunch of industry analysts.

1:07:55 And, uh, I've codified Matthew's 10 Tips.

1:07:58 This is a Salesforce document, so I can't share it with everyone, like, as a guest. So, you'll have to request access with your Google account, but I'll approve everybody who asks, and that's how I know if you were paying attention anyway. Matthew's Top 10 Demo Tips. You can also use this URL, which will get you to it. Uh, I won't go through all these in detail.

1:08:15 There's a lot of good content there. But, uh, and th-there's– we could spend hours just on how to do a great demo, but there's lots of things. I'll, I'll pick one, though, from this that I, I really, I find helps a lot, is physical metaphors. People really struggle to have a mental model sometimes for abstract technological concepts or data concepts, so physical metaphors help a ton.

1:08:38 Uh, the, the, the one I can think of that is the weirdest I've ever come up with, back when I was a Tableau Server trainer, is I said, "Here's how the worker nodes work with the manager node in a Tableau Server cluster. It's like a mommy bird and the baby birds, and the baby birds are all like, 'Cheep, cheep, cheep, cheep, cheep!' And they'll just eat whatever the mommy bird drops into their mouth.

1:08:57 And so in the same way, the manager node in a TabAdmin-era Tableau Server cluster pushes out things, and the worker nodes just receive it. 'Thanks, Mom!' And that's how Tableau Server works." And, you know, you could see on the eyes of the people in my training that they thought I was out of my mind. But I'll bet they remember it. And so using physical, physical metaphors help people map it.

1:09:19 I find this is really important in the AI era, very frequently mapping agent behavior to human behavior. Go, "Well, all right." Like, if I talk to Celia today about something, and then on Tuesday she was like, "Who are you again? What do you do?" Like, why does our agent always start every time as if you've never met me before? People expect the agent to behave like a human.

1:09:42 The more you can kind of human, um, yeah, bring humanity, uh, and, and something physical and tangible, a mental model they can map over, and there's much more. So, a little gift for you.

1:09:53 Uh, briefly, my favorite demo I've ever done, some of you may have seen this because I did it in Devs on Stage a few years ago, was, was called Tableau Gestures, a Tableau research project.

1:10:04 And, uh, uh, it's hard to tell a story about something you did that worked out well without sounding self-aggrandizing, but it is a decent example, so forgive. It's my birthday, right? Forgive this one. But I'm trying to show you how, how I attempted to leverage these, uh, principles. So, our research team says, "We got this really cool thing that you should show." "What is it?" They said, "It's called automated chironomia." I'm not making this up.

1:10:27 You can Google it. There's a whole research thing about it. Automated chironomia.

1:10:31 Okay? That just flies off the page at me. I just– I'm hooked. What is it? "Well, so what you do is you can do– it does finger detection and different, uh, types of finger detection, like snapping." And I was like, "Huh?" And they're like, "And then what you do is you can, you can go like this," and they, they showed how I could drag a chart like this out of thin air into a webcam like this.

1:10:57 They did it by overlaying on your video display. I should have included a video of this. I will, I will send you one to, uh… Maybe I'll find it in the chat. If anybody has… You can Google Devs on Stage from that year and find a link to it.

1:11:09 Uh, and, and so I was like, "Well, that's really cool." But it took us, you know, half an hour of talking about this before I understood what was cool about this.

1:11:18 And what's the problem that it solves? Well, the researchers just, you know, stumbled on something cool, and I realized, oh, I think the problem, and this was in like twenty-twenty-one and twenty-twenty-two era, the problem is that we have more virtual meetings than ever.

1:11:31 We have distributed teams. A lot of people are remote now coming out of the pandemic, and we're trying to find ways to communicate information that makes sense. The ratio of, like, the scale of the slide to the person is different in the virtual world. Like, right now, my slides are probably occupying eighty percent of your screen, and my little thumbnail is probably, like, five to ten percent.

1:11:53 But if we were in the old days, right, where we'd meet in person, like, I'm, you know, six four, and my slides are, you know, projector's not that much bigger. So, like, my, my, uh, hand gestures and so on, all the things we do that are the rest of the communication actually get lost, and the thing we're showing is, is predominant. And what this did is it overlaid it, so you could make your s- your webcam full screen, but still interact with the viz.

1:12:19 So the, the salt that I was kind of rubbing in the wound is, haven't we lost something about storytelling and communication in this all remote distributed era?

1:12:29 What if we could? And I said, "All right, introducing Tableau Gestures." And, and the thing that we arrived at, working with one of the really brilliant designers at Tableau, was that actually the way that the original demo was set up is you would drag the entire chart into the view, a finished chart. You'd configure it and then drag it into the view.

1:12:53 And we realized actually that's not how Tableau works.

1:12:57 Part of what makes Tableau so powerful is the instant feedback, the ability to explore from different perspectives, and you actually don't sometimes want to get the full finished picture at once, because it can be overwhelming. You want to tell a story incrementally, and this gave you the ability to do that. So we flipped the whole model around to where what you brought out was not a finished chart.

1:13:19 You'd pinch and drag out bits of data, and the viz would update based on that. And you can see, this is actually– it's hard to tell the scale here, but this is a, like, floor-to-ceiling jumbotron in a big, uh, conference room, so probably, I don't know, ten meters high.

1:13:33 And the… You can see my, my face is, you know, kind of looks like 1984 here, I realize. But it, you know, it's, it's, it's full size along with, like, proportionate to the story that I'm telling.

1:13:45 And, uh, that was easily one of the best, uh, demos I've ever done because we, we anchored it not in here's cool tech. L- and trust me, I nerd out about cool tech so bad, I have to remind myself of this all the time. It's let's anchor in a problem we have. Human communication has taken a big hit in the pandemic.

1:14:06 We're all on these webcam calls, like Zoom, like this.

1:14:09 Yeah, I can't see your faces right now, that kills me. I can't read your reactions. I don't know how it's… It's hard. I know you can see my slides full screen, and I'm tiny, so you can't see all my face. I'm Italian American, you're missing all the, you're missing all the hands.

1:14:23 What if we bring back some of that humanity but still have the benefits of this virtual world, or take advantage of that in the virtual world?

1:14:29 Introducing Tableau Gestures.

1:14:32 Seamless, exploratory data storytelling makes it so much easier for me to tell a story, right? A better and richer communication experience.

1:14:43 This one necessarily, isn't necessarily faster, but it's something more, something we didn't have before. So it ticked three of those boxes, um, and it absolutely landed. And, uh, just as a bonus, e- everyone who works with me on, uh, Devs on Stage knows this, you know, um, I, I grew up in a, a, a world where you'd, you'd say, you know, you gotta have three points in a poem.

1:15:05 Every good talk should have three points in a poem. Uh, do… There's this rule of threes. It's just a human thing. So I, unless there's a really good reason not to, I tend to say you should have three things, three beats, three story points. Uh, and usually a good demo, a good talk, kind of like a pop song, should kind of have like a good, better, best.

1:15:23 Uh, ba, ba, ba. Uh, bronze, silver, gold.

1:15:27 So when we did this demo and structured it, the kind of good was this bar chart that I could interact with just by moving my hand around in the air, and it could read my fingertip and point at things. And then the better was this, uh, scatter plot, actually. It was on the next page of the next scene of the demo. The scatter plot, interactive scatter plot.

1:15:46 So now we're go- the mental model is gradually expanding with me. This is all something new. We've never seen this before. First, we see the bar chart. Okay, I get that. I see it, it interacts. It highlights when you point at it. Oh, now we see a scatter plot. We know what that is. Oh, they're linked together. That's something new. Their mental model is expanding.

1:16:01 And then the wow, and you gotta watch the video. It's… I mean, as a presenter, it was one of the coolest moments ever, because I knew what was coming, and nobody else did. And when that globe spun, the 3D globe spun, the audience, it was incredible. A great moment.

1:16:15 Uh, and, and I gotta credit the people like Francois who, who suggested that kind of a structure. Like, don't just nerd out about, "Look, cool, I got a webcam chart." But anchor it on problems people actually have. Yes, this is cool, but why?

1:16:29 Uh, it's we've lost something human. What is the solution?

1:16:32 It's Tableau Gestures. It's bringing data into virtual meetings like this. What does it do for us? It's easier, it's better, it's more than.

1:16:41 Now let me show you. Ba, ba, ba. Three beats.

1:16:45 There you go. All right.

1:16:46 I told you just now you should have three beats, and, uh, I'm gonna break my own rule and we have a fourth bonus beat, and that is embrace the change. Um, I, we're, I'll, I swear I'll stop and we can get to Q&A. Uh, I love this photo. It's my favorite thing to work into keynotes. This is my daughter Estelle when we lived in the Netherlands at a place called Pony Park City.

1:17:05 Now, what I love about this is in a time of intense change, and this is a season of intense change, technology evolving every week, it seems there's something new. Uh, the question I have for you, for me, is, uh, are you Estelle, my daughter here on the right, or are you the pony?

1:17:21 And it may be that you are Estelle, and your stakeholders, your manager, your company are the pony. It may be that AI and this new way of thinking, moving away from purely drag and drop, purely passive, maybe all interactive analytics, uh, that's, that AI world is Estelle, and you're the pony. Maybe.

1:17:44 Maybe it's your corporate culture itself that is the pony. But, uh, regardless, uh, of which one it is, uh, I encourage you to be Estelle and not the pony, and embrace the change of this wild new world. I hope I've given you something of value. Thank you for a fantastic birthday present of telling some of my favorite stories and, uh, and spending it with great people.

1:18:05 Celia Fryar: Yeah, thank you everyone-

1:18:05 Matthew Miller: Now we got some time for Q&A

1:18:07 Celia Fryar: … who came. Yes, let's ask. All right, team.

1:18:11 Such good stories, Matthew. Seriously.

1:18:14 Matthew Miller: I hope I lived up to your billing. You gave me very high praise, so I-… hope it, hope it delivered.

1:18:19 Celia Fryar: Greatly appreciate it.

1:18:20 Matthew Miller: All right.

1:18:21 Eric Howard: So I'm a data dev. I want to know how you got the data from your moisture, uh, devices onto cal- the cloud in real time.

1:18:31 Matthew Miller: Yeah. Yep. Uh, uh, I'm on my work laptop, so I won't log into it. But I have a home hub called the Hubitat Elevation. Uh, and it's a ner- it's a nerd tool. I have, I act- I have two home automation hubs. One that's like user-friendly that my family uses, and then I have the nerdier one for me, uh, that you can like put custom JavaScript in. It's called the Hubitat Elevation.

1:18:52 Again, this is not an endorsement by Salesforce or Tableau lawyers.

1:18:56 Uh, but it's, it, it… I like it. It's nerdy. It's about 100 bucks. Stick it under a thing, uh, but I can SSH into it. It's got a decent UI. Uh, and so it's the bridge. It has all the radios you'd want, like Z-Wave, Zigbee, Matter, and then it has a maker API, which is dope. And I created a Google Cloud function that's the receiver. So, uh, it gets thousands of events a day for pennies, and once it's all in BigQuery, you can do whatever you want in Tableau.

1:19:24 Also, hey, Eric, good to see you.

1:19:25 Eric Howard: Yeah. Yeah.

1:19:25 Matthew Miller: I love the stuff you've been posting on LinkedIn, by the way.

1:19:28 Eric Howard: Thank you.

1:19:28 Homer Hegedus: I wanna, I wanna ask a follow-on question. Have you, uh, thought about, uh, making an AI agent that'll actually turn on the zones that need watering based on the moisture level?

1:19:41 Matthew Miller: So, Homer, it's a great question. Uh, you saw my family, but I haven't shown you a recent photo. I have teenagers- And I feel like they gotta be good for something, you know?

1:19:51 So, uh, there are lots of things I could automate by purchasing, um… I've, trust me, I have looked at the wireless, uh, the wireless, um, controlled home lawn stuff and the mower, but I'd rather just, you know, spend my Saturdays being like, "Joshua, can you please go mow the lawn?" I just, I'm, I'm a glutton for punishment, I suppose. Um- It build, it builds character.

1:20:12 When we moved here, I bought a push reel mower, like the old-school push reel mowers.

1:20:16 You know, if you've seen Calvin and Hobbes, I was like, "It'll build character, kid." And to be clear, we don't live on, like, five acres. We live on, like, a postage stamp suburban lot.

1:20:23 Um, but uh, no, we finally bought a proper mower. So no, I, I just go out there and water it when it needs it. Um, it's not that bad. Yeah, I can… I know, I've automated, like, everything else in my house. I have, you know, o- one of my, one of my favorites, you know, my dad would be like, "Shut the door. We can't afford to heat all of Rhode Island," right?

1:20:40 You know, every parent would be like, "Put on a sweater." Uh, and so what I finally did is any time they leave the sliding glass door open for five minutes, it shuts off the HVAC in the house. "Hey, who shut off the AC?" I did. Well, the bot did because you left the door open again, right? You're pumping all the AC out into-

1:20:57 Homer Hegedus: No, you did 'cause you left the door open again.

1:20:59 Matthew Miller: That's it, right? You did it yourself. So the automation's been really good.

1:21:02 Homer Hegedus: Sure. Yeah.

1:21:03 Matthew Miller: Um, uh, we're not supposed to talk about alcoholic beverages as Salesforce employees, so I'll just say, um, in the winter here, it gets really cold, and I get a notification when my fridge is… Sorry, my garage is cold enough to be a fridge, and then I get a notification when it's too cold to be a fridge, which is when you bring the beverages you might put in the garage inside.

1:21:21 Then I get a notification again in the spring when it's w- you know, warm enough to be a fridge again. Then I get a notification when it's too warm to be a fridge.

1:21:28 So for two brief seasons in Illinois, um, my garage is a place to store beverages, and the home automation takes care of it.

1:21:38 Um, yeah, I've been– I find I'm using a lot of this through, uh, MCP. I've got it all wired up through Claude so that I can, uh… And you know, it's, it's, it's funny how like in my own world, the, the usage of like Tableau and the semantic layer and the Pulse, all the story we sort of tell, like it's real. The best answers I get, I get w- trust me, I get way better answers going through Pulse than if I go straight to BigQuery.

1:22:01 The data's a mess in BigQuery. You should see it. Like it's, it's not pivoted. Everything's like attribute value. There's like data errors. The, in the Tableau layer, I was able to clean all that mess up, get to a proper like index humidity percentage, put Pulse on top of it to find the insights. And so I, I have my MCP go through Pulse to the data server, uh, data sources.

1:22:21 I almost never look at the dashboard anymore.

1:22:24 Homer Hegedus: Yeah, that was actually gonna be my next question, if you actually look at the graphs or you just-

1:22:28 Matthew Miller: On the Pulse stuff I do. Yeah, yeah.

1:22:30 Homer Hegedus: Yeah.

1:22:30 Matthew Miller: I get the Pulse digest every day.

1:22:32 Uh, and the one on, um, the, the Pulse metric on my home page, 'cause it's, you know, it's right there, right?

1:22:37 It's, uh, it's glanceable.

1:22:40 And that, you know, what gets measured, get, gets managed. I think we could do more. I've wanted to have an Apple Watch app for a long time. Uh, yeah, while, while we're looking for other questions, I might get myself in trouble and show you guys something pretty cool.

1:22:54 Any other questions?

1:22:54 Homer Hegedus: Oh, by the way, is, is your tree recovered?

1:22:58 Has all this technology-

1:23:00 Matthew Miller: Oh, yeah

1:23:00 Homer Hegedus: … actually solved the problem?

1:23:01 Matthew Miller: You asked that. Right, right, right. That's, that's, that's the… I should have told you. Yes, it's doing much better. Yes.

1:23:04 Homer Hegedus: Yeah.

1:23:05 Matthew Miller: Yeah. No, no surprise keeping it between 40 and 60%. It was just getting– It gets really dry here parts of the year, and it was, it was dying. And so, um, I, I did not intend for this to be a keynote demo. The, the keynote version of it, which I should share with you, is, uh, um… I'll find it.

1:23:24 You, I sh- what I did is I showed how all the other surfaces I use this in and also some that I could use, you know, I showed it like in a Google Slide, in a Google Sheet. I don't really use it there practically, but in like an enterprise setting you would.

1:23:36 Uh, but where this came from was I had started doing this in our, like, executive briefings after I'd done kind of the corporate demos and the corporate, you know, slides and whatnot, trying to humanize them as much as I can within the kind of parameters, you know, of my job.

1:23:49 At the end, I'd be like, "By the way, you know, can I just show you how I'm using this stuff?" People love that.

1:23:55 Like, executives from big banks, nothing wrong with banks, but you know, kind of stuffy people in suits, right? Were like, "Cool, can I see that? Well, what kind of sensor is that?" Uh, and, and so then the marketing team asked me to do that as the keynote demo in Europe, and, um, it was-

1:24:11 Celia Fryar: Can I, can I just ask a question?

1:24:11 Matthew Miller: It was, it was well-received. So I'll s- I'll put a link. I think it's gonna be one of these ones that you gotta request access to, but I can, um, I can grant access. Uh-

1:24:19 Celia Fryar: That'd be great. Thank you.

1:24:21 Matthew Miller: If I can find my way back to the chat. There you go. It's on, it's on Salesforce Plus somewhere as well.

1:24:27 Uh, question from Christian. Matthew, question, people who have never presented in a conference in a talk, what can you say to motivate them to, to do that? Oh my gosh.

1:24:35 Um, I have somewhere a video of m- my, one of my first talks I gave in a communications class in college, and, uh, it's awful. It's so bad. It's so wooden. Like, get up there and do it. You just gotta get at bats.

1:24:53 Uh, none, n- none of the great presenters I know was just born naturally comfortable standing in front of thousands of people. It's a, it's a learned skill, and almost everybody can learn it. Absolutely. I, I believe public speaking's a craft. You can join things like Toastmasters. There are tons of ways to, uh, to grow. There's a book I recommend often called Do You Speak Funny by…

1:25:19 or Do You Talk Funny by David Nihill. He's a comedian, but he does a lot of business presentations, and it's public speaking but the use of humor.

1:25:30 Uh-

1:25:32 Celia Fryar: Mm-hmm.

1:25:32 Matthew Miller: Do I ever use gen AI to provide feedback when I'm practicing? Yeah, that's a great question. So this is, again, not an endorsement, not a Salesforce-endorsed thing, but the best tool I've used for this, uh, and this is also not saying that we are or are not a customer of this company, blah, blah, blah, uh, all disclaimers. Uh, but a, a number of my former colleagues work there, a company called Yoodli.

1:25:56 Uh, this was started by two young entrepreneurs, uh, and their original motivation, it's still the DNA of the company. Yeah, there you go, Do- uh, Do You Talk Funny and, um, Yoodli. Uh, their motivation was, how can we take the kind of coaching like I got from Francois? Like, you know, who wouldn't-

1:26:12 Celia Fryar: Yeah

1:26:12 Matthew Miller: … like, pay to, you know, get time with, like, an executive vice president, a world-class speaker like that? So grateful for that, and many others who've mentored me along the way. Uh, how do you get, like, access to that? And they said, basically, there's… the problem is that it's usually, uh, um, sort of the privileged who get that. The underprivileged, who could benefit most from it, don't.

1:26:33 So they built it as a way, like, hey, AI is, you know, n- unbiased, and, uh, it is brilliant. Uh, again, personal endorsement, not corporate.

1:26:42 Uh, but I love Yoodli.

1:26:45 What you do is you set up a scenario, and you, uh, they have AI, like, it… You don't see, like, a bot, but you hear it talk to you, and it's ridiculously realistic. There's, like, a split-second lag after you s- present, but you can actually share your screen. And so you give it a rubric up front, uh, of like, you know, here's the beats that you wanna hit in a roadmap presentation.

1:27:04 You can define personas. So, um, they would, they'd be mad if they knew this, but I picked some customers who are kind of grumpy, and I codified them as, like, a, an agent. And so with, with a fake name. And so then you've got a meeting. I, I did one of them, one of, one of these, um, like, mock briefings where it's all AI, and, like, colleagues are listening to me do this.

1:27:24 And the AI was like, "So tell me why Power BI, you know, doesn't do this and this." And I gave a good example. And it's like, "Well, I need some data points on how Tableau's better than Power BI." So I gave an example, but I didn't have any data off the top of my head. I just told a story, you know?

1:27:37 And it said, "Yeah, you gave me an anecdote. I asked for data. You gave me an anecdote." Like, oh, man, saw right through me. You know? I didn't have any data handy.

1:27:46 So, uh, yeah, Yoodli's one of the best I've used for prepping, uh, in terms of getting, like, real live feedback. It… If you tell it to be savage, it'll be savage. Uh, and you give, you give it your kind of scoring rubric. Uh, I will also turn on, uh, Claude's spoken dictation thing and just riff. I often do that when I'm mowing the lawn, and I'll just talk for, like, 40 minutes back and forth, uh, and, and then it'll kind of build out a talk for me, uh, in the process.

1:28:14 I, I write all my s- LinkedIn posts by hand, though. I just want you to know that. Always write them by hand. The only thing I'll use AI for is, like, is this too wordy? It's probably too wordy. It's like, yeah, too wordy.

1:28:23 All right. I know we're at time. Any, any last thoughts, questions?

1:28:28 Celia Fryar: I think that's it.

1:28:30 Matthew Miller: Cool. I think-

1:28:30 Celia Fryar: And thank you so much. How do you spell Yoodli, by the way? I'm-

1:28:33 Matthew Miller: It's in the chat there. Um-

1:28:34 Celia Fryar: Okay.

1:28:34 Homer Hegedus: Yeah, Y-O-O-L-

1:28:35 Matthew Miller: Y-O-O-D-

1:28:36 Celia Fryar: O-O. Thank you. Perfect.

1:28:38 Matthew Miller: Yeah.

1:28:38 Celia Fryar: Okay, great.

1:28:38 Matthew Miller: So I promise you-

1:28:39 Homer Hegedus: And just, just to add a comment that, uh, you know, storytelling is tremendously powerful, not just, like, when you're, uh, you know, in front of a professional audience, but if you're an entrepreneur and you're trying to start a company, storytelling is the way you wanna start your pitch every time.

1:29:01 Matthew Miller: Good Don at Greenmor.

1:29:02 I got it from my great-grandfather. He'd lean forward at the dinner table and go, "Matthew, did you ever hear the one about the three-legged pig?" And you had, but you wanted to hear it again. I can't tell it professionally, 'cause it's not terribly correct for these days, but it's very funny.

1:29:15 Celia Fryar: Yeah.

1:29:15 Matthew Miller: So someday at TC, grab me and I'll tell you the story about the three-legged pig. And, like-

1:29:20 Celia Fryar: Great

1:29:20 Matthew Miller: … that's, that's where I got that from. All right.

1:29:22 Celia Fryar: Very good.

1:29:22 Matthew Miller: Thank you all so much. What a great way, what a blessing to me-

1:29:24 Celia Fryar: Yeah. Thank you

1:29:24 Matthew Miller: … to get to spend my, my birthday talking about my favorite stuff. Uh, hope you got something out of it.

1:29:28 Celia Fryar: We did.

1:29:29 Matthew Miller: Have a good one.

1:29:29 Celia Fryar: Thank you. Thank you very much. And happy birthday.

1:29:32 Matthew Miller: Yeah, buddy.

1:29:32 Celia Fryar: All right.

1:29:33 Matthew Miller: Bye.

1:29:33 Celia Fryar: Thank you, everybody, for being here.

1:29:34 See you in a month. Out. Take care.