Dremio Jekyll


Subsurface LIVE Winter 2021

Analytics for Everyone - Unlocking the Full Power of All Your Data

Session Abstract

Francois has spent the last decade as a key driver of product strategy at Tableau, where Billy served on the Board of Directors. In this fireside chat, Billy will explore Francois's key market learnings at one of the fastest-growing BI companies in history. They will discuss some of the twists and turns in the analytics market that have led us to where we are today, and also provide some insights into what analytics trends we can expect in the years to come – as well as how to spot them as they're forming.

Presented By

Francois Ajenstat, Chief Product Officer, Tableau

Francois Ajenstat is Tableau's Chief Product Officer. In that role, Francois is responsible for Tableau's overall product strategy and oversees the product portfolio, including product packaging, pricing and product positioning. He is also responsible for evangelizing Tableau's products with customers and partners and incorporating customer and partner feedback into the strategic vision of the Tableau portfolio. Francois brings a wealth of product management experience in the business intelligence (BI) industry. Prior to joining Tableau, Francois worked at Microsoft for 10 years in a number of different groups including SQL Server, Office and Trustworthy Computing. Before that, he worked for Cognos Corporation (acquired by IBM) leading strategic alliances with key industry partners such as IBM, HP and Microsoft.

The first time Francois saw Tableau, over 10 years ago, he recognized the potential to transform the industry and revolutionize how people work with data. "I knew I needed to be part of what was to become an entire new era in business analytics and it's been incredible to be part of the team that is making that vision a reality."


Billy Bosworth, Board Member at Tableau, (2015-2019) Board Member at TransUnion, CEO, Dremio

Billy Bosworth joined Dremio as CEO in February 2020 after serving on its board of directors since June 2019. Prior to Dremio, he served on the board of directors at Tableau Software (NYSE:DATA) for nearly five years, through their acquisition by Salesforce.com (NYSE:CRM) — one of the biggest software acquisitions in history. Billy also served as CEO of DataStax from 2011-2019, propelling the company’s growth from sub-$1M in revenue and 25 employees to more than $100 million in revenue. He is a frequent speaker on the topics of data autonomy, hybrid cloud, AI and distributed workforces and was an active member of the Forbes Technology Council for many years.


Webinar Transcript

Speaker 1:

Please welcome back Tableau board member from 2015 to 2019 and Dremio CEO, Billy Bosworth.

Billy Bosworth:

Well, hello everybody. If you caught my intro, then it’s nice to see you again, but if you’re just joining us, thanks for attending today’s session. I have with me my friend, Francois, [00:00:30] and Francois, it’s lovely to see you in a non-board meeting setting where we don’t have the pressure of all the board agenda ahead of us. But Francois comes to us from Tableau where he has been an instrumental part of that company’s explosive and phenomenal growth over the past 10 years. I think one thing you’ll find about Tableau is anybody who’s ever been associated with it comes away saying there’s something special about Tableau and Francois is actually one of those key people [00:01:00] who’s made that possible. So Francois, welcome, it’s great to have you on with me virtually today.

Francois:

Well, thanks, Billy. It’s great to be at this event and be here with you. I miss those boardroom times, I actually want us to be in person, but I think this will do in our virtual fire chat together.

Billy Bosworth:

I agree. It’s better than nothing, and it’s really nice to be able to see you again live and have this chat. So, Francois, let’s dive in. The essence of this conversation is you’ve had [00:01:30] such an amazing career from a learning standpoint and a perspective standpoint, and what we’re going to do today for our audience benefit is, we’re going to walk through and see what we can learn from the last 10 years, and then how do we apply that to potentially what’s coming in the next 10 years. And for you as an audience member, the whole point of this is to be able to see what trends we can pick up on to make sure that you’re in the right vein of technology shift as we go forward. So before we get into that [00:02:00] though, Francois, one of the nice things about these fireside chats is that it’s a little more human, a little more personal. So why don’t you spend a few minutes and just tell us a little bit about you, where are you calling in from? Tell us a little bit about yourself and a little bit about your background.

Francois:

Sure. Well, I’m calling in from a sunny, or actually very gray Seattle, and I’m the chief product officer for Tableau. So I’m responsible for product strategy, product direction, and I like to say ultimately making sure we build great products for our customers. [00:02:30] I’ve been at Tableau now for 10 years, actually over 10 years now, there was only a hundred people when I joined the company and we grew it into this incredible, incredible company. And then prior to that I was at Microsoft for 10 years, in the Excel team, in the SQL Server team, and Cognos before that. So just a long time in the data and analytics space. I just love this space, and I love all the innovation that’s going on right now in this industry. [00:03:00] And so, the last 10 years have been phenomenal, I think the next 10 years are going to be even more incredible.

Billy Bosworth:

Yeah. I’ve been around this industry 30 years and it never ceases to amaze me just when I think maybe we’ve hit a steady state of, especially with data, when we’ve hit a steady state there’s always some new, big things right on the horizon. We have a very mixed audience today. We have an audience from, about half of our attendees are coming from [00:03:30] large companies of over 5,000, but then we have half that are coming from companies as small as a few dozen people, and everywhere in between. And so one interesting thing, you said Tableau was a hundred people. And before that you were at Microsoft with a hundred thousand people, or a hundred million over at Microsoft, as it seems like. What drew you to a company that small coming from a company that big? What was it about Tableau that you saw that became that interesting that [00:04:00] you were willing to make a career bet on it?

Francois:

Yeah, it was definitely a drastic shift from the size of Microsoft to the size of Tableau, but I’m a hands-on person, I love product, I love getting my hands dirty, and I always like to look at what’s in the industry and looking at what is getting people excited? And I remember back then, I wasn’t at Tableau, I was at Microsoft, and I downloaded a version of Tableau and I just installed it on my laptop, and I got some data for my wife and I started [00:04:30] playing with it to try to understand what this product was all about, and I had my list of requirements. Okay. It did this thing, it did this thing, it did this thing.

After a little while I put that sheet to the side and I spent a couple hours just having fun, playing with the data, understanding new things and really getting insights that I wasn’t able to get before. And that’s really what drew me to Tableau, is being able to transfer the power of data [00:05:00] with, obviously in that case, the product Tableau, but transfer that power of data to even more organizations. And for me, that moment was so special and I fell in love, and that’s really where I put everything aside and I said, great, we’re going to go and make this big and have some fun.

Billy Bosworth:

So people like you were interested in unlocking the power of data before unlocking the power of data was cool, right? Clearly everybody knows today that, that’s the [00:05:30] key to all business success, all companies have to be data companies. When you think about now more of an industry question, 10 years ago, very different world, very different technologies existed back then, what did you see as the biggest challenges then when you were facing this issue of how do I unlock the power of this data? What obstacles were in our way 10 years ago?

Francois:

Well, it’s interesting, when you think back to 10 years ago, the world [00:06:00] was quite different, right? The BI and data world was predominantly an IT tool set, and it was IT that were producing reports for the business users. And it wasn’t really self-service, that wasn’t a concept back then. And so people felt disempowered, they just exported the data to Excel and they spent their time trying any way they could to unlock the value of that data. And so there was this shift from IT led [00:06:30] to really business user led. But the other thing that was interesting back then is the number of databases were quite small. It was pretty much Oracle, IBM, SQL Server, there’s a handful that people really adopted. Hadoop wasn’t mainstream by then, I think Hortonworks was invented, or was founded the year after that.

And so the big data hype exploded. The cloud world exploded. The number of sources [00:07:00] exploded. We didn’t talk about public cloud 10 years ago. Now it’s the default for everyone. And so it was a really, really different world back then. But the thing that has really come out in those 10 years is that data’s not a nice to have, data’s critical, it’s a must have. Companies that are data-driven are more successful companies. And especially in these COVID times where we’re [00:07:30] all digital, work from anywhere, having access to data and having the agility and the speed is the difference between thriving and barely surviving. And that is so apparent today were 10 years ago maybe people didn’t think of it in the same way.

Billy Bosworth:

Yeah. It’s the lifeblood now. And as I said, everybody agrees on that now. Interestingly one of the [00:08:00] side effects of the shift that you mentioned from the centralized team to the BU was, the first part of my career I was an application developer that lived inside of a BU, inside of a large organization. And so I was one of the people creating the problem. It, depending on your perspective of all of those exports, and then I would put them in an access database, or just a CSV file for Excel for my users. But then when I moved into management I realized one of the challenges was I walked into a meeting one time with my [00:08:30] CEO when I first was moving into executive management, and I was all proud and excited sharing all my data about how the business was doing, but then the CFO chimed in and said, that’s not the data I have. I’m showing you’re growing by this percent, not that percent.

And the CEO just went off the rails. It was like, okay, we’re done with this meeting. If you two can’t agree on the data, I was like, that caught me totally off guard. I was all empowered because I had my BU data, but it didn’t agree with the IT data that our CEO had. And so [00:09:00] I think that was one of the challenges in that shift. And I think it’s always a pendulum, right? It’s always working its way back and forth. So those were some of the challenges. What did you see as a result in the change area? What has changed the most to you in the past 10 years as we overcame those obstacles?

Francois:

I think that the most obvious change is that the concept of self-service, the concept of data is no longer just this geeky thing on [00:09:30] the side, it is mainstream, it is something that everybody talks about, just think about COVID, we all look at the numbers every day, we see whether we’re flattening the curve or not, that’s data, that’s visualizations. And we want to lean into it. We want to see how it’s relevant to me and my County, or for my age group and et cetera. And now we’re talking about vaccine distribution, it’s all about data. So it is everywhere. It’s all around us. You said [00:10:00] it right, where it’s a pendulum shift. We started from an IT led thing, and then the pendulum shifted to self-service business user. The reality is there needs to be more of a balance between the two. It’s really about self-service with governance. It’s about empowerment with control, and you need a little bit of both to be successful so that you don’t end up in the scenarios where people are fighting over what the data means.

Sometimes actually that fight, that conversation, is actually pretty [00:10:30] healthy, because it shows maybe the people that are setting up the data don’t have a full grasp of the business requirement. And so you need that communication element around it. And these challenges become even more amplified as, if you think about it five years ago, it was all about the rise of big data. Now it’s all about AI and how AI is changing everything, and absolutely it is, but it puts new constraints on the organization to be able to have access [00:11:00] to more data, to be able to trust it, to be able to have that speed and agility to deliver those insights. And that is really the thing that now that people are empowered, you need this collaboration between IT and the business, right? Because data just for data’s sake, it’s useless, right? You have to apply it to the use cases that you need. And it’s that connection between the two worlds that becomes really transformative to the company.

Billy Bosworth:

So you mentioned AI, and it’s hard to talk [00:11:30] about data today without talking about AI. And when we think about the last 10 years, that very nerdy term was always around, it’s been talked about for decades and decades, I think much longer than anybody would probably imagine, but it’s become mainstream, much of the data conversation in the world, right? This has now become very mainstream. And so what do you think is going on with AI [00:12:00] today that’s different from where we were about 10 years ago? Let’s take that same timeframe. What do you think some of the changes around that space have been?

Francois:

There’s so much. It is interesting to see how an older technology really gets a new light in the world of the cloud where compute is basically unlimited and we’re able to do so much more. I remember when I was a kid, my dad was actually a professor of expert [00:12:30] systems and data mining in university. It’s just the old form of AI today. But what’s really, really fascinating is in the world we used to really look at what happened, and with a lot of the technologies, we want to be more proactive. So we’re moving into why did it happen? What might happen, and what should I do about it? That’s the nirvana state where you can start automating business processes based on AI. [00:13:00] The only way you can do that is that the technology has to be really fast, the models have to be good, and you have to bring that across the board.

But I think that AI is really becoming, not just this side technology, it’s going to be infused in everything that we do. It’s going to make our products smarter. You have AI in your iPhone, for instance, right? That provides recommendations. You’ve got models that data scientists are able to run and do lots of experiments. But I think [00:13:30] that you also need to move from, if you think in the BI space, we went from a report factory to self-service, right? In the AI space we’re falling in that same trap where we’re going into a model factory and people are drowning in the number of change requests of the models, and we have to really find a way to democratize those models and empower people to make that work. And again, that only works if you have access to the data, [00:14:00] it’s fast enough to compute, right? And it’s accessible to a broader population of people that maybe don’t know how to code, or don’t know all the details. And it’s that explainable AI concept is really becoming much more mainstream.

Billy Bosworth:

And it’s making its way into public policy, isn’t it?

Francois:

Absolutely. There’s a lot of ethical concerns around AI. There’s privacy concerns. And so you really have to think about the ethics [00:14:30] around what we do, right? And make sure that you’re not biasing the models and refining the models. And so there’s a lot of considerations when you put that in place.

Billy Bosworth:

One of the things we talk about at Dremio is this notion of, we call it the data paradox, and the data paradox is this notion that as the insatiable demand for data escalates exponentially, the desire to democratize the data also [00:15:00] escalates, people want to put the data more out there. I’ve always viewed Tableau as one of the early pioneers in the democratization movement. We hear that a lot though. I wonder if you could help us understand what does that actually mean? It’s one of those words we hear a lot like AI and big data, and now another popular one is we’re going to democratize our data, but what does that actually look like for our practitioners? What should they be thinking about when they hear that?

Francois:

I agree [00:15:30] with you, the word democratization is a little bit overused these days, but really the essence of it is that we’re making it more easily accessible to broader population of people. Not everybody in the organization will be an analyst or a data scientist. A lot of people are just users. I’m a salesperson, I’m a customer service agent, I need data in my workflow, right? I have different sets of skills than the other folks. And so [00:16:00] the idea of democratization is really how do you broaden that access and make it accessible to everyone?

The dream has always been, can data analytics be as simple to use as Excel, or as pervasive as Excel. And you go back in time, right? In the nineties, nobody thought Excel would be commonplace, would be on everybody’s desktop. It was the tool for finance and maybe for the stats [00:16:30] geeks. Now we all use it, right? For lots of different things. That’s the dream. We are not there. There’s a lot that we need to do in terms of improving data literacy, improving data access, ensuring there’s trust in the data, and there’s also fundamentals around ease of use that needs to be put in place where I believe that we need a 10 to a hundred X improvement in ease of use and access of this data in order [00:17:00] for it to really be pervasive. So it’s a pretty bold statement to democratize it, but I think that’s where we really need to get to.

Billy Bosworth:

It introduces an immediate tension, doesn’t it? As soon as you open it up to the world, in your universe, wherever that is in your company, you’re going to get hammered with now all the privacy compliance, and all the things like GDPR that I believe is coming to the rest of the world in various forms and fashions. What recommendation do you make to people who [00:17:30] are embarking on these new projects of building, let’s say data lakes to democratize this data, how do they live in that tension of the control versus the accessibility?

Francois:

So when we were in person I would draw this grid, right? This quadrant for people to have this conversation. And you can think of in one access maybe its breadth of impact, so the number of people that have access to it, versus a risk of that impact. [00:18:00] So for instance, data that I’ll report to the SEC, upper right quadrant, right? Super, super important. I’ll go to jail if I get it wrong. Marketing campaign data, maybe bottom left, if the data’s not precise exactly, it’s directional, it’s good enough. And so you have to really think and adapt to the different use cases and apply the right governance structures accordingly.

And so sometimes we think purely about the risk [00:18:30] and the things that will go wrong, but not everything needs the same light. And so what I think about a lot with customers is how do you use agile methodologies for data rather than just waterfall. If everything’s waterfall, right? And I use those terms explicitly because they have long lead times, right? They require a lot of governance and control. What if you were more agile, and you’re building certain things with the business and you have that speed to unlock the data, [00:19:00] bring it to the users, understand the use cases and iterate as fast as possible. That’s really where we need to get to, but it is this balance and really understanding what the requirements are.

Billy Bosworth:

So you by definition, almost as a Tableau perspective, have to be quite agnostic to the architectures against which Tableau is operating, right? Whether it’s a data warehouse, or a classic database, or a data lake, you do have a certain agnosticism [00:19:30] about that, but I’m wondering what is your view, maybe this is more Francois versus Tableau head of product, but what is your view of what’s happening with these data lake architectures? This conference is centered around some pretty revolutionary changes in how we’re thinking about the data workflows. What is your take on data lakes and what’s happening with these architectures?

Francois:

This is another area that’s getting so much innovation and revolution for customers, and I think this is a huge opportunity. [00:20:00] What we always think about is, number one, do you have access to the data? How do you make as much data accessible to as many people as possible? If you have to keep creating copies of the data, that just increases cost, time and complexity. So, data accessibility, availability becomes really, really important. The second big thing is speed, right? The fact that you have data in a data lake is great because you can unlock use cases, but can [00:20:30] you make it fast? Can you make it so that people want to bring in more data, access it really, really quickly.

And number two is we always have to be, number three, I should say, you always have to be careful that a data lake doesn’t turn into data swamp. And this is why I think actually semantic layers, right? Having good cataloging capabilities on top of that data becomes so critical to help people decipher what is in there. And I think that sometimes we think so much of the storage part of look, I’ve [00:21:00] got data in the lake, look at all the things I can do. It’s only one piece of the equation having speed, broad access, and that common understanding is, I think the linchpin to bring that forward.

Billy Bosworth:

And where do you think we are today in that journey? I get asked this question all the time, they’re like, Billy, where do you think the market is in this evolution to the type of data lake architecture that you just described? Where are we in that journey as an industry?

Francois:

I [00:21:30] think we’re still in the early days, quite frankly. People are still designing data pipelines and data architectures like they did 10 years ago, right? Where data just moves and it goes into warehouse, and you create your pipelines in that way. And I think that we have this opportunity to make the data lake the system, right? Of analytics, the system of insight, where you get the best of both worlds, right? You get all the advantages of the data lake with the [00:22:00] benefits of the data warehouse that’s optimized for analytics. And that’s really, I think that key to bring these things together, instead of creating more silos of data, more duplication of data, how do you use the data lake more strategically, so it unlocks all the scenarios without, I like to say avoiding the tyranny of or, it’s not one or the other, how do you get an and? And that I think is where we’re seeing innovations from companies like Dremio [00:22:30] really enabling that power of and, where I get the power of the data lake and the power of analytics and data warehouses combined into one.

Billy Bosworth:

I agree with the and concept and just that whole mindset, because the reality is I think that in our technology world to believe that one architecture is going to become dominant in the enterprise infrastructure, I think is just unrealistic. I know [00:23:00] from our point of view at Tableau we talked about that all the time, was like the power of Tableau was the ability to work with all this stuff, but the question is, how fast can you get these new technologies adopted to start showing their worth? And on that end, we’re going to connect the dots to the beginning of the conversation here. And I want you to think about somewhere in our audience right now, there is a Francois 10 years ago, right? There is somebody [crosstalk 00:23:27].

Francois:

That’s really scary [crosstalk 00:23:27].

Billy Bosworth:

It’s a little scary I know, but they’re [00:23:30] out. And they’re trying to figure out what you saw 10 years ago, but they’re trying to see what’s ahead. What’s coming in the next 10 years? And so passing along your experience to our audience, how do they read the tea leaves? How do they, with this onslaught of technology, and this onslaught of innovation that you talked about, it can be a little daunting, right? I think you and I both know that. What’s your advice to people on how to get themselves very well positioned [00:24:00] to take advantage of what are going to be the most exciting impactful trends of the next 10 years?

Francois:

If I had the way of really predicting the future, I would be rich living on a yacht somewhere. That would be amazing. But I think that there’s a way of thinking through things. Number one is, you got to be curious. You want to be able to learn, stay on top of the technology trends, see what’s going on, look at what the startups are doing, and [00:24:30] where the technologies are, because that’s how you learn, that’s how you try new things, and that’s how you discover the little glimpses of where the world’s going to be. And sometimes part of that learning means that we have to unlearn, right? And I like to use the term having a beginner’s mindset. Don’t be shackled by the things we’ve done in the past. Oh, that’ll never work, right? That’s not going [00:25:00] to help you move forward.

Just have that beginner’s mindset. Think about what that would mean if you were brand new to that thinking. And again, sometimes experience is a wonderful thing, and it’s a great thing to do, but just asking the questions, understanding what it does, is really good. Another aspect is, you have to be agile and it means you can’t wait for perfection, because in the early days [00:25:30] as the technology evolves, of course, technology that’s been around for a longer time is going to be more mature, it’s going to have more other things, but what can you take from those seeds and help them grow, help them blossom. And that’s where I think agility and experimentation really come a long way, trying different things [crosstalk 00:25:50].

Billy Bosworth:

I’m laughing Francois, I’m sorry to interrupt, because I remember in our final board dinner that we had, you guys were going around like we’re accounting the early days of Tableau and [00:26:00] how raw the technology was, and how hard it was, and it was a great reminder of what you’re saying right there about be adaptable to, it’s going to have some rough edges on it, right? These new technologies are going to have tough edges, sorry to interrupt, but that was such a vivid memory of how hard it is in the early days.

Francois:

Yeah. But the thing that you have to get to is what’s the raw essence? What’s the core unique value proposition, key differentiator, of the thing that you’re [00:26:30] looking at? Because if you grok that, okay, how does that bring 10 to a hundred X improvement over what was the technology of the past?

Billy Bosworth:

Right.

Francois:

That’s really where, if I go back to Tableau 10 years ago, the technology was raw. It did one thing really, really well, in fact, so much better than anything else, did a lot of things extremely poorly. And as the technology evolved, it grew, but that one thing was so transformative to people’s [00:27:00] careers, to companies, that it propelled our success. And that’s the same thing on technology is, does that one thing, and I say one thing, because there is always this core value proposition, this core differentiator, how does that really provide so much more superpowers than what was the technology of the past?

Billy Bosworth:

I love that. I love the beginner’s mindset. We talk a lot about in the industry about growth mindset, which is very similar, but there’s a certain [00:27:30] humility about a beginner’s mindset to say that no matter how long you’ve been doing this, come with a fresh set of eyes, I think is what you’re saying. Right?

Francois:

That’s right.

Billy Bosworth:

And, boy, we have so many more learning opportunities today, don’t we?

Francois:

Oh yes we do. This is one of the reasons I love conferences like this one, because you get to spend the time to invest on yourself. You get to spend the time to discover new things. And it’s all just a click away, right? Remote learning has never been easier [00:28:00] because we’re all digital work from anywhere, but we actually learn from anywhere in the world as well. And just like you were asking me for the Francois emerging, and what does a future product leader, or a future leader look like, you can’t stop learning. You have to keep being curious. You have to keep exploring. You have to keep trying new things, and being open to discoveries. And that to me is the [00:28:30] thing that, as much as I hate the situation of COVID that we’re in, it’s providing more opportunities to do things like this that maybe we wouldn’t do because it would require travel, it would require a lot of costs. This is a great way of spending the time, learning, and discovering new things that you didn’t know existed. And so I really think this is a good opportunity.

Billy Bosworth:

Yeah. To think about this conference and what it would have taken to do two [00:29:00] of these in the six-month period for 10,000 new people, just think about that. The cost and the logistics that would have been, well, you know, you’ve been part of so many a TCs, Tableau conferences, that we know are monumental.

Francois:

And by the way, there’s learnings for you as the individual, but when you find that nugget of insight, you can share it with 10 other people just with a link. And [crosstalk 00:29:24].

Billy Bosworth:

That’s right.

Francois:

… other people on the journey, and get other people to discuss and learn together, [00:29:30] and that’s, it’s a really, really great opportunity that we didn’t even have as much in person, we had a different connection and you could learn from others in different ways, but this is really a great opportunity for this event.

Billy Bosworth:

Well, Francois, you don’t have a crystal ball, as you said, but I tell you what, you have done pretty amazing with the skills and tools and the intuitions that you’ve had. You’ve had an amazing career. I think the best is still ahead for you. You know I’m a [00:30:00] big fan of you and Tableau, and always will be. I love our partnership at Dremio, I think it’s amazing we get to work together again in a professional capacity as well, but thank you for these insights. This was really, I think, at a personal level and our audience I know is going to benefit from your learnings. And so Francois, thank you so much for your time. I’m sure the audience appreciates it. And I look forward to continuing our relationship into the future.

Francois:

Same here. Thank you so much. Love the partnership, love the opportunities, [00:30:30] and have a great conference everyone.

Billy Bosworth:

Bye-bye.