Tableau CEO: We’re the Google of data visualization

There’s a reason big companies put business intelligence at or near the top their technology priorities year after year. Meaningful business decisions demand the support of business analytics — and the visualization of trends, patterns, and data relationships provide insight you can’t get any other way.

Christian Chabot is on a mission to make that insight vastly more accessible and interactive. As CEO of Tableau Software, he aims to wrest business intelligence from the grip of specialists and hand it to a broad swath of business users, providing the tools to tap data stores and explore them visually by clicking and dragging instead of entering complex queries.

Chabot maintains that Tableau is “the fastest-growing company in software worldwide” and has been open about his aspirations to take his firm public. The company claims it will see more than $100 million in revenue this year and boasts thousands of customers, including such marquee names as eBay, Electronic Arts, Facebook, LinkedIn, and Wal-Mart. According to a company representative, the following interview with Chabot, which appears in edited form, may be one of the last pre-IPO opportunities for him to speak publicly about the company’s goals.

Although Chabot began his career in a consulting firm where he served as a “bona fide data geek,” he considers himself “an entrepreneur by identity.” His first startup, a digital mapping pioneer called BeeLine Software, was sold to Vicinity Corporation, which was in turn bought by Microsoft. He considers Tableau “the culmination of everything I’ve done.”

We began the interview by focusing on what Chabot calls the “the moment of invention” when Tableau co-founders Pat Hanrahan (a founding employee of Pixar) and Chris Stolte came up with the idea fusing graphics and databases.

Q: You have equated Tableau’s invention of visualization on top of databases with Dan Bricklin’s invention of the spreadsheet. Why do you think it ranks with that?

A: Well, we didn’t invent visualization. We invented visual analytics. It’s been true virtually since the dawn of computers, that you could take some set of data, like — here are my nine points, I want to chart them and put them into some procedural program, usually called the Chart Wizard, and pop out a visualization. That’s decades old. We didn’t invent that.

What we invented was a visual query language that actually lets you interrogate data using the picture. I can zoom and filter and sort and group and dive and compare and rotate and flip databases with a visual canvas.

That’s the key ability — for people to have a blank canvas and just see the schemas of the data they’re connected to and just throw the items they are curious about on this canvas and get instant visualizations and generate the query needed. It’s a revolution in thinking with data.

Q: That visual querying must require its own syntax. Is there a learning curve to that?

A: There’s a bit of a learning curve at the top, of course. But the breakthrough technology Pat and Chris came up with is called VizQL (Visual Query Language), which isn’t exposed to end-users. That is our internal language that lets us bridge the world of a simple user interface with serious high-performance database queries against the world’s biggest databases.

Q: And what you call your Data Engine gives you that high performance?

A: Yes, we have this really clever engine that, as you’re doing visual analytics — as you’re doing this process of zooming and sorting and grouping and filtering and comparing — automatically generates queries in the background.

We have this little setting that you can click. It says: Hey, do you want these queries sent straight to the database? Or do you want to actually go pull out a whole bunch of the data from the database into memory first so that you don’t bother the database system?

In our industry there’s been a religious war about which way you should do this. The right answer is both. It depends on your situation. Some customers have Teradata access and you’re darn right they want those queries going straight to data. That’s why they pay Teradata $10 million a year, because it performs. They have literally billions and billions of rows sitting in Teradata on big servers, with a stack surrounding it, and it’s all indexed.

Now other groups within the same companies might say — oh well, our data is in the SQL server database that hasn’t been upgraded in three years, and I think it’s really maddening to query it because it brings it to its knees. We say — great. We’ve built this in-memory data engine, so basically on a bunch of little commodity chips you can pull billions of rows into memory on a laptop that you got from Dell’s store.

Then we have customers who combine both. Sometimes, like at eBay, they’ll see major auction trends across eBay broken down into colored lines … then they say — oh, you want to overlay our conversion ratio from that one campaign we did? That’s over here. And they’ll pull it in memory and they’ll overlay those numbers graphically right on the Teradata numbers so you can even combine in-memory and live queries. That’s what is making Tableau a real Swiss Army knife for companies. That’s why sales are exploding.

Q: Conventionally, business intelligence has been like the waterfall development process. You send your requirements to business intelligence specialists, who generate a report that may or may not be what you want. So you’re talking then about eliminating the middleman and creating a direct relationship with the data, right?

A: That’s exactly right.

Q: So then who do you sell to? You’re probably not selling to IT people or the conventional BI guys. Are you selling exclusively to the business side that wants the answers?

A: It’s about half and half. You’d be surprised. In the early days of the company, it was always from the business side. It was the business owners saying, “I am fed up with our BI strategy. This is ridiculous. I’m going to go find something I can just go use and get this done today rather than wait until next February to do my conversion analysis.” I can walk the hallways of literally any Fortune 500 company and find rampant frustration with their current BI strategies. People are tearing their hair out.

Q: It’s near the top of the spending priority list year after year.

A: Exactly. And for the first time in Gartner history, they have gotten a different answer to one of their classic question: What’s the most important criteria in your business intelligence buying decision in the coming 12 months? Last year, for the first time, No. 1 was ease-of-use, hands down. Give me something my people can use because I am fed up!

Back to your question: IT is actually the buyer about half the time, because for most IT organizations processing change requests for business users isn’t the fun part of the job. That’s not what really gets them excited. Right now a lot of people in IT live in a world where … the board meeting is tomorrow, so the day before there will be an unending stream of follow-on questions from the VP of marketing for her slides. IT does not enjoy that.

With Tableau, IT can say: Look here, answer as many questions as you want. If you have any problems let me know, but my team has set this up. Knock yourself out.

Q: Self-service, right? It’s the consumerization of IT.

A: Right. And I think people who don’t really understand what’s happening are thinking: Oh, this is the business picking their own thing. It actually was like that seven or eight years ago, but it’s just not like that anymore. IT is increasingly driving these things because they’re trying to move their strategies toward more self-service.

Q: Let’s talk a little bit more about your business. What’s your typical sell cycle like into an organization?

A: We use a land-and-expand model. That means customers of all different functions and groups are usually trying our products, our full-function free trials online, then adopting them initially for some group or project, having some success, and expanding from there. And we embrace that. I mean we’re obviously trying to reinvent business analytics technology, but we’re also trying to reinvent business intelligence selling and the business intelligence customer experience.

Before Tableau, where could you get a full-function free trial of a business intelligence platform from the home page of a website? We invented that. Before Tableau, where could you get free corporate training videos for a BI platform as free YouTube videos? We invented that. What other BI product has a price list that fits on one line? We invented that.

Q: I would imagine that the incumbents would come back and say — oh, they’re not really business intelligence, you know, it’s just this lightweight thing. It’s just presentation-layer stuff, there’s no real analytics. Do you get that kind of pushback?

A: That’s what they say every day and our sales are exploding. I think that’s what Siebel said about Salesforce, isn’t it? We’ve seen this movie before.

Q: If you talk to the business analytics folks they’ll say, well, there’s the database side of the stuff, there’s the presentation side of the stuff, but the real nut is in proprietary analytics — and high-end predictive analytics.

A: We just launched some new predictive analytics capabilities, so we are now playing in that arena. But playing in that arena doesn’t mean go add the nth algorithm. I mean how long has SQL Server had its data mining package available? And what’s the customer adoption of it? I’ve literally never met a user of that product in my life. In fact, if you ask Microsoft they’ll tell you — sales are virtually zero. The last thing people want is: Give me 500 algorithms only data scientists can understand. Because my data scientist req has been unfilled for three years.

At Tableau we’re really inspired by Google. When you sit down at a Google search engine, you have a question every time. How do I drive to my son’s soccer game? Who wrote “Moby Dick”? What time is the debate? You have some question.

So what is the Google for databases? The answer is Tableau. Because when you sit down with a database you have a question. And that question isn’t some crazy data mining thing; it’s usually basic things.

Q: So you can distribute this capability throughout your customers’ companies. Your model facilitates that?

A: Even the full Tableau Desktop version can be used by virtually anyone. We use Tableau within Tableau, as you can imagine. Every department uses it. Our sales team is over 300 people. Every single one of them uses Tableau Desktop to dive into their performance and their accounts.

Q: But there’s still room for the data scientists, right?

A: Yeah, but I think honestly all the excitement about the data scientists is in the wrong direction. We already had data scientists. That was the problem: We had this small priesthood of people where everything gets slowed down because of the resource constraint. So I’m telling you 18 months from today talk of data scientists will be gone. It’s totally the wrong thing to be focused on. The thing to be focused on is empowering your people, not your data scientist.

Q: It’s always been a two-pronged thing. The data scientist is the one who knows the tooling and the data analyst is the one who knows the data. Right?

A: Tableau is for people who understand data. And what a total absurdity that the people who understand data largely cannot operate the systems that help them analyze it.

Q: Why aren’t the incumbent players going to come after you and trump you?

A: It depends on the company. Most of our partners, such as Teradata, are not interested in going in our direction. That’s just not what they do. For the Oracles and Microsofts, it’s the innovator’s dilemma 50 to 70 percent of the revenue of every single one of those major companies is from the maintenance and services stream on the old school stack that requires all the development. They’re in checkmate. They can’t cannibalize it.

Q: What does “big data” mean to you within the context of what you’re doing with your company? Is it even a meaningful phrase?

A: Sure, I think it is. At any given time you might call big data the datasets and data repositories that are not easily accommodated by the dominant database technology of the time. The dominant database technology of our time is the relational database, and it’s not accommodating these massive data collections.

A new generation of tools has grown up to try to better accommodate that data expansion. And Tableau, as you can imagine, has run into this ecosystem and embraced it. We want to actively partner with everyone doing that and be the new standard for business analytics.

The reason Tableau is the fastest-growing company in software worldwide is because the current generation of database and BI technologies is too hard to use. Enter big data. Big data databases and big data programs are way more development-intensive than even Business Objects ever was.

Q: You’re talking about Hadoop, right?

A: Hadoop is a programming platform. So in a big data context, Tableau is even more relevant. It’s really been a tailwind for us.

Q: In the case of Cloudera, they have no interest in playing at the application level. They’re all about having the best Hadoop distro and the best Hadoop management tools. They’re not interested in moving their way up the stack.

A: It worked for Teradata. It worked for Informatica. I think it’s a brilliant strategy.

Q: What are your ambitions for the company? You have been open about your aspirations to take the company public. I was told this might be one of the last interviews I could get in advance of that.

A: I’ll give you the big picture. Two years ago, the company was 195 people. Starting this year, we were 350 people. And today we’re 740. Our revenue has grown accordingly, and although I can’t disclose the number right now, I can say we’ve broken through the $100 million revenue mark this year.

So 740 people, 100 percent year-over-year growth, breaking through the $100 million barrier, and opening offices all over the world. We’re up to about six offices now in geographic expansion. We believe things are prime for disruption and for a better way forward for people. And that means being an international, publicly traded company with thousands of people and hundreds of millions of dollars of revenue and hiring plans all over the world. We’ve always wanted to be a public company and we have taken a lot of steps to get the company ready.

Q: In terms of expanding your technology footprint, which direction will you go?

A: Tableau 8 comes out in Q1. We’ve made some advancements in data integration because Tableau does an impeccable job being the Swiss Army knife that reaches out to every file format and gets it hooked up really easily. We’re taking more and more steps to make sure that you can blend all those results together in a way that’s crisp and clean and easy.

In terms of a road map, that road map will never end. There is always going to be something more to do on the data integration element.

We’re big believers in that, and we’ve made big investments in mobile. Most mobile analytics products, they’re just a joke. It’s like, look at my chart on my iPad. On stage in San Diego I demonstrated drag-and-drop query of a big data store right in the palm of my hand, on stage. And I was zooming into it and diving and pivoting and drilling, right there with my finger, against a massive data collection.

Q: Do you see any difference in user behavior when people are literally getting their hands on the data?

A: That’s a good question. I don’t think anyone knows how tablets are going to evolve exactly, but right now we tend to see the common-sense situations. You know how often people are in a meeting, a bunch of numbers are being shown, and someone’s got an idea, but the prepared numbers aren’t quite right. The VP of sales says — wait a second, don’t decide to put our conference in Barcelona too fast because I’m pretty sure that we don’t have any customers in Spain. It suddenly becomes this ridiculous situation. Just being able to say — oh, hold on a second, drag, drag, drag — actually Spain is bigger than you thought.

What’s so important is that it’s active. It’s not “read the report on the iPad.” It’s literally get in there and interrogate a little bit and pivot and save and send and make it actionable. Because I think a lot of what’s been lost in analytics is that it’s been so development intensive — everything’s a project. They’re used to not being able to use data to inform a moment when the moment is happening. Now they can do it with a product like Tableau, and the tablet is a delivery vehicle.

Q: With tablets, people pick technology they like. Why should it be any different with business intelligence?

A: That’s exactly right. That’s what the consumerization trend is all about.