Looker’s Audacious Moonshot to Outgrow Oracle
You may have considered Looker to be just another business intelligence tool. After all, that’s the bucket analysts at Forrester and Gartner put the company in. But talk a bit with its founder, Lloyd Tabb, and you quickly realize that BI is just a part of the story. In fact, the company has much bigger goals in mind.
Tabb is a master technologist who translated a love of creating programming languages early in his career into a string of successful startups that left their imprints on the Internet age. After leading development of dBase at Borland, Tabb went on to write the first Web application server at a company that would be acquired by Netscape. After leading development of Navigator and helping to found Mozilla, Tabb went on to found LiveOps, one of the first commercial crowdsourced ventures.
Throughout his career, Tabb built an untold number of tools that let users access data. But all too often, those tools were bespoke, one-off creations that lost their value outside of a narrow set of use cases. What the world really needed, Tabb figured, was a universal tool that could relate to any data.
The idea wasn’t well-received. “I wanted to build a new programming language to describe data, and everybody thought I was crazy,” Tabb told Datanami at Looker’s recent JOIN conference in San Francisco. “They’re like, ‘You’re going to make it visual, right?’ No, I want to write a language. ‘A language? Nobody wants to learn a new language!'”
Around 2011, Tabb sat down and started writing the new language. The main goal was to provide an abstraction layer above SQL that could be used for describing how different pieces of data relate to each other, along with functionality for creating dimensions, aggregates, and calculations.
That language would later be called LookML, and it would form the basis for the Looker data platform. LookML works with and generates SQL, but it also addresses a key weakness in the workhorse language that has led to numerous problems with how people manipulate and consume their own data.
“SQL is Assembly language for data. It’s Turing complete. You can actually do any relational data transformation in SQL because SQL is complete,” Tabb said. “But it was architected in a way that has no reusability and no encapsulation. You’re always re-writing the query from scratch every time. Programming is about reusability, and so Looker takes SQL and makes it reusable.”
Instead of writing SQL queries by hand, Looker customers describe their data relationships using LookML. When a customer wants to query the data, they rely on LookML and the Looker platform to automatically generates the SQL syntax, which is then pushed down to the databases for processing on an as-needed basis (it currently supports more than 35 relational databases).
This approach provides several advantages beyond the reusability factor, including correctness. “The fact that we described it up front lets us do all kinds of crazy stuff,” Tabb says. “If I have a revenue declaration in a Looker model, it’s declared in exactly one place. So if revenue is net of tax, and I change it to be net of tax except in California, that calculation will ripple through everything, and every single report will update because it’s declared in exactly one place.”
By contrast, if that revenue calculation was defined using SQL, there’s the potential for the data to go haywire. “If you save that query, and then somebody changes it later, you have no way of going back to change it there,” Tabb says. “So if you run that query again, it’s going to be inaccurate. That’s the standard. That’s what people do. That’s the thing that leads to data chaos.”
Building a Platform
LookML provides the technological foundation for the company’s data platform. The most visible applications currently sitting on that platform are the BI and visualization tools that Looker sells. They’re compelling in their own right, as they allow non-technical users to select data sources, apply filters, and view results, all in an intuitive drag-and-drop manner.
But BI isn’t the only goal that Looker has in mind.
“I think of the platform as the bedrock, and we’re building buildings on top of the bedrock,” Looker’s Chief Data Evangelist Daniel Mintz said at the user conference. “To date we’ve had the BI tool, which is by far the tallest and most full-featured building…Now we’re building more buildings. It’s a horizontal platform, but it’s a very wide platform on which we can build a whole lot with.”
Other buildings going up on that bedrock include things like Data Blocks, Viz Blocks, and the ActionHub. The Looker Block concept originated with version 4, which debuted in 2016 and gave customers the opportunity to use pre-built data sets and visualizations. Looker provides a handful of blocks, such as weather and demographic data, while some are built by Looker’s 160-strong partner network.
The ActionHub concept debuted with last month’s launch of Looker 5, and provides users with the capability to work with data originating in outside systems, such as Salesforce, Zendesk, and Google AdWords. Instead of tabbing out of the Looker’s browser-based user interface, users can make changes to connected applications directly within Looker.
While BI may get the headlines, the driving force behind Looker is the capability to build a variety of data-driven applications atop the company’s platform. Next year, the company plans to add more pre-built applications to its platform, including marketing analytics, IT operations analytics, and Web and mobile event analytics. A virtual reality (VR) app that started as a hacking lark has also garnered traction among Looker’s customers, who pay $3,000 to $5,000 per month for 10-user subscriptions to the platform.
“We know we have a platform on which you can very easily build these kinds of applications,” Mintz said. “We’re following their lead and saying, huh that’s fascinating. We don’t expect most people to do all that work. We can build that once and give it to lots of people.”
‘The Big Thing’
This concept of the data platform is a fairly new one, and Looker doesn’t think it has much competition at the moment.
“We do things differently,” said Tabb, who is also Looker’s CTO and chairman. “We’re standing alone in a room. We’ve taken such a different approach to it. We’re not following anybody. We’re building what we think is right.”
Tabb doesn’t take offense to the fact that most people, including industry analysts, have wrongly classified Looker as a BI and analytics tool, when in fact it’s building something that’s arguably much bigger.
“It’s the only place for them to put us because there’s no data platform [category],” he said. “We’re a way of dynamically asking just about any question out of data, and I don’t think anybody else really does that.”
For the time being, Looker will own this category for themselves, and by all accounts, the payoff has been handsome. While the privately held Santa Cruz, California company doesn’t disclose its financials, a company representative publicly said it had revenues in 2016 in the “low to mid tens of millions of dollars,” according to the “One Million by One Million” blog. If the 400% growth rate cited by the company holds up very long, it won’t take long to earn back all of the $177.5 million it has received in venture capital funding. It was reportedly valued at $850 million during the last valuation.
Tabb compared Looker’s growth rate to a “moonshot” and says the company is actually growing faster than planned. “We haven’t had any Apollo 13 moments,” he said. “We’re going to the moon. It’s awesome.”
Tabb’s confidence is based on the fact that every major company needs to have better access to their data to compete in our increasingly digital economy. As the sizzle around Hadoop turns to fizzle, attention seems to be turning to platforms like Looker’s, which can help users get value out of data sitting in multiple data repositories.
The momentum was also evident at the recent JOIN conference. The company’s second annual event was held at San Francisco’s Palace of Fine Arts Theatre (the old Exploratorium building), and was exponentially larger than the company’s first conference held in New York last year. Many representatives from Looker’s customer base, which includes names like Docker, eBay, Sony, Uber, and Warby Parker, were on hand.
“I just keep seeing the logos that we’re getting and the traction that we’re getting and the adoption and love that we’re getting – I don’t think we have any barriers to scale,” Tabb said. “We’re the big thing. We’re not the little thing. I don’t think there’s anything like us. We’re on a trajectory to be an Oracle-size thing.”