The Center for Data Innovation spoke with Frank Bien, president and chief executive officer at Looker, an analytics and business intelligence (BI) software company headquartered in Santa Cruz, California. Bien discussed the importance of making it easier for business users to ask questions using data, as well as how to foster a data-driven culture in a business.
This interview has been edited.
Joshua New: At the heart of Looker is LookML, the modeling language that powers the Looker platform, which one of your customers described as “revolutionary.” What about LookML makes it stand out from other languages?
Frank Bien: The proliferation of self-service visualization and dashboarding tools has been great for getting data into the hands of business users, but rather than simply build another dashboarding tool, we wanted to tackle the bigger problem of making giant amounts of data available in a very reliable and usable manner. LookML is essentially a modern and reusable language that builds on SQL, which is the de-facto business language but it is also difficult to understand and componentize. LookML is a higher-level language that sits on top of SQL and makes it much more extensible and reusable. By doing so, it can expose really interesting elements of data from really large data stores and put it into the hands of business users
Right now in most businesses, the data people have lines out their doors of people wanting data extracts to put into self-service tools, which causes a lot of problems. Looker built a single platform that could handle this whole process, and that’s all thanks to the data modeling made possible by LookML at its core. The end result is non-technical users can explore and make use of data in entirely new ways.
New: Looker has partnerships with a handful of cloud computing and data management companies, including Amazon Web Services, Teradata, and Microsoft Azure, which your customers use to host their data. Could you explain how Looker fits into this infrastructure?
Bien: That’s the world I came out of initially. I came from a database company called Greenplum that EMC acquired, which went on to become a company called Pivotal. We saw this giant amount of hype around big data, but there weren’t a lot of really successful stories of people using giant data sets. However, that era did give us big, fast, inexpensive analytic databases on an entirely new scale. It was a revolution in data infrastructure but I would argue there was only a slow evolution of the tools that went on top. Looker is all about unlocking the value of these vast data stores.
If you look at tools like Google BigQuery or Amazon Redshift that let you deploy petabyte-scale data repositories in minutes, the whole world of how BI used to work gets flipped upside-down. You used to have to extract data from a database, put it in a warehouse, extract it again, heavily transform it, put it in a data engine, and then finally do analysis on very small amounts of data. These new databases allow us to operate on data where it sits, but there weren’t tools to do this very well. Looker leaves the data in these fast warehouses so it can leverage their power to transform, aggregate, enrich, and analyze the data while it stays in these systems. Looker sits on top of the database, between the users and their data, and transforms their business questions into queries that run directly on these immensely powerful databases. We greatly simplify the analytics stack because there’s no longer a need to remove and shrink the data to conform to the needs of the smaller databases. The result is that you can answer much bigger questions and get a lot more value out of your data because the “plumbing” required is much faster and more flexible.
New: In March 2016, Looker partnered with messaging and collaboration tool Slack to launch something called Lookerbot. What does Lookerbot do?
Bien: This is part of a much larger trend that I think people will get a lot of value out of. Today, Looker lets you describe all of this data in your databases and provide it to people in your business so they can use it. Normally, this would go into dashboards and visualizations or other things that most people think of as BI tools, but that doesn’t make sense anymore because it treats BI like a separate work process, kind of like driving while looking in the rearview mirror.
With Lookerbot, we wanted to bring data out of these traditional BI tools and into people’s daily routines. Collaboration tools like Slack are one area where that’s happening. As an example, assume some team members are talking about a customer issue in Slack. Rather than stopping the conversation going on in Slack and doing a separate unrelated work process in a BI tool, Lookerbot brings data right into the conversation in Slack. This enables conversations like, “Hey, something’s going on at customer X, do you know what it is?” and Lookerbot can pull data about that customer into a chart right in Slack. Then the conversation keeps going and the whole context and history is captured. The Slack Lookerbot brings data right into the conversation so that users can use it to make better decisions.
New: A lot of Looker’s customers are well known for their nontraditional business models. For example, Warby Parker is known for it’s at-home, “try before you buy” approach, Dollar Shave Club sells shaving products through a subscription-based service, and thredUP acts as an online marketplace for secondhand clothes. Is there a particular reason Looker and these companies have paired so well?
Bien: As I mentioned, Looker goes beyond dashboards and charts that most of industry uses because we wanted to be differentiated by doing something more. We wanted to start working with people changing their industry with data, but a lot of these companies are up against giant, entrenched competitors. So if companies like Warby Parker and Dollar Shave Club wanted to win, they had to understand what was happening with buying behavior, or with their supply chain, or what their customers wanted, better than the competition. We offer a business-centric data analysis and were able to help these companies develop their data culture.
Tom Tunguz of Redpoint and I actually wrote a book about this called Winning with Data. If you can put reliable data into the hands of core business users, you make it easier for them to ask questions, which means they can start asking more questions. This is the central part of that data culture, where you have people making decisions based on data rather than just guessing. Once you demonstrate the value of data culture in e-commerce, other companies take notice. Now, more traditional brands like Kohler and The Economist are trying to do the same thing. The tech companies have been leading for sure, but the need for real access to data is spreading rapidly.
New: You’ve written about the need for the role of “chief data officer” to evolve into the role of “chief analytics officer.” Could you describe the difference? Why is the latter more beneficial?
Bien: Whether people call this role the chief data officer or chief analytics officer in the future doesn’t really matter, but traditionally, the role of the chief data officer has been creating the infrastructure and repositories of data assets. This really took off in the advent of big data hype. In contrast, the role of the chief analytics officer is helping people answer questions and empowering the business with data. Generally, a data officer is information technology-centric and an analytics officer is business-centric.
It’s important to note that these aren’t mutually exclusive roles. In fact, I think most chief data officers today really rely on their analytics capabilities. But going forward, companies should want to move on from just caring about infrastructure and start focusing more on how to build a data culture.