Interview: Lloyd Tabb, Chairman & CTO, Looker on Front-line Analytics and Data Democratization

Lloyd Tabb is founder, Chairman, and CTO of Looker. For the past 25 years, he has been designing, building, and leading large-scale software projects. Early in his career, Lloyd designed database products for Borland. With the birth of Mosaic, the first web browser, he recognized that the browser would redefine the way databases are accessed and shared. As a founder of Commerce Tools, later acquired by Netscape, Lloyd built on that vision by designing one of the first web application servers. At Netscape, he was the principal architect on Netscape Navigator Gold (which later became Composer) and a lead on Netscape Communicator. Lloyd was also a founder of Mozilla.org, Netscape's open source organization, where he helped acquire and lead their Open Directory Project.

Here is my interview with him:

Anmol Rajpurohit: Q1. First of all, I would like to start with congratulating you for the success of your innovative product. Looker was mentioned in the list of 20 red-hot pre-IPO companies in 2014 B2B Tech by IDG Connect. According to you, what are the top three features of Looker? What other unique benefits of Looker are worth mentioning here?

Lloyd Tabb: Looker is focused on empowering data analysts to share their analytical capabilities across an entire organization. We’ve created a development environment where they can, at a base level, construct a reusable data model that forms the basis for high-level objects, such as explorable views, dashboards, and filtered dashboards. The model is something like a data mart, but coded in LookML, which is the modeling language built into our platform. LookML is powerful enough to do much of what is traditionally done as part of an ETL process.

AR: Q2. Currently, there seems to be a lot of talk but hardly any action around "decentralization and democratization of data across the organization". How do you expect Looker to change this trend?

LT: There’s a saying that insanity is doing the same thing over and over again and expecting different results. At Looker, we’re doing things differently, and we’re driving new and better results. By coupling a web-first interface with a modeling layer that enables a simplified—but complete—end-user view of the data, we’ve been able to effect real change in customer organizations.

AR: Q3. How does Looker help organizations tackle the talent crunch for data scientists? What sort of learning curve is involved here for Business Mangers to get started on using Looker?

LT:

The data teams in companies are typically made up of smart folks with brains that gravitate toward numbers and economics. They’re people with innate curiosity and some basic technical abilities. I don't see that changing much. What’s changing are the tools they use.

Looker is very easy to learn for a data analyst—they can learn it in 30 minutes. Business users can learn to query Looker in even less time. No matter what kind of talent you have in your organization, Looker makes your people better and more effective at their jobs.

AR: Q4. How do you define "Front-line Analytics"? Can you share a few use cases of how Looker empowers decision-makers?

LT: Looker can help anyone in the company. In a web world, customer acquisition is a cost and revenue center; tracking customers to their source and then following the lifetime value of those customers are tasks that any “front-line” decision-maker can tackle. The analytics team builds the model, but a regular finance person or marketing person can pull their own data, do the analysis, and watch it over time. They can also look for other trends: Do people who use coupons have a lower lifetime value? By how much?

Marketing managers can drive micro-focused campaigns that target very specific customers. For example, they can send an email only to people who have ordered more than 10 times but have not ordered within the last 60 days.

AR: Q5. How do you look at Looker's competitive landscape, particularly companies such as Tableau, QlikView and Birst? What do you consider as Looker's sustainable strategic differentiation?

LT: Looker takes a discovery-first, web-first, and model-first approach that operates 100% in database. Rather than creating pretty, pixel-perfect reports and dashboards delivered on a periodic basis, we’ve created an environment that fuels our users’ curiosity. With Looker, they can see all of their data. If they have a question, they can dig deep and get to an answer. They can pursue an interesting thread. What they learn makes them—and the business—smarter.

Looker also takes deep advantage of very powerful databases. We run on Amazon Redshift, HP Vertica, Pivotal Greenplum, and the like. These are huge computational clusters, and our LookML models can see everything in them. The legacy systems you mention have their own data engines that are only looking only at a subset of the data.

AR: Q6. Why do you think it is high time for firms to shift away from the cliche of "Daily Active Users (DAU)" (or similar metrics) and move towards more meaningful characterization of users? What approach would you recommend to analyze event data for business insights?

LT: The problem with Daily Active Users is that it doesn’t really tell you anything about who your customers are. Growth does matter, so you need to measure it. But you need to measure more. If you're collecting event data, you have the opportunity to convert that data into insights about your users. You can start by sessionizing your data, so you can calculate things like time-on-site and understand how to approach measuring engagement. With engagement metrics in hand, you can characterize users and do all sorts of cohort analysis. For example, you can show usage by month pivoted by user signup month. Bucketing users by lifetime usage, you can see if big lifetime spenders spend more or less per transaction over time.

When you can measure engagement and characterize your users, you have what you need to build an audience and build a business.

AR: Q7. Data Scientist has been termed as the sexiest job of 21st century. Do you agree? What advice would you give to people aspiring a long career in Data Science?

LT:

Don't stop being curious. The job is discovery. The tools will change, but the discovery-oriented way of thinking won't. If you’re the kind of person who gets causation-vs-correlation, then you’re the right candidate for a career in data science.

Helpful majors are econ and computer science (and maybe math or even accounting). You need to be competent at computer science and math, but more crucial is economics—the study of cause and effect.

AR: Q8. What was the last book that you read and liked? What do you like to do when you are not working?

LT: I like Ben Horowitz's new book The Hard Thing About Hard Things. I love Seth Godin and most of the things he's written. Behavioral economics is an area of intense curiosity for me. When I'm not working, I'm usually thinking about working. I love my work.

Related:

• Big Data Landscape, v 3.0, analyzed
• Interview: Paco Nathan, Chief Scientist at Mesosphere
• Don Zereski, VP, Local Search & Discovery, HERE (Nokia) on Location Analytics and Architecture Evolution