Interana Tapped by Microsoft to Analyze Bing Searches
Interana, a big data startup that develops a database and related tools for doing analytics on time- and event-based data, today announced that Microsoft selected its software to analyze queries made with the Bing search engine, with the goal of better understanding user behavior.
This is a major win for Interana, which was co-founded by former Facebook engineers and came out of stealth mode only nine months ago. Microsoft has a long history of developing its own products, and does not often license third-party products. Microsoft had been using an internally developed system for analyzing how the behavior of Bing users changed over time, but evidently liked Interana’s solution better.
“They tried Interana and it just did a lot of exactly what they needed,” Interana co-founder and lead developer Lior Abraham tells Datanami. “It’s a pretty big deployment. Definitely a win and a validation that building a product that’s geared to solving the problem of getting answers out of data at scale.”
Interana’s solution is based on a distributed column-based database designed to process vast amounts (i.e. trillions of rows) of time-series data, such as clickstreams, call detail records, change logs, or sensor readings. Instead of storing time-related information secondarily, as is the case in many databases, Interna treats time as a first-order principal, which enables time-related queries to run very quickly. On top of this core, Interna built a graphical interface that allows non-technical people to query and explore the data, without special training or SQL knowledge.
Abraham and fellow co-founder and CTO Bobby Johnson worked on time-series related problems while at Facebook. In a previous interview, Johnson recalled how Facebook would run a monster SQL query every night to figure out which users used the social media platform over the past week. As the log and event data kept piling up, the limits of relational technology became readily apparent. The two joined with Johnson’s wife, Ann, and together co-founded Interana to build and market a time- and event-based database designed for massive scale.
Interana’s software can query trillions of rows of data (more than a petabyte) and return an answer fairly quickly, thereby enabling users to explore their time- and event-based data and iterate upon it quickly. When you get to a certain scale, other types analytic databases just can’t keep up, says Interana director of marketing Mark Horton. “That was a huge consideration in looking at our platform,” Horton says. “Being able to do what we do at the scale we do, we’re one of nobody I know of. That’s what caught Microsoft’s attention.”
Microsoft engaged Interana shortly after the company emerged from stealth mode last year and began a proof of concept. “They obviously generate massive volumes of event data through the Bing search engine,” Horton says. “They really needed something that would enable them to get to the analytics much faster at scale than what they previously had before.”
While details of the implementation were not available (Microsoft did not give press interviews), it’s clear from speaking with Interana that Microsoft has big plans for Interana’s technology. Microsoft will use it to generate better insight for conversion analysis, engagement analysis, retention analysis, and enabling root-cause analysis, in addition to conducting A/B tests.
Craig Miller, Microsoft’s group engineering manager for Bing Experiences, says Interana “met the challenge of delivering the insights we needed from our massive volumes of event data” during the POC. “Interana’s scale is impressive. It took only minutes to get answers to questions, opening new possibilities as to what we can do with analytics at interactive speeds. We look forward to continuing to work with the team.”
Getting the second largest search engine in the world as a customer is a huge win for Interana, which is based in Menlo Park, California, and has raised $28 million over two rounds of funding. Other previously announced customers include Sony, Asana, Orange Silicon Valley, Tinder, Sonos, and BloomBoard.
“They’re analyzing their actual production data, the actual Bing searches–everything the user does through Bing on every Microsoft property,” Abraham says. “They’re using it for both exploration and transparency into how people are using the product, and then running experiments to make decisions at massive Bing scale.”