It’s Time for Behavioral Analytics to Be Simpler, Interana Says
There are all sorts of patterns of human behavior hidden in your Web logs, call records, and sundry IoT data. The trick, of course, is identifying the patterns without expending enormous resources. For the folks at analytics software vendor Interana, the secret to success lies in enabling non-technical workers to explore and query time-series data in a fast and user-friendly fashion.
With another $18 million round of venture funding announced this week, Interana is set to further build around its core proprietary technology, which includes a highly compressible time-series database and Web-based UI that lets users explore time-series data sets without the need for a data science expertise or a SQL programmer.
“A lot of the value we’re adding is the way our unique data store works,” says Interana CEO and co-founder Ann Johnson. “The magic is in between the data store and the UI. Because we don’t use any of these SQL-types intermediaries, we’re able to do it much more efficiently and much more intuitively.”
Interana has been slowly growing a reputation in the field of behavioral analytics since Johnson co-founded the company with two Facebook engineers—including her husband, CTO Bobby Johnson, and lead developer Lior Abraham—in late 2012. At Facebook, Bobby Johnson was in charge of executing massive SQL queries that spanned multiple pages in order to answer relatively simple questions, such as “Which Facebook users have been active for five out of the last seven days?”
“Time-series questions aren’t a good fit for SQL,” Ann Johnson says. “SQL has no way of looking at how things flow through time and really interacting with individual things through time.”
The Interana database makes time a first-order principle, while storing other items, like the names of users or the IP address of a computer, as secondary attribute. Along with some super-fast compression algorithms, these qualities help to enable the Interana database (which runs on distributed Linux clusters) to deliver interactive responses to queries that need to look at millions or billions of rows of data.
“We have a really deep, cool technology that enables us to do these really fast queries,” Ann Johnson says. “The reason we can do it so quickly is because we realize the nature of these queries is very different….It differentiate us and frees us from a lot of constraints of the data warehouse and the entire index, the SQL overhead or anything with Hadoop, which isn’t really made for fast data retrieval.”
In Sik Rhee, general partner at Interana investor Vertex Ventures, says the company is poised to disrupt a congested analytics market and give companies a better way to find insights in data.
“Figuring out better ways to ask the right questions of their data is an iterative process, and unlike other offerings that tout themselves as self-service but have many hidden intermediaries, Interana provides true self-service analytics, giving users that critical ability to ask questions without a gatekeeper,” Rhee says in a press release.
Interana’s software is used for all sorts of behavioral analyses that have an element of time to them. Many customers use the software to build interactive dashboards that make it easy for them to explore how certain metrics are changing over time, such as the number of calls made by particular groups of people, or the popularity of certain parts of a website or mobile app.
“A lot of our customers are interested in things like daily active users or monthly active users to a website,” Ann Johnson says. “You can also look at funnels. Let’s say you want to understand how many clicks it took to get to check out. You can monitor how many steps are in that funnel.”
There really aren’t many shrink-wrapped solutions for answering these sorts of time-related questions at big-data scale, Johnson says. Many will struggle with complex and expensive SQL queries running against a relational data warehouse, while others will roll their own queries in Hive or MapReduce in a Hadoop environment.
“As soon as you’re looking at loosely 20 billion records or more, your only options are Hadoop or Interana,” Johnson says. “I would say anybody with a large IoT data set is writing custom programs to look to answer their questions.”
Many of Interna’s early customers have been ecommerce companies, which is not surprising, considering its aptitude with clickstream analysis. But as the company matures, it’s increasingly being brought into more enterprise accounts, including one major tire manufacturer.
Johnson expects that trend to continue. “When we go into a lot of these customers, they’ll have three to 10 people per department or company, who are able to use their data tools,” she says. “When they install Interana, they now have 100 to 300 people who are able to access it. So it’s really taking the bottleneck of the SQL programming out of the loop and allow people to really just understand, deeply, their data.”
The $18 million in funding comes from existing investors, including Battery Ventures, Data Collective, Allen Company, Fuel Capital and Index Ventures. The new investment brings Interana’s total funds raised to $46.2 million.