CoolaData, a Tel Aviv, Israel-based startup, has raised a $7.4 million series A round for its behavioral analytics service that’s built entirely atop Google’s cloud computing platform. Greylock IL and Carmel Ventures led the round.
The company came about a year ago after co-founders Tomer Ben Moshe and Guy Greenberg were running a consulting business and realized a “tremendous hunger for what [they] call unified behavioral analytics,” Ben Moshe said during a recent interview. What that means is companies want to take all the data they’re collecting via web applications, combine it with internal data and then do things like cohort analysis, clustering and other statistical-analysis functions on top of it.
CoolaData’s platform enables this by ingesting the data, processing it all using a custom semantic engine that understands the connections between various pieces (e.g., that two things are talking about the same customer) and then exposes it via a SQL interface or what CoolaData calls an “analytic document.” The latter, Ben Moshe explained, is like a Word document for analytics that includes a variety of visualizations as well as those statistical analyses. The documents can embed in other applications or platforms and receive third-party data directly via API or internal data from the CoolaData system.
CoolaData’s take on SQL is customized to include “behavioral extensions,” Ben Moshe said. That means they can sort the data not just by value, but also by sets of similar users or patterns of behavior over time, for example.
As great as all the analytic capabilities and data-ingestion and processing models might be — and they’re definitely CoolaData’s secret sauce — I think the decision to use Google’s suite of cloud services under the covers is just as important from a business-model perspective. Just like startups in years past would take advantage of cloud servers from providers like Amazon Web Services, they’re now beginning to take advantage of those providers’ various data-analysis services to provide new types of data-based applications without reinventing the wheel.
“While it is young,” Ben Moshe said of Google’s cloud, “it is very robust.” In the case of CoolaData, Hadoop and HBase are running on Compute Engine, App Engine serves web applications, BigQuery acts as the columnar database and data is stored in Google Storage. We recently covered another startup, Pondera Solutions, that uses the BigQuery, the Google Prediction API among other Google tools to power a fraud-detection service.
In fact, although CoolaData is technically a Google customer, it’s also a technology partner, he added. It’s one of the user companies Google enlisted to help make its cloud platform more appealing to business users rather than just developers.
Ben Moshe said the company hopes to make its platform generally available in the next few months (it’s currently used by a handful of beta customers in various online industries) and to add connections to additional third-party data sources such as Twitter and advertising platforms. The ultimate goal, he said, is to move more toward predictive analytics and a broader range of capabilities (e.g., sentiment analysis) as more data sources and developers come on board.
And, technically, as long as Google keeps adding more services and tools to its cloud platform, CoolaData can keep evolving its backend system to handle new features.