Today’s guest blogger is Nick Terner, backend team leader at CoolaData, based in Tel Aviv, Israel. CoolaData provides fully actionable, unified behavioral analytics solutions for online companies. The company recently raised $7.5 million in Series A funding.

CoolaData delivers something different to the world of big data analytics. Many companies in the space offer solutions designed to answer the “what” and the “when” of consumer online behavior. We uncover the “why.” Why did a customer cancel an account? Who is likely to uninstall my app? Why did a customer abandon an online shopping cart?

To uncover the why, we need to examine the sequence of events that occurred before a customer made a certain decision. For example, did they go to an online retailer’s customer service site and not receive the help they need? And to better understand the why, we need to look for patterns in online behavior, often times using historical data.

All of the above requires that we analyze a great deal of data – current and historical – using innovative methods like cohorts, paths and funnels. And to produce actionable big data analysis in real-time, we use Google Compute Engine for the data enrichment process. Compute Engine allows us to take the information, analyze and learn from it in order to determine actionable insights for our customers quickly.
CoolaData is building a unified behavioral analytics solution that is infinitely scalable, powerful and dynamic, with Google Cloud Platform as the foundation. Hadoop and Storm run on Compute Engine, and Google App Engine serves our Semantic Layer web API. Google BigQuery enables analysis of the huge datasets we collect from a variety of sources via Google Cloud Storage. So far, gaming, e-commerce and FinTech companies are the initial adopters. We also participate in the trusted tester program for Google Compute Engine, which has allowed us to use new features, while at the same time providing recommendations for others.

Our development team decided to use Compute Engine when we built our platform over a year ago. Google’s predictable workload performance was key, but unified authentication across the cloud platform and encryption were also major factors. With Compute Engine, we know exactly what we’ll get when we provision an instance. The Compute Engine team offered great support and allowed us to plan out our future capabilities. In addition, Compute Engine now has transparent maintenance, which allows us to perform regular proactive maintenance without impacting running VMs. We expect this will make Google a reliable, high-performance environment, and regular host patches will also ensure better security than other providers.

Compute Engine’s predictable performance is a huge benefit because it allows us to plan for growth. As we move out of beta testing and prepare for the general availability of our solution in early 2014, we know we can onboard new customers smoothly – even if they bring us huge amounts of historical data. We also know that we can easily add connections to more third-party data sources. The more information we can analyze, the better we’ll be at helping them not only understand the reasons behind a customer’s decision, but also predict and influence the choices that customer and other similar customers will make in the future.

-Contributed by Nick Terner, Backend Team Lead, Cooladata