Social and mobile analytics startup Kontagent has expanded its business to include a data-mining service powered by Hive, the SQL-like interface for querying data stored within Hadoop. It’s a smart move by the company, and one that other cloud-based analytics providers would be wise to replicate.
What makes it so smart is Kontagent’s realization that out-of-the-box analytics dashboards can only do so much. Even though most startups working in the space have some skilled data scientists on board building the backends that power the pretty metrics and dashboards users see, the applications will always be limited to a select number of broadly applicable and pre-determined features. But that means a lot of wasted data, and a wasted opportunity to facilitate users’ curiosity.
Kontagent CEO Jeff Tseng explains the rationale nicely in a press release announcing the new service, called kSuite DataMine:
Our core product is an online dashboard that answers hundreds of common questions about user behavior in social games and mobile apps, but we wanted to empower data analysts to dive deeper by creating ad hoc queries. The challenge is that the kSuite platform captures as many as 50 billion events for our customers each month. We needed to provide a more flexible big data exploration solution for analysts and business intelligence professionals to dive into the raw data.
Hive is the natural choice to power the offering. Assuming Kontagent stores data in Hadoop, its backend is already setup to handle the new service. Additionally, Hive’s SQL-like nature means a minimal learning curve for customers once it has been abstracted underneath the Kontagent UI.
This is the kind of thing we might expect to see down the road from companies like Parse.ly — which has suggested it might open up customer data to custom analytics — or even more-established firms such as Clickable. When users get an itch to go exploring their data, why not be the one to let them do it?