Cloud Computing: How Birst thinks it can make data warehouses useful again

Grazed from GigaOM. Author: Derrick Harris.

Analytics startup Birst wants to plant a cloud seed in your data warehouse to make it accessible to the business users who actually need it. No, it’s not trying to move any data to the cloud, but rather is implementing tried-and-true cloud product architecture to rethink how users interact with their companies’ centralized and valuable data.

The problem, according to Birst Founder and CEO Brad Peters, is that using data warehouses has traditionally been a cumbersome process. Only a limited amount of data — perceived to be the most valuable data — goes in, and doing any work on that data within the warehouse usually requires the assistance of the IT department. As a result, data warehouses have essentially become staging grounds from which data is shuttled via feeds to various departments that analyze it against data on their own systems…

There’s a strong argument to be made that all this distributed data is a bad thing because it means every department is working in isolation. There’s one data warehouse and dozens of departmental databases, and little connection between any of them. This concern is also driving EMC Greenplum’s vision around Chorus, its big data collaboration software that lets co-workers share and discuss their datasets, and other efforts (e.g., VirtualWorks) to provide at least a federated view of organizations vastly distributed data assets.

Birst thinks it can change companies’ data management practices with a new product called Distributed Business Analytics that lets them drop a multitenant instance of its analytics product right into the data warehouse. That way, Peters said, users can create their own sandbox environments with their departmental data and join it with warehouse data to run analytics centrally. They can also share data, create ad hoc teams and otherwise collaborate.

It’s like Birst’s original software-as-a-service offering, which has resided behind the corporate firewall for a while thanks to an appliance version of the product, but focused on (indeed, within) the data warehouse. SaaS products must be easy enough for business users to manage from start to finish. “If our end users couldn’t put data in and do something useful with it, we’d be out of business,” Peters said.

Although it’s targeting data warehouses because that’s an area where companies have made such heavy investment over time, Peters said the new product can work with any centralized data store, including Hadoop, that generally requires IT intervention to access and analyze.