Fast-growing analytics startup Birst, Inc. is announcing a new way for organizations to democratize business intelligence, enabling individual users to customize their views and applications without compromising the central data store.

A concept the company calls “Networked BI” is designed to “virtualize the entire BI ecosystem, transforming every aspect of an organization’s approach to analytics,” Birst said in a press release. The features are likely to have the greatest immediate value in geographically dispersed organizations, where local and corporate data need to be combined, and in vertical markets.

In a conventional analytics scenario, customization would require the corporate IT group to make copies of data, which would then need to be kept in local silos versioning and synchronization problems. Birst says its approach eliminates replication and inconsistency problems. Users can also share customizations with each other.

“There’s not only one copy of the data but also only one copy of the metadata,” said Pedro Arellano, senior director of product strategy at Birst. “You can build the fabric on your own without relying on the centralized team.”

In making the announcement, Birst highlighted research by Gartner, Inc., which said in January that democratized BI will tax the ability to IT organizations to handle demand. It said new capabilities were needed to enable business users to “access, profile, prepare, integrate, curate, model and enrich data for analysis and consumption by BI and analytics platforms.”

That’s a need Birst says it’s addressing with Networked BI. “We can support an end-user coming into Birst with no business model defined,” Arellano said. “That can be the most time-consuming part of the process.”

Users can combine data from different sources in a protected sandbox and share the results. For example, a user can mix data from an Excel spreadsheet with records downloaded from Salesforce.com and share the model with the rest of the company. Others can then embellish upon it.

Changes are logged in the manner of a wiki or Google Document, Arellano said. “Each user can see other people’s changes,” he said. “Every construct is contained in a semantic layer and these definitions are shared.”

Birst says the model achieves an ideal balance of central control and individual experimentation. “It works from the top down – with central management delivered to decentralized groups – and from the bottom up – with decentralized groups working on their own,” Arellano said.

Birst will make the virtualization features available as an optional upgrade to its basic subscription plan. The company doesn’t publish price information.

Creative Commons image by Creative Tools via Flickr