'Birst'-ing Onto the Cloud-BI Scene
Published: July 20, 2010
by Alex Woodie
Companies in the market for a full function business intelligence product they can access from the cloud may want to check out Birst. The San Francisco company, which launched a new BI interface for smartphones last week, says its customers get the equivalent of a fully integrated BI suite like Cognos, but at just 20 percent of the total cost, and a fraction of the complexity, thanks to automated tools and the software as a service (SaaS) and cloud delivery method.
Birst positions its software as an end-to-end suite of BI applications designed to help people analyze their data and make business decisions. Customers can view and manipulate their data through pre-defined reports or dashboards served through a Web 2.0 interface, or even tap into an existing OLAP cube, while an array of administrator tools (such as report designers, database connectors, and ETL engines) keep the routine work to a minimum.
Birst was launched in 2004 by the principals behind Success Metrics, which developed fully hosted BI solutions aimed at companies in the pharmaceutical, insurance, and retail brokerage industries. As Success Metrics' BI technology advanced and open source software became more powerful, the company decided it was time to launch Birst, and take a gamble on self-service, cloud-based BI.
According to Birst CEO Brad Peters, the market is clearly shifting to cloud-based BI. Peters, who worked in the analytical side of the house at Siebel Systems and Oracle, witnessed how Salesforce.com utterly decimated the market for on-premise CRM software with its brand of SaaS-based CRM. Now we're at the same point on the road when it comes to BI.
"The on-premise CRM solutions that were designed as point solutions and then given to IT to cobble together were incredibly expensive to install and to maintain, and had ridiculously long deployment cycles," he says. "The same types of problems exist in business intelligence. Why would I want to assemble my own car, if I could just drive one off the lot? If you have your own mechanic, and he likes to do sheet metal welding, that's what you're going to do. But an increasingly number of people want to buy the same pre-built product off the lot. I do think SaaS is the next great architectural shift in software."
The Birst approach cuts down on the extraneous work in several ways. The first and most obvious has to do with hardware and administration. Servers, operating systems, database, and all the ancillary necessities--management, backup, recovery, etc--is handled by Birst. The company contracts with the managed services provider Rackspace for space in a data center, and manages its application for customers. Reliance on open source technologies, including the high-speed analytical InfoBright database, allows Birst to leverage the advances of others.
The other major way that Birst cuts down on the obscene price tags and deployment timelines that typically accompany large BI roll-outs is by streamlining the development and deployment of the data schema used in the data mart or data warehouse. In effect, Birst will prepare the data for BI for you.
Peters refers to this key piece of Birst technology as the unified metadata layer. "We can automate the construction of a full star schema in the cloud," he says. "This allows you to map metadata onto either your local data or data you load in Birst, and then quickly turn around a business or semantic view of that data to your end users, who can build reports and dashboards."
Considering that building a single star schema to answer a single business question costs the average company $3 million to build, $500,000 per year to maintain, and takes months (according to one Gartner study referenced by Peters), the fact that Birst can automate this aspect of data warehouse construction can mean a lot, especially to small and mid-sized companies that can't justify this kind of expense and have no experience with data warehousing.
Customers can host their own data--or use a mix of on-premise and Birst-hosted data--while still using Birst's tools to create BI views that transcend the physical location of the data. "Generally what we see is a bifurcation," Peters says. "Customers that are more sophisticated keep it local. They may have some data in a secondary system that isn't structured, maybe in Excel or mainframe flat files. But it's not loaded in a data mart yet. They can keep it local, while using Birst to federate it and allow you to create a single view of all that."
Birst has 25,000 active users across 100 customer organizations, ranging from small companies with just a few tables to companies with billions of rows of data, Peters says. A fraction of these customers are using Birst to analyze data housed in IBM System i (AS/400) servers. While the company doesn't offer a pre-built connector yet for DB2/400, it can use standard ODBC mechanisms to get at the data, he says. Standard connectors are offered for DB2, Oracle, SQL Server, and MySQL.
The other advantage that Birst has is price. The average cost per user is less than $1,000 per year, considerably less than the Cadillac-level BI products that have dominated the market for so long. And because Birst operates in a cloud environment, the company can offer some pay-for-use pricing that scales up and down according to usage.
Just as Salesforce.com ushered in a new era of SaaS-based CRM offerings, Peters predicts that Birst will break down the barriers to BI. "Once customers realize that on-premise BI has a bunch of headaches, they'll realize that Birst is a compelling alternative," he says.
Post this story to del.icio.us
Post this story to Digg
Post this story to Slashdot