In-Depth

With Birst, Former Siebel Veterans Tout SaaS-y BI

Birst targets the important (but small) projects that are ill-served by existing BI tools, its developers argue

If, as some folks maintain, there's a silent plurality of business intelligence (BI) users whose needs aren't being addressed by conventional BI or spreadsheet products, they're suddenly getting a lot of attention.

Last week, for example, start-up Lyzasoft launched, touting a completely client-based BI offering that, officials claim, targets the unique -- and heretofore ill-served -- needs of analysts and power users (see http://www.tdwi.org/News/display.aspx?ID=9129).

This week, Birst, a software-as-a-service (SaaS) BI deliverable created by a pair of former Siebel Systems Inc. veterans, debuted.

In contrast to Lyzasoft, Birst parent company Success Metrics touts a SaaS-ified -- or completely off-the-client -- approach to help BI users. Like Lyzasoft, however, Birst targets the not-insignificant group of BI users typically involved with seasonal, one-off, or idiosyncratic analytics projects that aren't "important enough" for their internal IT departments.

Individually, these projects might not amount to much; incrementally, proponents argue, they can have an impact on the corporate bottom line.

Birst uses SaaS to empower user self-service. Only self-service can free frustrated users to pursue otherwise neglected or low-priority projects on their own, argues Brad Peters, CEO of Success Metrics. In a more general sense, Peters maintains, SaaS itself is a good solution for the twin bugaboos of enterprise BI: deployment complexity and lagging user adoption. He claims that Birst -- which Success Metrics has quietly developed (and, to a limited degree, marketed) over the last several years -- grew out of his and co-founder Paul Staelin's experiences at Siebel, where they led that company's formidable Siebel Analytics group.

"We'd sell them the [Siebel Analytic] software and we'd come back and see them at [a Siebel conference] and say, 'How have you guys done with it?' These are cases where they were talking about transforming their business processes, and all of this cool stuff. They'd tell us that they had four reports functioning at that point," he comments.

"There was a big gap between vision and reality. It basically came down to the complexity [of the pre-packaged] applications," Peters continues. "We had thought that building pre-packaged applications would solve that [adoption and deployment] problem. We'd say, 'Let's build a sales analytics application or a services analytics application. That'll make it easier for them to deploy.' In reality, that [approach] did nothing to address the complexity."

Birst aims to be all things to almost all users. It encompasses reporting, ad hoc query and analysis, and dashboarding/presentation components (via Flash/Flex technology from Adobe Software). Birst is self-service SaaS, according to Peters: it's designed for individual business units to get quickly up and running and analyzing, he maintains. To that end, it touts an integration wizard -- which supports Microsoft SQL Server, Microsoft Access, Microsoft Excel, or any kind of delimited format -- that automatically populates the remote Birst repository.

Birst itself creates (and optimizes) a data model and generates metadata (including dimensions with attributes, as well as time-based and comparison metrics), according to Peters. "We felt we would deal with that issue [i.e., integration] if we could automate a good chunk of that. If we could have something create data models for you automatically in the cloud. Create all of the metadata that you otherwise would have to create for yourself, and make it very easy for people to take raw data and get value out of that without having to bust open the raw plumbing and all of that stuff that goes on," he indicates.

One of the biggest selling points of SaaS is its relatively painless deployment model. This lowers the bar for entry, and gives customers an inexpensive way to test drive a SaaS offering without being locked into it: if it doesn't work as advertised, or if it doesn't address their needs, they can simply unsubscribe.

Success Metrics goes this proposition one better, offering a limited-use version of Birst free of charge. It includes 10 MB of free storage and support for up to five users. Capability-wise, "Free" Birst offers basic charting capabilities and a moderated user forum, Peters says. A "Basic" package ($99 monthly) offers 25 MB of storage, supports up to 10 users, and -- in addition to basic charting and a user forum -- offers "advanced" charting capabilities and email support.

Birst is available in "Professional" ($199 monthly; 100 MB of storage; 20 invitations; additional analytic capabilities and group management features) and "Groups" ($1,999 monthly; 1 GB of storage; 50 invitations; all of the "Professional" amenities, plus a dedicated database, automated report delivery; and phone support) flavors, too. Lastly, there's an "Enterprise" variant, which -- for $100,000 yearly -- offers subscribers up to 1 TB of storage; support for unlimited users; and predictive analytic/custom reporting and dashboarding capabilities, among other features.

SaaS BI first made a splash a few years ago, with entries from the former Siebel (now Oracle Corp.), Business Objects SA (now an SAP AG company), Informatica Corp., Oco Inc., Visual Mining Inc., and a host of other players. There's a sense, however, in which SaaS hype seems to have outstripped SaaS substance. You don't hear as much on the SaaS front from several companies compared to the zeal with which they promote their on-premises offerings.

Peters, for his part, says that one reason SaaS BI hasn't delivered on the hype is that comparatively few companies are actually offering BI-as-a-service. Instead, he argues, they're offering hosted BI via an Application Service Provider, or ASP model. "The challenge is that no one's really done SaaS BI. There are folks who are doing BI in an ASP model. There are folks who are doing SaaS, but what they're offering is prepackaged reports," he says.

The Birst approach, Peters stresses, offers full-fledged reporting and analytics in a SaaS model. It includes canned reports or dashboards and lets users (or their IT department overlords) create their own. More to the point, he says, it's designed to appeal to users who feel neglected by, or somehow left out of, their company's larger BI strategies.

"We think initial [uptake] will start in the departmental area. Obviously, we support some very, very large enterprise-scale deployments," he concludes, stressing that Birst -- which Success Metrics has quietly marketed to unspecified customers ("We have some very, very large, enterprise-scale deployments today," Peters avers) -- can scale to support "terabyte-scale" deployments.

"We see it [being adopted] typically in departments within organizations, either big organizations or small organizations where [users] don't control their own destinies, so they're relying on other folks to get value," he concludes. "The challenge they have is that they have to put in requests to some centralized organization, but the list of things that people want out of [this central IT group] is five times larger than they're able to deliver, so these [users] are left basically with dumping stuff into Excel and trying to do it that way."

Must Read Articles