In-Depth
BI for the Masses, Birst Style
SaaS BI specialist Birst practices what it preaches, officials say; it is, first and foremost, a general purpose BI platform.
The demise of software-as-a-service (SaaS) BI specialist LucidEra put new light on the still-incubating SaaS BI and DW markets. It also raised questions about the viability of the SaaS model -- questions, argues Brad Peters, CEO of SaaS BI specialist Birst Inc., that are, to a degree, inappropriate.
Peters says that LucidEra wasn't a SaaS B player in the strictest sense of the word. The chief problem with LucidEra BI, he contends, was that it was too specific: its targeted analytic applications limited its applicability to wider audiences.
For these reasons, Peters argues, LucidEra wasn't so much a SaaS but an ASP (application service provider) vendor.
Peters says Birst practices what it preaches. It is not a purveyor of analytic applications. Barring a couple of small exceptions, Birst does not market pre-fab, vertical- or industry-specific applications for pipeline analysis or any of a score of other practice areas. Instead, Peters positions Birst as a general-purpose BI platform -- in other words, precisely the kind of proposition that LucidEra itself was supposed to be.
"What Birst is is essentially taking the same kinds of user experience you got from Siebel analytics -- [so things like] schema generation and dynamic data warehouse construction under the covers," he observes. "You can pull in data from any different source; [the Birst service] will construct facts [and] dimensions for you -- and have the biz model flexibility to navigate data in the same way that offerings that are on-premises do."
Peters' criticism of the niche nature of LucidEra's analytic application isn't new. For example, unstructured data guru Seth Grimes, a principal with consultancy Alta Plana, has pointedly described LucidEra as a "sales analysis platform" -- rejecting its description as a "general-purpose" BI platform.
In a posting on the personal blog of BI industry veteran Timo Elliott (who doubles as senior director of strategic marketing for SAP Business Objects), ex-LucidEra employee Alex Moissis -- who also works in marketing with SAP Business Objects -- speculated that LucidEra foundered when it was forced to narrow its focus (away from that of multi-purpose platform designed to support a variety of analytic applications) "down to a single application area."
Moreover, LucidEra's competitors have made similar allegations; still-gestating SaaS BI start-up GoodData, for example, contrasts its catholicity with LucidEra's analytic "nichiness." That being said, Peters can claim especial credibility in this regard: he's a veteran of the former Siebel Systems Inc., where he helped launch Sales.com, Siebel's ill-fated ASP CRM entry.
In other words, Peters says, he knows what doesn't work when it comes to application hosting or (and he's on more tenuous ground here) SaaS; putting aside the reasons for Sales.com's demise -- which Peters insists had a lot to do with Siebel's unimaginative design and marketing -- the ASP model itself simply didn't make sense.
"Why aren't there any ASPs? Think about it. If I had you install MySiebel and you're going to run it [in a hosted context], how is it different? Fundamentally, you have the same knobs to turn, the same bells and whistles, the same software architected for the same deployment models; there's no differentiation there," he says. Factor in the operational strictures of your typical application hosting or ASP deployment -- where customers expect more aggressive service levels than they do from their internal IT organizations -- and the "cost justification of ASP got wiped out."
The genius of SaaS, Peters says, is that software deployment is engineered into the application service itself.
"Saas when done as SaaS, not as an ASP, is reengineering the entire software stack for delivery in a service-oriented model," he says. "There's a big difference between SaaS and [hosting applications on] virtualized hardware, and really the difference is that I've moved my hardware to the cloud and can provision it faster -- that's capital provisioning as opposed to application provisioning."
In other words, Peters contends, virtualizing an application instance and then hosting it -- which in a sense amounts to a nifty reprise of the defunct ASP model -- doesn't get you any of the benefits of SaaS. "If you came from the platform part of the world, you believe that just by making the OS virtual, you've solved the problem, but that isn't what happened. It took an application company [Salesforce.com] to build a CRM system that anyone bought."
That isn't the approach taken by Birst, Peters says. "Fundamentally, the entire stack is hosted, but we rewrote it from scratch to tie together the data management, the data integration, the OLAP and ROLAP, the dashboarding, reporting, and scorecarding all in one layer," he says.
Birst's selling point is that all of its capabilities can be rapidly deployed. That's also where most of the development heavy-lifting came in, Peters indicates: "The fact that you've turned the deployment of the application serviceable is where the expertise comes in. What that requires is people who know how to deploy apps and have spent a considerable degree of effort to make this thing work -- but look at the result! When you spin up salesforce.com, it's like that; but if you try to spin up SAP -- wow, nobody's ever said SAP is to spin up."
Outside of a few analytic applications (specific to wealth management and insurance), Birst doesn't market industry applications (ala LucidEra or SaaS competitor PivotLink); Peters says Birst prides itself on its emphasis on the basics.
"We focus more on the traditional [BI] stuff -- like reporting. We actually do reporting, because it's important, and because it's what our customers need," he says.