Good Data, for example, said it would grant free access to its SaaS-based, multi-tenant analytics engine for up to six months to any former LucidEra customer that wanted it. Birst, PivotLink and other SaaS BI providers made similar gestures.
Industry watchers, meanwhile, were quick to point out that LucidEra's demise was a result of the vendor's narrow focus on Salesforce.com data and pipeline analysis and was not due to any inherent flaws in the SaaS BI delivery model itself.
While SaaS-based BI may have been a good fit for LucidEra's customers when they signed on with the now defunct vendor, circumstances change, Evelson said. On-premise BI – either a vendor product or homegrown applications – might now be a better option for those whose reporting and dashboard requirements have evolved.
SaaS BI is ideal for small companies whose corporate data already resides in the cloud or that have just a few, relatively straightforward "out of the box" applications and data sets, according to Evelson. Many LucidEra customers – none of the several contacted for this story would comment -- probably fit one or both of these scenarios when they chose the SaaS BI route. It is relatively simple to put SaaS-based reporting tools on top of such applications and data sources, Evelson said.
But as companies grow – adding multiple and often customized financial, CRM and ERP applications, for example -- it becomes more difficult for SaaS BI to support them, he said.
"All BI SaaS vendors offer basic import capabilities of source data from a variety of sources like relational databases, spreadsheets or text files. But data rarely exists in such neatly organized formats, and in one place," Evelson cautioned in a report released earlier this year.
Similarly, many SaaS BI vendors are geared to particular vertical market reporting and analysis – data from Salesforce.com in LucidEra's case – meaning it is not always possible to expand SaaS BI software to other uses, such as reporting and analyzing financial, human resources or inventory data, according to David Hatch, an analyst with Boston-based Aberdeen Group.
"The ability to customize the solution is very limited. It's going to exist based on how the SaaS provider implements it, not on how you want it," Hatch said. "The problem is outgrowing their level of expertise and their ability to meet an organization's reporting requirements. Whether or not they go out of business [as LucidEra did], that's really a horrible situation to be in."
Forrester's Evelson also said that some departments and workgroups turn to SaaS BI when their company's on-premise BI applications of choice don't meet their specific job requirements or when waiting for IT to support their project would take too long. Former LucidEra customers -- or any company -- in that situation should reevaluate their company's "internal enterprise BI solution," Evelson said, which now may be up to the job.
Companies that do ultimately decide to go the SaaS BI route should hedge their bets, when possible, with escrow accounts, said R. Ray Wang, also a Forrester analyst. An escrow account, in the context of enterprise applications, is essentially an agreement between the customer and vendor that if the vendor goes bankrupt or is acquired, the customer is entitled to recover its user data and metadata, the applications' source code, documentation and any application executables, Wang explained.
Escrow accounts are especially critical for SaaS enterprise application customers, he said, because many vendors in the space are not yet profitable, and it is impossible to know which ones will be around in five or more years.
Wang said very few SaaS enterprise application vendors that he's aware of currently offer escrow accounts, but they are beginning to catch on. As SaaS gains mainstream acceptance and vendors begin selling to CIOs and CEOs -- not just department heads, who are more concerned with vendor viability -- more SaaS vendors are likely to begin offering escrow accounts more frequently.
In any case, all companies negotiating with SaaS BI vendors should ask about them. "What good is your data if the vendor goes bankrupt?" Wang said. "Not much!"