One of the biggest knocks on Software as a Service (SaaS)-based business intelligence (BI) applications is that they are less than ideal at integrating data from multiple databases and applications, both on- and off-premise.
Case in point: LucidEra announced it was shutting its doors earlier this year for what analysts say was the company's almost singular focus on integrating and analyzing customer billing and account data from just one source, Salesforce.com.
Richard Daley, CEO of open source BI vendor Pentaho Corp., which recently acquired LucidEra's Web-based reporting tool but not its assets related to the SaaS delivery model or Salesforce.com integration capabilities, doesn't think vendors that offer strictly SaaS BI have much of a shot at success.
Referring to SaaS BI vendors like Birst and PivotLink, "I wish those guys luck, but I don't see it [SaaS BI] as a big business," Daley said.
And some analysts agree. While SaaS BI deployments are gaining traction at some small and medium-sized businesses, as well as in departments and workgroups with limited IT budgets and resources, most large enterprises have shunned the SaaS delivery model for BI.
That's because SaaS BI is not ideal for environments with multiple, customized data sources -- which is the case in virtually all large enterprises and even mid-sized companies – according to Boris Evelson, an analyst with Cambridge, Mass.-based Forrester Research Inc. In addition, large enterprises usually have the resources for internal BI deployments, allowing them greater control over their data.
But, LucidEra's focus on Salesforce.com data and analyst criticisms notwithstanding, SaaS BI technologies are increasingly able to tap into multiple data sources, both on-premise and in the cloud, according to the vendors. This development is the result of work not just of the SaaS BI vendors themselves, but data integration vendors as well.
Earlier this month, for example, Pervasive Software unveiled a new product, Universal CONNECT , that the vendor says can integrate data from on-premise and cloud-based applications and data sources, including ERP, CRM and BI applications.
And, last week, SaaS BI provider PivotLink inked a partnership with Boomi, a specialist in cloud-based data integration. According to the two companies, the deal will allow PivotLink customers to integrate data via Boomi's AtomSphere application from more than 70 sources heretofore unavailable, including CRM applications from Siebel, ERP apps from SAP, Oracle and NetSuite, and HR management tools from PeopleSoft.
"This partnership is evidence that the cloud universe is maturing rapidly as point solutions give way to end-to-end analytic platforms that provide universal access to all the data needed for decisions," Quentin Gallivan, CEO of PivotLink, said in a statement.
According to Boomi CEO Bob Moul, Boomi's AtomSphere is a self-contained runtime engine that houses all the integration processes and mapping logic. Users can create connections from internal and cloud-based applications to PivotLink's BI tools via a Web-based portal with drag-and-drop functionality.
The runtime engine also checks for changes to integration processes and automatically makes updates, Moul said.
"We built Boomi from the ground up to be a SaaS multi-tenant platform," he said. Referring to the gaps in connectors among SaaS BI platforms and on-premise and cloud-based apps, he added, "that's exactly the gap we're trying to fill."
Other SaaS vendors are tackling data integration issues on their own. Brad Peters, CEO of SaaS BI vendor Birst, based in San Francisco, takes issue with the notion that SaaS BI tools are more difficult to integrate with multiple data sources. Birst has developed an ETL service that lives in the cloud and can integrate and model data from just about any source, he said.
"You should be able to do what you do behind the firewall in SaaS," Peters said. He said Birst has customers using the SaaS-based ETL tool and BI application to do complex data processes traditionally done with on-premise data warehousing tools, such as hierarchy flattening and modeling multiple survey data.
Still, not all SaaS BI vendors have sophisticated data integration components. And in some cases, IT and BI staff will need to customize data integration software to tap into SaaS-based BI applications, according Forrester's Evelson. There are also issues "getting huge volumes of data into the cloud simply due to bandwidth limitations," he said.
For companies that do choose to evaluate SaaS BI, Evelson said they should ask if the vendor provides flat file import capabilities – ideal for integrating Excel spreadsheets. They should also ask about SQL-based import capabilities, which are good for integrating data from relational databases and data warehouses that need little transformation; and/or extract, transform and load (ETL) capabilities, which allow for complex transformation of data being integrated.