DMA's systems integrator implemented row-level security, meaning that a customer can only see the data they are authorized to see, he said.
About 1,400 user-generated reports are now in the system, and most of DMA's customers "are blown away" by the capabilities, he said.
Other SaaS BI users report similarly satisfactory results.
Children's Choice, a chain of corporate child care centers, has been using a SaaS BI tool from Birst for several months, said Dan Lawler, chief financial officer.
Staffing costs are a key concern for Children's Choice. While regulations tend to dictate certain employee-to-child ratios at the centers, the company is always looking to avoid overstaffing, Lawler said.
It is using Birst to surface information from the company's custom-built time and attendance application, which tracks the comings and goings of both staff and children, he said.
Back at headquarters, Lawler can use Birst to monitor staffing trends at the centers. "This allows us to help [center managers], to send them little nudges and notes saying 'Hey, you may be overstaffed in this classroom," he said.
Like DMA, Children's Choice had little interest in hosting a BI application in-house, and worked with a partner on the implementation.
Security was a key concern. Birst is not extracting any personnel data regarding employees, nor other sensitive material, such as pictures of children who attend the centers, he said.
Birst should serve Children's Choice's needs over the longer term, according to Lawler. The company has 40 centers now and is hoping to add eight to 10 more each year.
"Our growth is not so aggressive that we don't think someone like Birst can handle it. We think they're going to meet our challenges," he said.
But it is far from guaranteed that every SaaS BI project will be successful or even appropriate for a company's needs.
For businesses that have just a handful of packaged applications that haven't been heavily tweaked, pushing information from them into a SaaS BI application should be fairly straightforward.
However, if a company has a wealth of customized applications and data models, the time and cost savings presented by the SaaS deployment model are sunk. That's because it would be necessary to expend plenty of effort cleaning up that data for use in the BI application. Forrester estimates that data preparation can take up to 80 percent of a typical BI project.
"In such situations you have to ask yourself why it would not make sense to just go the last mile, and create an internal BI application, leveraging all the clean data and metadata you’ve already created," Evelson wrote.
Forrester's report goes on to make a plethora of recommendations about how to choose a SaaS BI vendor, whether it's a platform giant like Oracle or one of the dozens of pure-play vendors in the market.
An application's breadth of functionality should be fully considered. Some customers may only need an application to pump out reports and dashboards. But full-featured BI software adds extras like text analysis, ad-hoc querying and individual workspaces, the report states.
Potential hidden costs present another problem. While SaaS typically provides the benefit of lower up-front costs, some application vendors may require customers to buy additional licenses if a BI application is going to access it, Forrester notes.
Security could be another deal-breaker. Beyond general queasiness with hosting potentially sensitive data on someone else's server, there are regulatory hurdles as well. Some countries don't allow certain types of client data to leave its borders, Forrester notes.
DMA, for one, was not concerned about security issues, since PivotLink meets the SAS 70 Type II auditing standard, Evelson added. "I would expect they do a lot better at protecting data than other companies. Their business model dictates it."
Finally, as with every emerging industry, there are more winners than losers. Your SaaS BI vendor could go out of business, as startup LucidEra did last year.
Therefore, a plan B is a must, according to Forrester. Companies should not only place the application's source code into escrow, but also periodically test the process of moving data in and out of the system or to another BI application.