Serving the Underserved: Is On-Demand BI the Answer?

By David Hatch, Guest Author

In my last post, I discussed how On-Demand BI may not be just for SMBs, based on some recent research findings, and that corporate plans for On-Demand delivery of BI applications (SaaS, Hosting, BI-Appliances) may be gaining ground within organizations of all sizes more rapidly than has been generally reported. This is being driven by a reported lack of BI skill sets among non-IT information consumers, and challenges surrounding data quality and integration.

The combination of limited skill sets and data-related issues, with an increase in On-Demand BI capabilities (Moore’s law at work: faster Internet, improved data compression, and advanced zero-footprint application options) has catalyzed the emergence of a variety of solutions from both new and entrenched vendors. Over time, reporting and analytics solutions have grown from an exclusive tool for power users to a critical business application for all knowledge workers. Yet, as information access is demanded “when, where and how” knowledge workers want it, companies are struggling to find the BI and IT skill sets required to provide and support a geographically dispersed organization with a wide variety of business roles and information needs.

On-Demand BI represents several new options for delivery of actionable information, including software deployment techniques such as software-as-a-service (SaaS), BI appliances (two flavors – software and software/hardware combination), hosted/ASP (dedicated application hosting on a per-customer basis), and hybrid options.

The BI market’s established vendors, such as Business Objects, Cognos, Oracle/Hyperion and SAS are introducing On-Demand options, while smaller providers like SeaTab, Oco, LucidEra, Dimensional Insight, and OnDemandIQ are offering BI as a complete On-Demand solution. Additionally, many software providers have developed or acquired robust training services to address skill set limitations.

Consulting firms such as Accenture, Avanade, BearingPoint, EDS, and Hitachi Consulting offer focused BI services aimed at alleviating the strain on an organization’s IT resources while delivering complex large-scale solutions that are customized to a company’s needs.

Aberdeen research has revealed that the primary users of information are also the most underserved when it comes to accessing BI applications. Not surprisingly, “line-level knowledge workers” are identified as one of the top users of information; however, they are also the number-one group identified as being underserved by current BI capabilities. Notably, all groups were identified as being underserved by BI applications at a greater rate than they were reported as being primary users of information.

What does this say? The research suggests that traditional approaches to expanding the use of BI applications to the enterprise have largely failed. Large, complex installations are proving to be beyond the existing BI skill sets and data integration and quality capabilities of most organizations. The research also shows that there is a “buy it” vs. “build it” dichotomy developing. Companies are either taking an outsourcing approach, and attempting to find external solutions to the current challenges, or they are bringing in training and institutional programs (Six Sigma or BI Competency Centers) to raise the levels of existing skill sets.

It is yet to be determined if an outsourced approach, via On-Demand BI offerings, will alleviate these obstacles. The ability to customize a SaaS BI application to align to an individual business strategy is not as flexible as an on-site implementation approach. But, if users are not able to access or fully use existing custom applications, this issue becomes less important. It will be interesting to see if an On-Demand approach continues to gain traction with BI users, or if the most successful implementations continue to arise from on-site customized applications. I will continue to monitor this activity through upcoming research and survey analysis.