BI OnDemand - Is it Good Enough?

A strong driver from increasing the value of outsourced solutions has been hosted, on-demand, software as a service [SaaS]. To date, the three main concerns about both SaaS and cloud computing environments are:

1. Availability
2. Integration with the Enterprise IS infrastructure
3. Security

Many are asking whether this environment makes sense for BI. There are two main streams of BI usage. Analytics and Operational BI [OBI]. In the case of BI analytics, the BI system need not be tightly integrated with the other enterprise processes and the business will not stop if there is unexpected downtime. This takes care of two of the main concerns of SaaS environments - availability and integration. With security in data centers today probably better than most corporates, I personally don't feel that security is a major issue today as it once was.

The other obstacle BI applications in an On Demand enviornment must meet are functionality and customisation or configuration to meet specific customer needs. Essentially the application cannot be customised, as there is a single instance of the application version being shared by all service customers. However, in terms of configuration, most BI onDemand providers are continually expanding options. Many meet the needs or smaller and medium size enterprises in most industries. In the case of larger, or more specialised businesses, onsite installations that can be customised more closely meet their needs.

BI differs in requirements from other popular on-demand applications, such as CRM, Sales automation, Customer support, Collaboration, Content management. Since requirements in these areas do not vary significantly from one business to the next, it is easier for SaaS application developers to create a one size fits all environment. With BI, there is more likley to be a need for a custom-tailored structure, making it difficult to standardize the offering, thereby negating the savings from standardization and automation. In addition, the former apps group tend to access data from a single data source, whereas BI requires data from right across the organization and beyond the corporate walls.

However, On-demand BI vendors have risen to the challenge and addressed these issues in various ways, including:

• Predefined Data Integrators - for specific applications, mapping data into a standard database design, for standard reporting and analysis.
• Wizard Driven Interfaces - allow users to set up data integration rules and define reporting requirements.
• Automated tools - that automatically import client data, analyze it and create appropriate data models and integration processes.
• Specialized analytical databases - conventional database structures in products such as Oracle, DB2 and SQL Server do not lend themselves to analytics as well as columnar and in-memory database engines. Automated tools allow simpler designs, easier automation, adequate performance without optimization, simpler adjustments and run on less expensive hardware than conventional technology.
• Automated opportunity discovery - one of the major roadblocks to BI adoption is that users don't know what to do with the insight they gain from the results on BI analytics. Automated tools mine data and automatically deliver suggestions that reduce sales cycles, improve customer satisfaction and define marketing campaigns.

In summary, whilst BI SaaS offerings tied to specific data sources and applications are most likely to succeed in the short term, longer term, the greatest potential lies with automated tools and analytical databases. These tools significantly reduce the work load and risk of integrating data and processes and can be substantially cheaper and more flexible than conventional alternatives.

One thing we can be sure of, BI on Demand is here to stay, and it is only going to get better, very quickly. If you are yet to be convinced - keep watching. Subtle advancements often bear significant improvements.

Back To Top

SaaS BI Index | NetSuite | Cognos Now! | SAS | SaaS BI Analysis