Gartner identifies the fatal flaws of business intelligence

At Gartner’s recent Business Intelligence Summit in London, UK, analysts warned that although business intelligence (BI) offers great business opportunities, most enterprises are still failing to use it strategically. Gartner’s analysts identified the most common pitfalls of BI implementations and advised an audience of more than 750 attendees how they might be avoided.

Seven major flaws

1. If we build it, they will come
Gartner’s research director Bill Hostmann pointed out that data is fundamental to BI, but how the data warehouse is set up and run is also critical. Data warehouses are often created by the IT organisation with little or no business involvement. Consequently, users within the organisation often perceive them to be of low value.
“Too many IT departments build a data warehouse based on the assumption that once it is built, users will automatically see the benefit,” he said. “BI applications require a clear and intimate understanding of the business itself and it is only by working on business and IT issues in tandem that the real value of BI is realised.”

2. Managers need to negotiate the numbers
According to Frank Buytendijk, a research vice president at Gartner, too many people hide behind spreadsheets because they are used to them and because they know how to manipulate the numbers to satisfy the politics of their organisations. He advised enterprises to use the pressure of compliance to achieve greater things, such as cleaning up the many data silos, creating more ownership around performance data and eliminating many of the thousands of spreadsheets.

3. Data quality problem…we don’t have one
Gartner predicts that through 2007 more than 50 per cent of data warehouse projects will have limited acceptance, or be an outright failure, as a result of lack of attention to data quality issues. Furthermore, many organisations fail to see that they have an issue with data quality, focusing rather on identifying, extracting and loading data.
Hostmann is clear that data quality issues need to be addressed on an ongoing basis and that enterprises need to accept that these are not just IT issues.
“Consistency and accuracy of data remains the responsibility of the business departments operating the systems, not just the IT department,” he said. “New regulations and corporate governance have increased the demand and attention for data quality. The problem is that most companies continue to take data quality as a given.”

4. Our enterprise applications vendor will deliver the best solution
“All too often, enterprises assume that a one-stop-shop solution will be both the most cost effective and best solution for the business,” said Gartner’s research vice president, Lee Geishecker.
She warned attendees at the summit not to assume that major enterprise-application providers offering BI solutions and tools will always address all information requirements. “Although some enterprise-application providers can save you money overall, it isn’t always the case. Always compare your enterprise application vendor’s solution with that of a market leading speciality vendor,” she concluded.

5. Darwin was right – BI projects need to evolve
Business demands projects to be short and simple, and to have an immediate return. However, Hostmann said that this could easily lead to the wrong scoping of projects.
“Just as using the same templates time after time mean plans rarely evolve, building in the same limitations in to a new system is one of the greatest inhibitors to success. BI needs to evolve but BI projects should not - they should start and stop and not evolve,” he added.

6. We can outsource the whole thing
Gartner predicts that through 2006 less than 10 per cent of enterprises, where outsourcing could be a viable strategy, will be ready or able to outsource their BI applications and operations completely. Gartner said enterprises must define their BI key competencies and capabilities in order to determine what to in- or outsource. “As ever, the golden rule of outsourcing applies; avoid the temptation to outsource everything and only outsource things that are not a core competency,” Geishecker warned.

7. Just give me a dashboard!
A management dashboard should be seen as the finishing touch. “Companies must have a solid and stable BI infrastructure in place first.They should then create a networked approach where these new technologies are able to communicate with other BI technologies inside and outside the organisation, as well as with other technologies such as business process management and application integration,” said Buytendijk.