Subscribe
About

The seven deadly sins of BI

There is much BI disillusionment, driven by several consistent issues and errors.
Corey Springett
By Corey Springett, Strategic business manager at Progress Software South Africa.
Johannesburg, 22 May 2008

Report cards are in from all over the world, and they are not complimentary towards the current state of business intelligence (BI). Here are some of the negative comments being conveyed regarding BI:

* Response times are too slow (Nigel Pendse, as reported in the seventh BI Survey), and represent the greatest hurdle to user acceptance. Pendse reports that user times average between six and 10 seconds against small corporate data sets, and this while users are enjoying two-second response times from Google interrogating the entire Internet. This so-called Google Effect is causing users to reconsider their use of BI systems.

* Multiple single versions of the truth, as reported by DM Review (June 2005). This despite the promise of BI that it would give executives one, unquestioned view of the corporate truth.

* Business and IT are not communicating: users set up their own systems, or continue to use Microsoft Excel, and IT resents the burden users place on it (DM Review, June 2005).

* Limited adoption of BI, despite millions invested in it. User adoption ranges typically from 5.5% to no more than 12%, even though organisations acquire enterprise-wide licences costing millions (BI Survey 2008).

* 87% of BI projects in the UK do not live up to expectations (National Computing Centre, July 2007).

Users need far too high a level of skills required to make BI work, especially at the back-end.

Corey Springett is strategic business manager at Progress Software South Africa.

There is widespread disillusionment with BI, driven by a number of common and consistent issues, themes and errors. They can conveniently be termed the seven deadly sins of BI, and they will form the basis of seven Industry Insights to follow this one.

1. Skills: Users need far too high a level of skills required to make BI work, especially at the back-end.

2. Time: It takes too long to get started with BI, and it is inordinately difficult to do ad hoc queries. One of the issues here is the length of time and the level of skills required to design and produce reports. This places an intensive burden on IT and engenders frustration in users, who cannot easily and quickly gain access to the information and insights they seek. In addition, many organisations find the latency in BI unacceptable, and they are looking for near-time BI.

3. Cost: BI skills are costly and scarce, and tools and implementation run into the millions.

4. Ease of use - Users need a significant level of skills to ask the right questions, which means only a select few can make use of BI. In some cases, business users need to be able to write their own SQL instructions.

5. No mass adoption: Less than 10% of people in organisations make use of BI due to some of the factors addressed above.

6. Companies must have a data warehouse to do BI: A data warehouse can be a showstopper, with its technical requirements, cost and length of time to deploy. It should be relatively easy to do BI, without the need for a data warehouse.

7. BI delivers a single version of the truth: This is never the case, for a variety of reasons, which I'll explore in detail later in this series.

There is a new approach to BI, one which addresses all of the issues outlined above. It rolls out quickly, runs against production data without the need for a data warehouse, allows users to ask questions in natural English, removes the need for IT to program complex reports, has a Google-like interface, responds quickly to complex questions, and can help democratise BI with its intuitiveness and high performance.

Over the next seven Industry Insights, I'll analyse the seven sins of BI, and the alternative solution to them.

* Corey Springett is strategic business manager at Progress Software South Africa.

Share