BI projects: what's the best approach to success? Top 5 tips on setting up and exploiting the modern datawarehouse.

There are many reasons why data warehouse projects fail. Many cite bad Vendor choice or poor design. But look closely and most Business Intelligence (BI) or datawarehousing projects fail because of more structural reasons. The organisations that succeed in the use of BI create the right foundations for success. Here are some of the key ones:

1. Get it together

There is no technology that benefits more from close integration of stakeholders than Business Intelligence. This can be formalised by the use of a BICC or other structures, but the key is to understand that IT needs to provide a flexible and responsive service to the business within defined boundaries. By formalising the link between IT and the business, results are faster and of higher quality.

2. It's agile.

Business Intelligence is not the same as application development. A common error is introducing structured methodologies which stifle one of the key elements of good BI--agility. It's interesting to see that in most organisations there's a positive relationship between use of BI and the ability of IT to deliver rapid and successful change. As a result, iterative development almost always has a key part to play in a successful BI project -and managing it successfully by keeping quality high is key.

3. Get your IT team mix right.

Delivering agile BI requires that your IT team has the right mix of skill profiles and on-off shore resource. A core of the team benefits from having skills throughout the BI delivery technology stack (design through to Database, ETL and Front End BI tool skills). This helps the iterative process as well as with quality and development speed. The person who runs your BI team from the IT side should also understand the business and be a strong communicator. The further removed from the business your developers and analysts, the lower the agility, quality and relevance of your BI.

4. Process.

Strong processes in an agile environment are paramount. From data quality and reconciliations to business ownership of the data through to rigorous quality checks. All and more are key to ensuring the optimum mix of agility and quality. In particular, policies and procedures where users help maintain the quality of data they work with themselves are key.

5. Start Smart not big

Big bang corporate datawarehouse implementation successes are very rare and for good reason--because invariably they break most of the rules above. Whether you are starting out on the road to Corporate BI or reworking what you have, ensure that the subject area you are choosing is:

* high profile and core to the business

* limited in scope

* using well understood and reconcilable data

A success in an area such as this is not only rapidly achievable, but will provide the spring board to get BI pervasive within the organisation and provide valuable lessons for future deliverables.

Ben Garside, Director of BI specialist consultancy, Morgan Benjamin