Nigel Pendse from the OLAP Report gives practical advice to avoid the numerous pitfalls when buying BI products

In the article linked below Nigel Pendse has a few tips for people who are thinking of purchasing a BI product. He first outlines some of the pitfalls that are there if you get it wrong:

• At worst, the project could be a complete failure, with all of the software and services costs written off.
• Some of the project will survive, but the benefits will be minimal. This will ensure even less incentives for replacing this failed system.
• Some projects have big ambitions for company-wide rollout. Because of the failure, only a small portion of employees get to use the product, but if a wider rollout is required, the entire selection process will have to be carried out again.
• A vendor’s enthusiasm may encourage the selection team to purchase unnecessary licences when a simple solution (perhaps using excel!) may work just as well – or even better.

A summary of the tips follows (in no particular order). Sometimes it is not necessary that all of these steps are carried out in your company – you should make sure that your vendor has done it for you.

• Ensure that both end-users and your IT department are involved in the evaluations of the BI products and in the selection of the eventual product.
• Conduct a formal multi-product evaluation and don’t pick a product because of the company who develops it!
• Assess your needs – predict what your users really need instead of what they say they want.
• Sometimes your company may need more than one product – be open to this type of solution. It may save you money in the long run. Some power-users will have different requirements than most other users – don’t try to buy several licences of an expensive product just because it suits the power users. A simpler cheaper product may suit most of your employees.
• Don’t evaluate too many products in detail – some products can be eliminated by just looking at your needs.
• It is not necessarily the best option to buy from the largest vendor.
• Examine the financial costs – a good list of headings under which to do this is available in the article.
• Test at least two of the products – most of them have free licences for trial periods.

Don’t test every product you investigate – some of them should be ruled out at an early
stage. Three products should be the maximum you test.

Read the article here

http://www.olapreport.com/How_not_to_buy.htm.

[Update – you now have to subscribe to get access to the report at that link. According to a link on Timo Elliot’s blog, you can now access an older version of the article (more about OLAP than BI) here.]