Seven Common BI strategy Mistakes

By Clive Margolis

In a recent study Gartner, Inc., a leading IT research and advisory company, lists defective strategy – or not having a strategy at all – as one of nine “fatal flaws” in Business Intelligence (BI) project implementations. A good BI strategy, on the other hand, is the key to a successful implementation, helping to maximise return on IT investment.

How do you know if your BI strategy has defects? Below are seven telltale signs – how many of these apply in your organisation?

1. BI system not seen as relevant
This often indicates a lack of consultation across the organisation. This results in a lack of buy-in, and potential users do not see what the BI system has to do with them. If the beneficiaries of the system are not involved in the design the system is unlikely to be successful when implemented. The BI strategy, if it exists, does not include the requirement to widely consult internally.

2. Seeking strategic advice from the vendor
An over-dependence on the software vendor can leads to a conflict of interest where the vendor siezes the opportunity to get what it wants (more sales) at the expense of what the organisation needs (an appropriate use of the software).

A common example of this would be the overselling of an OLAP cube solution where a reporting application would do the job better for the users concerned. OLAP cubes provide highly summarised information in hierarchies, and there are many advantages to this, but ease of reporting is not always one of them. Furthermore, the number of users for whom OLAP cubes are appropriate is almost always less than those who require a standard reporting application.

3. Continuation of the past
Doing things “as they have always been done” is the enemy a of good BI implementation. It can all too often lead to attempts to cut budgets, and often does lead to adoption of inapproprate decisions such as reporting from spreadsheets where what is needed is a fair appraisal of the available BI delivery tools. The use of existing solutions simply because they are currently in place is a sure sign of a strategy defect, and it is vital to be aware of the latest BI methodologies and applications.

Changing one’s habits is often not easy, and may well require training, but a certain degree of change is necessary if you are to maximise the benefits of BI to your organisation .

4. Hot-air balloon management
When budgets start to get out of control, you start to see a lot of pet projects thrown overboard. This is what I call management by “hot-air balloon”. As with a hot air balloon, throwing things out of the project make it a lot lighter and more able to float. The trouble is, this process also makes the BI project a lot less functional and effective.

The problem, more often than not, is failing to correctly assess the cost of the project in the first place. Could the project have been split up to begin with to make costs more manageable? The message here is, it’s better to promise less and deliver, than to promise the world and let people down when economic reality hits home.

5. Information silos
An information silo occurs when data from one computer system is unable to communicate with another. This is the type of problem that a Business Intelligence project with a good strategic plan is able to identify and correct.

Siloed information is very expensive to an organisation because the benefit of information within the organisation is hidden to parts of it. The issue is usually resolved by making the strategic decision to create and maintain a data warehouse. A data warehouse collates data from diverse places into a series of one or more ‘data marts’, in the process reforming and recalculating the data to make it more useful to Business Intelligence.

6. Inaccurate metrics
Metrics (I use the word as BI vendors such as IBM Cognos use it) monitor month-on-month changes in performance and are often based on calculations between different items of data. They can become inaccurate in a number of ways. For example, the data may be entered incorrectly, or miscalculated due to the metric being defined incorrectly. Or the calculation itself can simply be incorrect. Often there is no single definition of a metric – a strategic error – or they may be over-reliance on an inherently unstable data source such as a spreadsheet. These are strategic errors, because they can be remedied by policy decisions such as avoiding spreadsheet input wherever possible and publishing a dictionary of metrics for the organisation as a whole.

7. No ongoing development of the BI system
Unlike, say, an accounting system or a CRM (customer relationship management) system, Business Intelligence systems add value by being reviewed and enhanced on a continuous basis. Projects have a ‘lifecycle’ which involves reviewing your BI needs regularly. New metrics and KPIs (key performance indicators) are discovered, these need to be added to the data warehouse and made available to the appropriate managers. New data marts which were not in the original project scope might come on board in a later phase.

These developments should be considered as part of the overall strategic thinking so they can be correctly prioritised and budgeted. This gives BI a more central, and less peripheral role within the organisation, which allows it to reach its full potential in adding value to your organisation.

The role of Business Intelligence within the IT environment is continually evolving and increasing. By adopting effective BI strategies you maximise the potential that BI has to offer in terms of productivity, cost saving and in numerous other ways.