So what is Business Intelligence actually? Why do 80% of corporate BI projects fail?

Probably one of the worst assumptions you can make as a consultant (or employee) is assuming that your customer (or employer) knows exactly what you are talking about when you use terms that are in common usage, but that have no universal definition. One of those terms is Business Intelligence. 

Your personal understanding of Business Intelligence will likely be biased according to your particular business discipline. As an IT person, Business Intelligence may mean the technology to present information, while as an operations head, Business Intelligence may mean access to information relating to Six Sigma or Total Quality Management methodologies. To the CEO, Business Intelligence may mean access to information concerning whether the organisational strategy is on track, and some key operational figures. An internet scan presents various Business Intelligence definitions. Here are but three:

• Wikipedia ... a set of theories, methodologies, processes, architectures, and technologies that transform raw data into meaningful and useful information for business purposes
• CIO.com ... an umbrella term that refers to a variety of software applications used to analyze an organization’s raw data. BI as a discipline is made up of several related activities, including data mining, online analytical processing, querying and reporting
• Gartner ... an umbrella term that includes the applications, infrastructure and tools, and best practices that enable access to and analysis of information to improve and optimize decisions and performance

The more we look for definitions, the more we find a technology leaning to a definition for Business Intelligence, which probably makes sense given the inclination of the industry to develop interesting names for regular business processes, such as Business Intelligence (BI), Customer Relationship Management (CRM), Customer Experience Management (CEM) and so many more. There are however one or two common threads amongst most of the definitions, and these are DATA and meaningful INFORMATION. Whichever definition you choose, as long as it has these two elements in, you're unlikely to be too far off the norm in terms of a reasonable definition for Business Intelligence.

- x - x -

From a holistic business perspective, an Enterprise Architecture model (a methodology which provides a particular interpretation of the strategy of the business in terms of elements that will be required to change in order to realise the business strategy), begins to demonstrate why so many Business Intelligence programmes fail, especially if they are considered as technology projects instead of as (strategic) business projects. For starters, here are but a few Business Intelligence issues that would emerge as part of an Enterprise Architecture approach to data and information:

• Technology is not able to define what should be measured. Business specifies this ... are we measuring strategy performance, tactical performance, operations, or everything?
• Technology is not able to define how things should be measured. Business specifies this ... for example, the age old definition problem of when exactly a sale has occurred and how it is measured differs for every business, and has nothing to do with the type of technology implemented 
• Technology is not able to define who should be able to see what information. Business specifies this in the context of its data governance policies. If it doesn't have any, best it gets a move on and define them as data governance is a critical aspect of proper information management
• Technology is not able to define why a particular piece of data or information is important. Business assigns a context to data and information. For example, while working capital may follow a particular cycle in a business, only a person familiar with the operations of the business will know whether the working capital figures are in line or whether they have gone awry
• Technology is not able to define when and how often an element of data or information should be updated or refreshed. Business determines this. For example, cash flow is often measured daily, if not hourly or even realtime depending on the operation, while the business' fixed assets may not change for many months at a time and thus only needs to be updated once a month
• Technology is not able to intepret why a piece of data or information is important, and what it means. Again, business determines the context. This is not only true in a reporting context, but also in an analytics and data mining context

In the 'regular' definition of Business Intelligence, an entire context of an Enterprise Architecture is missing. Some may say it's implied in the definition ... perhaps ... but an awful lot of Business Intelligence projects I've come across have given these areas but scant attention. Indeed, many simply try to recode existing operational reports without asking key questions such as whether all of the reports are relevant, accurate, or sufficient. Given this, it's no surprise really that so many Business Intelligence projects fail ... as the regular definition  only gives a fraction of the true picture of the true scope of a Business Intelligence programme.

For success in Business Intelligence projects, it is not only important to properly understand what is meant by Business Intelligence, but it is also important to understand how Business Intelligence fits into the operating context and strategy of your organisation. In enterprise strategy speak, if you implement Business Intelligence without a proper appreciation of aspects such as the context layer of the Enterprise Architecture, then your Business Intelligence project is more than likely to be one of the 70%-80% of corporate Business Intelligence projects to fail (Gartner ... see below).

After 20 years of being involved in Business Intelligence in some form or other (from the days when it was still called Management Information or Enterprise Information), note that there are many more reasons that Business Intelligence and similar projects fail, such as poor scoping, poor expectations management, poor project management, poor stakeholder management, poor measure definition, poor source data management, lack of a single source of data, lack of commonly agreed measure definitions, poor presentation layer development, poor business problem definition and probably the worst, failing to see Business Intelligence as an integrated part of the fabric of the entire organisation.

The purpose of this article was to highlight that it is important to understand that the mainstream definitions of Business Intelligence are arguably incomplete. Use them as a guide, but they exclude a whole business world of issues which need to be addressed before anyone even begins thinking about any single technology aspects! Get the business part right first, then the technology part will be sure to be a success, ranking in the 20% of positive outcomes for Business Intelligence projects.

Between 70% to 80% of corporate business intelligence projects fail, according to research by analyst firm Gartner.

A combination of poor communication between IT and the business, the failure to ask the right questions or to think about the real needs of the business, means most business intelligence projects fail to deliver, the firm says. IT departments make the mistake of looking at business intelligence as an engineering problem that requires a specific package solution. Business intelligence is not a crystal ball that pops out the answer. People in IT need to stop approaching business intelligence as a vendor or engineering solution, or as a tool. 

Source: www.computerweekly.com Jan 10 2011