Having yesterday been somewhat disparaging about the efforts of others to delineate the reasons for BI projects failing, I realised that I had recently put together just such a list myself. By way of context, this was in response to being asked for some feedback in a subject area where I had little expertise and experience. Instead of bailing out of answering, I put together a more general response, a lightly edited and mildly expanded version of which appears below.
Please note that there is no claim on my part that this list is exhaustive; in common with all humans, us IT types can be very creative in finding new ways to fail, I am sure there are some out there that I have not come across yet.
- The objectives of the project not being clear. By this I mean the business objectives. There are two layers of problems, the actual business issues may not be understood well enough to form an effective response and, if the business knows what it needs to do in general terms, IT may not fully appreciate this for a number of reasons (mostly down to lack of communication) or may be unable to translate this into a suitable programme of work (possibly because of a lack of knowledge of how the business operates). Where IT is not part of the senior management team, or sees itself as a department apart, this issue is more likely to occur.
- Strategy formation being skipped. If you don’t understand what a project is meant to be about, it is difficult (to say the least) to form a strategy. However, if the test in point 1. is passed, then it may be tempting (or there may be pressure applied) to get to the end game as soon as possible without either forming a strategy for the project, or fitting this into both over-arching business and IT strategies (which one fervently hopes are complementary). As I know all too well, the strategy formation step can be tough one and people may sometimes be keen to skip it. The current economic climate may lead to this happening more frequently and my opinion is that this will be storing up trouble for the future.
- Fragmented systems’ landscapes. Related to the above, it is often very difficult to make progress when there is a patchwork of different systems and approaches throughout an organisation and little desire to address this short-coming. Often some sort of revolution (albeit sometimes a quiet one sustained over many years) is required to move forward. Sometimes this requires some crisis, internal or external, as virtually every organisation is inherently conservative; no matter what their marketing spiel may claim to the contrary.
- Lack of Change Management. Projects often also have an organisational change aspect (what else are they for?) and the problems here are: a) people do not like change and resist it; and b) many otherwise able managers are not experienced in change – indeed we tend to educate most managers to be efficient in a steady-state environment. Even when this aspect is recognised, it is often underestimated and work does not start until too late in the game.
- People. Aside from these, the main other problem is people. Projects, even small ones, are difficult and not everyone is up to running them. Simple errors in execution can derail projects that otherwise tick all of the boxes.
Of course any passing Gartner analyst is more than welcome to rip this to shreds if they see fit.