Project leaders must know their limits

Know your limits

Overconfidence is an epidemic in corporate life, and it has been understood and documented for over a decade. The poor performance of many mergers and acquisitions has been attributed to chief executive overconfidence, although I think the researchers failed to understand that many acquisitions represent an attempt to bolster share price and deliver a handsome bonus to senior management.

But I digress. You would expect IT departments in large organisations to also be plagued by this particular behaviour ­ and you’d be right.

The yardstick for measuring overconfidence is confidence limits.

Question: how wide is the milky way? If corporate life has got into your bones you’ll no doubt be trying to produce a good guestimate. Well I’ll put you out of your misery: it’s one hundred thousand light years, give or take.

It really doesn’t matter if you got the right answer, what is more important is the way you tried to determine your confidence limits. If they were tight, then I’m afraid you are probably the overconfident type ­between zero and infinity would have been a perfectly good answer.

The most blatant manifestation of overconfidence in IT is seen when projects are allotted a budget and timescale. Eighty per cent of projects run wildly over budget and their allotted time ­ maybe the way we determine these metrics is more at fault than the actual delivery process.

Perhaps the most surprising demonstration of overconfidence is the appetite for large IT projects in many organisations. These tend to take the shape of large infrastructure projects and the deployment of application suites.

The risks associated with such wild adventures are very large, and yet it seems that senior management will rubber stamp such activity quite willingly. The nicely presented proposal and slick PowerPoint presentations simply serve to spread the overconfidence from the proposer to the board ­ and it often works.

A long list of benefits will be detailed with a delivery date and costings to the nearest penny. This is overconfidence in its most acute form.

I really don’t want to knock down this particular house of cards without proposing an alternative ­ although I’m fully aware that any method that cannot be accommodated within a spreadsheet will be dismissed out of hand.

Since large IT projects are risky ventures, and since there are many unknowns, it might be more sensible to try to accommodate the unknowns in determining the cost and timescale of a project. You could propose an expected budget to satisfy your CFO’s need to have a figure in a spreadsheet. However, you will also deliver confidence limits with some idea of probability.

Obviously there will be a lower limit and an upper limit and some notion of the likelihood that these will hold. I have actually met organisations that work this way, but they are the exception.

For most organisations, there will be all kinds of barriers to a more mature approach. IT will often find itself bidding for work that might otherwise go to an external supplier and will want to be seen as competitive and confident. And we all know what that means: unrealistic timescales and budgets. External suppliers will have no such problem ­ particularly if they are funded by some fairly senior figure in the organisation. When things start to go wrong, the internal sponsor will juggle budgets and perform other amazing tricks to make sure the project is delivered.

Overconfidence is a form of corporate insanity, but the chances are that you are stuck with it. Even so, it is useful to know that the condition has been diagnosed and a cure is available.

Martin Butler is the founder of analyst group Vanilla Research