Why IT projects fail

Why do big IT projects fail? For that matter, why is there a high percentage of failure in most large consultancy 'led' change projects.

  • E-Mail
By  David Westley Published  June 24, 2007

Why do big IT projects fail? For that matter, why is there a high percentage of failure in most large consultancy ‘led' change projects. I have an answer - and, no it's not that all consultants are idiots (they are not), nor that they spend their time adding complexity to projects to maximize billing hours (which is, of course, absolutely untrue).

At a conference this week I was fortunate enough to be able to listen to the author of The Undercover Economist, Tim Harford, who uses the basic principles of economics to explain everything you ever wanted to know about modern life - or at least why coffee is more expensive in some places rather than others, why cheaper printers are actually more expensive to manufacture (I'll let you know at the end of this column) and why it doesn't only help to be mad (well, irrational) to play the stock markets, it's a requirement.

For purposes of this column however, it is Harford's take on the effectiveness of economies, and companies that is relevant. For illustrative purposes he used a ‘simple' cup of Starbucks coffee to explain what he meant.

Do you know how to make a Starbucks Latte? You may think you do, but have you ever milked a cow? Do you know the most effective way to get coffee beans to grow? And, can you really build machinery that can spit out hot air to steam milk...

The truth is something as simple as a Starbucks Latte is, when you break it down, very complex to put together. However, despite that, there is no one in charge of the Starbucks Vente Latte process from start to finish. Lots of people are responsible, at each part of the value chain, without a single presiding intelligence.

Of course now that I have written it, that seems obvious: the idea of someone being in charge - from start to finish - of distributing Lattes to London, Dubai or New York is ludicrous.

But is it really? In planned economies that's exactly what used to happen. When communism collapsed, economists from the Soviet Union flooded the United Kingdom - the home of Marx, Hobbes, Keynes, Locke and Smith - in the belief the country with the world's greatest economists could be mined for information as to how to unlock the great Soviet economy. One of the key questions they asked was: "Who is responsible for supplying milk to London?"

As supermarket shelves were left empty in the communist East because of great centralized plans that went awry, in the capitalist West thousands of companies were involved in getting hundreds of milk products to the supermarket shelves - some succeeding, some failing. Importantly, the failure of one company did not break the system - it just represented an opportunity for another entrepreneur to exploit.

Central planning invariably fails because it does not have the thousands, millions even of ‘feedback' channels that a completely decentralized, market-led economy has.

IT, and indeed big structural and change management projects often fail because they are plans to deliver milk to London. No matter how clever the planners, no matter how many MBAs they have, IT projects would be best left to an internal market - with multiple parties competing at different stages of the process. How to make that happen is too big a question for me, but a very nice bit of research for some clever PHd - or big research consultancy.

Finally, I promised I would let you know why cheaper printers (and cameras, etc) are sometimes more expensive to produce? The answer is that there is an additional manufacturing stage - the one where the manufacturer goes in afterwards to break something. With printers that invariably means disabling a microprocessor to make the product slower.

Why would a printer company do that? Simple - they have to find a way to get those people who are willing to pay more to pay more, while still having a lesser product for those people who are price ‘sensitive'.

Add a Comment

Your display name This field is mandatory

Your e-mail address This field is mandatory (Your e-mail address won't be published)

Security code