Too Big To Succeed

Feb 25, 2010 09:16 · 835 words · 4 minute read

William Heath has put up a post called “What the smart government IT supplier needs to say in 12 weeks’ time”. I started a comment there which grew to the size of a post, so I figured I might as well put it up here as well.

He asks the question:

“Now: here’s the crux. Britain’s new post-election government may be pretty hostile to its IT suppliers. Whichever colour it is it faces the same problems, but let us assume for sake of argument it is Conservative.

Relations have not improved since the unseemly spat between Intellect and David Davies over ID System contracts. Big IT suppliers and their big bills are definitely seen as “part of the problem” in Tory HQ, as is the trade association, and an ineffectual (overpromoted/overpaid) CIO culture and the excessively big, out-of control IT projects they have cooked up.

What is a smart government IT supplier to do in this situation?”

I suspect that the big IT houses are going to be having more and more conversations with people like James Gardner, a former banker who is now Chief Technology Officer at the Department for Work and Pensions.

Earlier this month he posted about his experiences of spending a week in a Job Centre somewhere in the Rust Belt of Scotland.

Apologies for quoting from his post at length, but I think this is a significant illustration of a mind shift taking place:

“But here is another thing I’ve found in this Job Centre, and it is something I’m not surprised about.

Staff build their own stuff to get around the limitations of systems we provide. There are Excel based spreadsheets which are used for diary management (“oh, I can’t have this open too long, otherwise no-one else will be able to make appointments”). There is email based workflow, where each step is a new inbox that gets manually monitored. And there’s any number of self-made data capturing things that are used for statistics and business reporting.

And all of it is stitched together with another technology: paper. They create their own forms, and their own paper based systems in order to supplement their jobs.

Consequently, the work is processed in a highly efficient way. I’d make a guess that each JobCentre does things slightly differently, depending on how good their custom additions to each of our centrally provided processes are.

If there was ever proof needed that decentralisation of the core is a good thing, then I’ve been immersed in it for the week so far.

I wonder what would happen if we put the appropriate end-user computing tools in the hands of these people and said “design the perfect Job Centre system”. My guess would be something good.”

Having been involved in the peripheries of Big Projects in the past, I’ve often wondered if the reason that they fail is linked to their sheer size and the capacity of an ordinary human being to cope with the scale.

Beyond a certain size, it seems that the probability of success by any definition tends to zero, and no amount of tinkering with the political complexions or terminology or methodology-of-the-month will change that.

At some point in history, the processes that cause these problematic systems to be created in the first place either didn’t exist, or were paper-based. That suggests a couple of questions.

Perhaps if the processes didn’t exist before IT, they can’t exist after IT - because they’re too large and complex to be administered in the first place.

And maybe moving from paper to digital processes doesn’t actually increase efficiency once you account for the eleventy-billion pound cost of the digital process itself.

Sure, there’s a superficial improvement by virtue of being able to call up a record on the screen rather than retrieve it from a shelf somewhere - but the true cost of the operation is a lot greater than the immediately-apparent interaction would suggest.

I spent half an hour yesterday trying to update a gas bill online, and failing because the British Gas system didn’t like the combination of address and name on the account. Talking (eventually) to a call centre agent, it turns out that this is a routine problem for them.

Viewed simplistically, the online process is a Good Thing, because it’s more (apparently) efficient. But viewed holistically, it’s a disaster, because it’s creating additional work for both customer AND organisation. The issue is that the process won’t ever be seen holistically, because the organisation isn’t taking into account my contribution.

What all this suggests - it’s just my gut feel, I’ve no empirical evidence to back this up other than a hunch - is that the days of the monumentally collossal top-down design is over, simply because they’re too expensive and too complex to work.

That the future is more likely to consist of patchworks of systems - and a realisation that apparent inefficiencies aren’t actually curable without spending more time and money than was wasted in the first place.