I just came across this report from the UK National Audit Office that examines the UK experience over the past 8 years as they tried to move many administrative activities into a shared service center.

In a nutshell, the study concludes that, despite significant cost and effort, the planned benefits of the initiative have not been achieved. By creating complex services overly tailored to individual departments, the government has increased costs and reduced flexibility. There has also been a failure to develop the benchmarks necessary for measuring performance and to date, the government has spent far more on the initiative than anticipated.

In some ways, I’m not surprised. The whole point of having a shared service center is to monitor performance continuously and optimize to common processes. If you’re customizing at every step and you aren’t tracking performance, you shouldn’t be surprised that outcomes aren’t what you expected. One government agency that has done this quite well is NASA — fifty six services consolidated across ten distinct centers with rigorous measurement and transparency in place.

So what’s the moral of the story? Shared services can work in government, but require some very deliberate steps.