Friday 3 January 2014

Tackle the most difficult cases first

Just a few weeks before Christmas, there was a thrilling interview by Decca Aitkenhead in the Guardian with Louise Casey, the head of the government's troubled families programme.

I realise that 'thrilling' is not a usual epithet to use about this kind of thing.  I found it so because it was authentic and because I admire Louise, which is why she played such a key role in my book about the future of organisations, The Human Element.

Perhaps I also imagined some echo of the thesis I made there when she said:

"All of what we do turns on something very simple: the relationship between the worker and the family."

I believe that to be the case, and it is a revolutionary doctrine to be articulated from inside government because so many of the approved systems imply something so different, and - as I believe - very much less effective.

What I originally found so interesting about the way that Louise Casey worked, back in the days when she was tackling rough sleepers, was her refusal to accept more than the most basic official targets.  She did this, as she does everything, through sheer force of personality - but the reason is illuminating: because projects of that kind need to concentrate their resources on the most difficult cases first.

If you do that, then the less difficult cases tend to get swept up in the process, but if you are bound by targets then you tend - encouraged by the less broad-minded consultancies, McKinsey for example - to focus on the easiest first.  Especially if this is a payment-by-results contract.

What happens then is that the most difficult cases just get further entrenched, while government resources are poured into supporting mild cases.  You can see this absolutely everywhere: it is part of the tragedy of UK public services in the Blair/Brown years, and it has still not really been tackled.

So when I heard that the Universal Credit's delivery chief has decided to test out the system on the easiest cases first, you kind of suspect the game is up.  The system thinker John Seddon suggests it should be the most common cases, and his newsletter which discusses this is not online, so I can only quote from it, explaining why overpayments happen:

"Overpayments occur because successive governments have industrialised tax and benefits services; no longer do claimants have a relationship with people who administer credits and benefits; many of the errors are down to systemic failures in administration: badly designed services, which, for example, take so long that peoples’ circumstances change during processing and ‘back offices’ that apply rules (instead of meeting people) which fail to take account of variety, and some are due to the difficulty people have accessing services. In fact underpayments are as big a problem as overpayments, but Duncan Smith doesn’t mention that. Although he doesn’t say so I think Duncan Smith wants us to believe we have overpayments because people are naughty (‘scroungers’). There may be a few, but an IT system is the least efficacious way to catch them."

If you notice the parallels here between what Seddon says and what Louise Casey says, then I hardly need to go on - the vital importance of face to face relationships, as the best and ultimately cheapest way to deal with diversity.

If UC is going to be tested on the easiest cases first, the writing is already on the wall - which is a pity, because it is a radical and important idea, disastrously implemented.

And if you are still wondering why the DWP is boasting that it has an IT asset worth £150m after writedowns, take for a moment the case of the US insurance giant MetLife.  They have a new app that allows them to see into all their 70 different and incompatible databases and see what each customer needs.  Guess how long it took to build?  Ninety days.

Anyone who persuades the government that we need to mortgage the nation on huge new IT systems that can't deal with variety, and therefore take years to kick into some kind of miserable life, may now need to be locked away where they can do no more harm.  It is like selling crack to an addict.  There is only one solution: just say no.

Subscribe to this blog on email; send me a message with the word subscribe to When you want to stop, you can email me the word unsuscribe.


BruceK said...

This problem is in no way restricted to large organisations. I work at a small but highly target driven company, and the sidelining of difficult cases which you describe is a familiar problem.

One of the worst aspects of this is trying to cure the disease with more of the same - setting up another target, which inevitably gets manipulated in turn, and then another, and then another.

Mark Pack said...

Doing difficult cases first is one way to avoid sidelining them, but it's not the only way.

Especially when building IT projects, tackling the easy cases first can have much merit. It gets you a working system sooner, and by speeding up or automating easy cases it can free up staff time for the difficult cases.