For most in Government IT, disaggregation has been a hot topic and is a live goal for nearly all of them, even those busily extending their contracts with incumbents so that they can buy time to disaggregate properly, as I wrote in June 2013 for instance.
Concentrating power in big, slow moving central organisations has, traditionally, been a bad thing. As an organisation grows, so does its bureaucracy. Government has, then, repeatedly broken itself down (Departments and Ministries ... agencies and NDPBs) in an effort to separate policy from delivery and get closer to the customer, with varying degrees of success.
Political fiefdoms have, at the same time, been created to satisfy egos (ODPM) or to pretend to the outside world that real change was happening (the story of dti on its journey to the current BEIS for instance). Alongside that, functions have moved - Child Benefit between DWP and IR (now HMRC) - and Tax Credits, whilst benefits, were sited in HMRC rather than DWP, to the great consternation of HMRC staff on day one (and for many days thereafter).
GDS, perhaps accidentally, perhaps as a result of a flood of cash in the Spending Review, has become that big, slow moving central organisation. I'm sure it wasn't intentional - they saw gaps all around them and took on more people to fill those gaps. Before they knew it, they needed a bigger office to fit in the 900+ people in the organisation. Along the way, they forgot what they were there for, as the NAO said.
On data, all we know for now is:
"Data policy and governance functions of the Government Digital Service (GDS) will transfer from the Cabinet Office to the Department for Digital, Culture, Media and Sport (DCMS). The transfer includes responsibility for data sharing (including coordination of Part 5 of the Digital Economy Act 2017), data ethics, open data and data governance."
The real issue here is not that "Data", whatever that is in this context, has moved from GDS to DCMS, but that we lack (still) an executable strategy. We have a trite "transformation strategy" that is long on words and short on actions (see "No Vision, No Ambition" on this blog), but we have no real framework to evaluate this decision, to move "Data", from one department to another.
An executable strategy would lay out not just the what, but the why, the how and the when. We would be able to see how changes were planned to unfold, whether incremental, revolutionary or transformational ... and when a decision such as this was taken, understand the impact on the that strategy and whether it was good or bad (and sometimes, decisions with known bad impacts are taken for good reasons).
Mike Bracken, writing in the New Statesman, is emphatic that this is a bad idea - one that runs against what everyone else in the world is doing. His closing take is that:
"the UK seems to have made government a little bit slower, more siloed, harder to reform and more complex."GDS is hardly the rapidly responding, iterative, agile organisation that it set out to be (and that it certainly was in its early days as I've said before) ... so maybe this little bit of disaggregation will free up the remaining (and still large) part to get moving again.
Over the last two decades we've had several goes at this - OeE, eGU, OCIO and then GDS. Each worked for a while and then got bogged down in themselves. New leadership came in, threw out some of what was done, took on some different things and did the things that new leaders generally do when they come in (say how rubbish everything was until they came along and then proceed to do much the same as had been done before only a little differently).
I suspect, though, that this isn't enough of a change. We need a more fundamental reform of GDS, taking it back to its roots and to what its good at. So maybe it is the beginning of the end and maybe that's no bad thing.