Disruptive Proactivity

“The public generally do not understand the difference between operational and statistical uses of personal data” say ONS. They are right. Unfortunately, neither do other people who probably should.

The ODI has recently published a draft data spectrum which recognises that there are a variety of areas of tradeoff and differences.

The Chief Data Officer is not about statistical uses of data – it is serving very different user needs (user needs, not departmental needs) – it is/was about using data to improve operations. Why do you think Whitehall was so terrified?

ONS is a creature of statute – the culture, processes, and worldview are all on that basis. ONS produces statistics “which serve the public good”, and is brilliant at creating and publishing Official and National Statistics on the basis of data it gets given.

An effective CDO will force departments to produce/use/release data on areas they don’t want to release. While that is in theory doable via Official Statistics, in practice, it rarely is, absent outside drivers, and ONS will never have the political leverage to force publication.

Official Statistics are required to have a detailed process – it’s what makes the construction statistics better than counting cranes. The National Statistician’s remit is clear, the remit is defined, and clearly advocated for. It would be unwise to throw that away. Having ONS “do” operations is probably strategically unwise. What the (open) data communities have long asked for, is something beyond that.

For example, there have been questions raised about the “user need” for the GDS performance metrics. We know that every department hated those metrics being public, because they’d rather hide them (and, at first, they did). The CDO should be about data coming from operational processes, to improve those processes. That is not Statistics or statistical; but it is the basis for agile process development – to know what works, you have to test and measure. The CDO, being integrated with GDS, had the ability to use data to drive transformation that wasn’t purely digital. And it is from that, it is clear why the CDO role is was so controversial, and is now being gutted by those who quite like the status quo.

ONS is about analysis and fact; the CDO should be about operational detail and improvement (what that means in practice is an entirely different conversation – there are of course relationships between the two). The National Statistician is entirely politically independent; whereas the CDO can respond to such interests.

Experienced staff who should have known the difference between operational and statistical uses, should have understand both open data and GDS principles enough to see the difference, yet may not have entirely recognised the effects of that difference. ONS may want the public to understand the difference (p20), perhaps they could start with their own Directors…

We have had the ineffective and deservedly discredited “data sharing open policy making process” that Paul Maltby crashed onto the rocks of Whitehall intransigence and pre-existing hidden agendas. The new CDO will require the operational power to tell DWP that no, you can’t secretly take data from across Whitehall while refusing to reciprocate, and nor can you demand the ability to buy more services from Experian without providing any evidence that the existing projects to do the same thing have worked at all (they refused), and it’s probably no longer ok for Government to randomly use data without telling people what you’re doing and why. The CDO role needs someone who understands data, but also someone who can deal with the political crap, and win (sometimes).

It’s no wonder Whitehall was so glad to see the back of Mike. They don’t care about that new fangled web thing; but working out transaction costs? the civil service wanted no more of that, thank you very much.

Money and territory, as always.

Someone asked on twitter yesterday who should be the next CDO. And while none of the people suggested either want or should in any way get the job (anyone want care.data imitators all across government?), that’s not to say that it’s widely understood the role they play. Open Data is a part, but likely a relatively small part, since Open Data is already expertly handled by other organisations with their own remits. Join up, don’t replicate. Not every group in this space has delivered or shown a user need for its remit.

Given the need for political capital and an understanding of resources, neither of which are likely to come from the Cabinet office where the role has to live, a non-civil servant might be a very good idea this time round.