I used to lead the Open Data Institute’s work on data institutions. The team both piloted data trusts and explained that a range of approaches existed – including things like data representatives and data cooperatives – that can change how decisions are made about data. Hopefully to make those decisions more trustworthy. There are many other people working on data institutions in the UK, in Europe and around the world. I’m often surprised by how many.

Over the last couple of weeks I have been talking with people about data institutions. Many of the conversations surface similar implicit assumptions.

There can be only one

In many of the conversations people assumed that there could be only one data institution within a particular context. They had not thought about whether and when there might be multiple.

Some data institutions will exist to steward data for which you might want there to only be a single source of truth[mfn]I know. I do love a bit of epistemology and discussions about the nature of ‘truth’ but that would be an unnecessary diversion in this blogpost[/mfn] – for example the list of Prime Ministers of a country, the list of websites that exist, or who you are married to.

Many others will steward data or have a purpose where there might be multiple things doing roughly similar jobs but, perhaps, with different methodologies or priorities. Maybe one has a purpose of “for the benefit of the people of Newcastle”, another has “for the economic benefit of the people of Newcastle” and a third has “for the benefit of the businesses of Newcastle”. A single word can make a big difference.

Sometimes there should be only one data institution but multiple will exist. That’s life. We live in a wonderfully imperfect world.

Being open to the need to work with other people and other institutions is a better starting assumption than there being only one. Institutions might compete with each other, cooperate with each other, or both, but do expect it to happen.

Rip it up and start again

Another assumption was about the need for something new.

The way we steward data at the moment is not working, therefore we must need a new institution to fix the problem, right? Maybe…

Sometimes we need to fix things that are not working, or at least try to make them better. An existing institution might provide vital services, it might contain valuable knowledge, or it might do things that – shock! horror! – are only loosely related to data. Creating a new institution might break existing and important things.

I do not know of a good methodology to help people decide when to try a revolution and when to try evolution, but do make sure that it is a conscious decision

You forgot government

Many people thought that they needed a new type of data institution – like a data trust or data cooperative – when actually they might just need to improve a simple, old-school democratic institution like a bit of government.

I am very conscious that I live in the UK, a high-income country with an old and (relatively…) stable democracy. Not everyone does. I’ve worked a lot internationally, but mostly in similar countries. In these countries we have many institutions that are already legally responsible and democratically accountable for stewarding data for a particular purpose.

There will be institutions responsible for land registries, local places, criminal justice systems, welfare payments and – in a country with a national health system like the UK – health and social care. Perhaps, rather than working around those government institutions you need to use democratic processes to change their behaviour to make them more useful and trustworthy. 

Some people seemed to forget the government and implicitly assumed that they needed to take responsibility into a new institution that they would build and run.

Sometimes we do need to take responsibility away from the government, but at other times we need to add new responsibilities to government or just make existing bits of government work a bit better.

Again, make it a conscious decision.

Building institutions takes time

Building institutions takes time. Not just your time, but other people’s too. It will take even longer if you do not think about why you are doing it and do not surface and challenge assumptions about what any new institutional arrangements should look like.

Assumptions like whether there will be multiple institutions, whether there should be something new, whether the institution should be part of the government, what approach you need, or even whether that approach is suitable for your particular context.

Making those assumptions explicit and challenging them is likely to help you move a bit faster and be a bit more effective at actually making people’s lives a bit better.