It has been an exhausting week, with three separate trips into London and more face-to-face time with colleagues and friends than I’ve experienced in a while! But also an exhilarating one…
Strategy day
I’m going to start with our Strategy day on Wednesday because that really was the centerpiece of the week.
Jonathan, our new Advocacy and Campaign Director started work on Monday, so Wednesday was a great opportunity for us all to get to know each other, align somewhat around our goals and language, and think through what our goals should be over the following month.
Getting to know you
I’m a fan of the book The Five Dysfunctions of a Team. It argues that the foundation of a high functioning team is trust, because when you trust other members of your team (at a deep, personal, level) then you can disagree with them without fear, and that enables you all to focus together on your organisational goals rather than holding back to protect yourself.
Tim and I have known each other for a while, but never actually worked together, and Jonathan was someone we had only had a few conversations with during the interview process. My main goal for the strategy day was to build trust quickly.
The Five Dysfunctions of a Team also argues that trust is created through being vulnerable with each other. So before the strategy day, I set us all the task of creating a personal user manual, containing:
- A bit of personal information eg about history and family
- A description of our values and what we value in others
- A description of both our strengths and our blind spots where we need support
- A summary of what motivates and demotivates us in our work
- Our preferred working patterns and styles
- Our preferences about how we like to be communicated with
- Our pet peeves
I’d intended for us to talk through these on the day, but we all read each others’ before the day itself, and instead used the time chatting over lunch! So I’ve now set us the task of each adding comments asynchronously to each others’ personal user manuals to highlight areas of overlap or complementarity.
One of the things that’s most liberating for me about starting this initiative is being able to be pretty unapologetic about expressing myself through the organisation. Unlike at ODI, I don’t have to be concerned about the expectations or sensitivities of the founders or funders or a team I didn’t select. But while this is freeing, it comes with its own challenges: there is no distance between our goals and values and my own, so as I talk about them I am talking about myself. This is naturally a more vulnerable position. But I’m trying to just lean into that as an advantage!
Practical culture
A second exercise I led was about how we all wanted the organisation to work. I got us to silently brainstorm practices we wanted to have in place around communication, meetings, documentation and other day-to-day ways in which the organisation would operate.
We’ve agreed on a rhythm of touch points:
- Daily stand up (on plans for the day) and stand down (on what actually happened) messages in Discord
- Weekly video calls on Monday mornings (on plans for the week) and Wednesday afternoons (to check in and for discussion)
- Weeknotes like these for personal reflections in public; we discussed a bit how these aren’t official organisational communications, but a more exploratory space
- Monthly in person meeting days to plan for the month and in-person brainstorming
We discussed having a quarterly rhythm as well, but haven’t quite resolved what they should look like. I’ve seen other organisations have reading weeks, co-working sprints, and finishing-things-off fortnights. We might play those a little by ear and see what we need at the end of the summer, and the end of the year.
One thing that struck me as we talked was that we were all keen to build a high trust environment. We want it to be ok to talk about taking a walk in the middle of the day; to take all your holiday (or more); to publish something without getting it signed off, because you’re confident that’s not needed. No doubt there will be bumps along the way but it was lovely that those values of trust, care and collaboration came through the discussion from all three of us.
Yeah but what about the strategy?!?
Ok, I’ve focused a lot on the internals of how we work rather than the majority of the actual content of the day, which was much more focused on why we’re here and what we’re doing. So let me just run through some reflections from the other exercises we went through…
How does collective data governance help?
First, we spent some time going over a draft theory of change I’d put together to try to express the rationale behind Connected by Data and the activities I’d pinpointed for us to do.
This theory of change used the Nesta template, which starts with problems and ends with impact. Jonathan rightly questioned the connection between them: for example how would unfair personalised pricing be addressed through collective data governance? The logic/rationale is gradually clustering in my mind into three buckets:
- Many of the problems we’re seeing with data manifest at a collective level – for example, we can only see where personalised pricing is unfair by comparing the prices encountered by different communities – so these can only be addressed collectively. In particular, the market’s invisible hand rarely favours the marginalised without some strong inducement.
- People don’t make wise or informed decisions about data, for a whole bunch of reasons to do with cognitive limits and biases, being nudged through dark patterns, and the inherent complexity of the modern data ecosystem. Collective decision making bodies should have a perspective and processes that helps them to be wiser and more informed, and create a safe environment for individuals to operate within.
- Some of the decisions we have to make about data-driven systems, such as what “fair” pricing actually means, are not straightforward, and depend on fuzzy, debatable concepts. Those decisions – such as what counts as public good – should be made by the people affected by them rather than by the organisations whose interest it serves.
Who does it help?
Jonathan also pushed us to be more specific about who would actually be helped by the changes we want to see. This involved us writing short pen portraits onto post-its: the smallholder farmer who is offered crop prices based on satellite imagery of their fields; the immigrant who avoids using the health service because they’re worried about how data about them might be used; the researcher who painstakingly spends time opting out of cookies as they search the web.
There were two things that I think it was difficult to capture through this exercise. One was the everyday experience that affects all of us: it was a lot easier to come up with stories about people who are already marginalised, but we don’t want these issues to feel like “someone else’s problem”. The second was that these stories didn’t easily capture the less tangible impact of a lack of sovereignty or agency, and the feeling of exploitation and disempowerment that comes with it. Instead, they focused on concrete outcomes, such as unfair pricing. It feels like the more abstract, emotional, response is one that would be resonant but perhaps harder to communicate in ways that prompt action.
Who has the power to change things?
I had a number of conversations with people over the past week – with Tim and Jonathan, obviously, but also with Renate Samson, Tom Steinberg, Sushant Kumar, and others – about Connected by Data being a (time-limited) campaign rather than a long-running organisation.
I find this framing really energising, but I’ve also realised that we need to get more clarity about what the longer term (5 year) success in that campaign looks like. I’ve worked through things like the roadmap for this year, but in part because we only have funding so far for this year’s work, I haven’t been more specific than that.
With that in mind, the discussion Jonathan led on “Who has power to change things?” was simultaneously helpful and daunting. This was a tough question to answer, even from a parochial perspective of the UK. On one hand, I do think that organisations that are stewarding or using data could already do quite a lot to incorporate collective data governance if they wanted to, so I was inclined to say that Chief Data Officers had that power. On the other hand, it could certainly be accelerated if it were encouraged by the ICO, and even more so if it were enforced by them.
But the power to shape ICO’s power comes from DCMS. And our legal framework is (and will continue to be) heavily influenced by the European Union, and international policy-shaping bodies such as OECD. And in practical terms the effectiveness of any changes is going to be dependent on the big international tech players such as Google, Meta and Apple, which really need internationally aligned regulation, which needs political impetus at the level of the G20 and UN.
We need to do more work on this, but my current feeling is that there is no one locus of power, and that’s part of why the problems with the data ecosystem are so difficult to unpick. Getting clarity about what we choose to target and how we pick our battles is going to be important. Which is why one of my big tasks coming out of the strategy day is doing a bit more thinking about picking domains to focus on.
Electoral data
On Tuesday, I went to a workshop organised by the UK Democracy Fund within the Joseph Rowntree Reform Trust, looking at electoral data. The main point of the workshop was to look at some emerging recommendations about what would be needed to build stronger data infrastructure for each stage of the electoral process – from data about electoral infrastructure such as places we can vote, through data about candidates, the election process itself, voter registration, and actual votes.
The whole discussion highlighted to me what a poorly governed state a lot of this data is in. Democracy Club is doing a superb job with data that really should be better stewarded and shared by public bodies. At the same time, electoral registration and data about who votes (the marked register) are made available in ways that many voters will be unaware of – and possibly not supportive of – while being unavailable for researchers or civil society actors who are trying to increase voter turnout.
Perhaps unsurprisingly, none of the recommendations focused on asking the people affected by this data collection, use and sharing, what they thought. So that was what I focused on. As with justice data, there are some really interesting challenges here in balancing individual privacy (people don’t generally like being profiled for marketing) and public good (political parties should be able to campaign, which is essentially “marketing” themselves to voters). Making these decisions will require nuance and they will have a lot more legitimacy if they are decided by those affected rather than technocratically.
17 Rooms
The final event of my week was a “17 Rooms” session on building a gender data cooperative. This was particularly interesting to me as it had a focus on building on existing women farmer cooperatives in India: not creating new data institutions, but building data into institutions that already exist.
As with many such discussions I still felt the purpose of such a gender data cooperative was a bit unclear. We discussed benefits at different kinds of levels, which I think are distinct and might favour different designs, but two advantages of pooling data felt interesting to me:
- enabling individual farmers to benchmark themselves and learn new farming approaches from other members of the cooperative
- enabling the cooperative as a whole to present an aggregate view of their members to the outside world, for example to get loans, insurance or other financial products
This latter relates to privacy-enhancing technologies that provide for local processing of data without the underlying data being shared or copied. There is a technological push towards federated data processing (including federated machine learning) but for this to work we also need federated data governance. And this ties back to the issues I discussed last week about how and whether different communities might adopt different governance structures, but perhaps with the same governance “interface”.
Anyway, this last week was a lot. I’m grateful now for a little time off, around my birthday and taking advantage of two bank holidays in the UK for the Queen’s Jubilee. So no weeknotes next week.