I wrote some month notes in January. It was satisfying putting them together, and some people seemed to appreciate them, so here are some for February.


Strategic work

Programme planning

As I wrote last month, we’re currently putting together plans for five programmes of work that we will structure our work around over the next few years, on data literacy, data assurance, data for challenges*, data institutions, and evidence & foresight.

* We’re not sure about the name of this programme, because it sounds like it’s focused purely on putting the right data together for innovation challenge prizes, where we mean it to encompass more extensive data initiatives or programmes of work that might also cover things like creating common standards or policies, and not include challenge prize activities. Naming things is hard.

I’m really pleased with how things are progressing with most of those plans, and in particular how the programme leads are taking (and feeling) ownership of them. In case it’s helpful (I struggled for a long time around how to describe programmes of work), here’s the structure of the programme summaries I put together to help guide the programme leads:

  • One paragraph summary
  • Background: the problem statement for the programme, laying out the challenge in the world the programme seeks to address
  • Goals: description of the impact and outcomes of the programme; we’re trying to split these into external outcomes (that advance our mission) and internal outcomes (that advance our sustainability)
  • Ecosystem: description of the other organisations in this space, and the stakeholders for this programme, including ODI’s USP relative to other organisations
  • Activities: description of the (long running) workstreams within the programme; it’s proving useful to include a standard “running the programme” workstream here which can cover activities like programme boards and upskilling the team
  • Structure: description of who does what within the team and programme board
  • 2021 Plan: description of the programme priorities, KPIs, milestones, timeline and budget for 2021; for example if a programme has a research workstream this is where we describe what we’ll be researching this year and when

We’re also putting together questions-options-and-recommendations papers for the Board, so that when we discuss the programmes with them in March we have some thought through, meaty and meaningful discussion topics rather than just presenting the plans. These are focusing on questions of scope, focus, scale, and ambition, which I think are the right kind of questions to be taking to the Board.

The piece that I’m most worried about is KPIs. It’s so easy to come up with KPIs that are easy to measure but meaningless, or vice versa, or that push you in the wrong direction (eg sacrificing quality and impact to meet an unrealistic quantity target). I’d really like someone external to review, challenge and shape the KPIs that we’re putting together, so if you or someone you know would be up for that (can be funded, but is under time pressure), please let me know.

Talking about our commercial work

A second strategic area that I’ve been pushing at ODI is about why we do commercial work (that is, transactional work that generates margin for us, mostly for private and public sector clients) and how we talk about it.

From what I have seen, non-profits fall into three camps:

  1. Those that don’t do commercial work at all. These organisations rely on grants from philanthropic investors, and sometimes government. If there is transactional project work it tends to be limited to work for NGOs, charities, and other “good guys”. They tend to be (and have space to be) “outsider” organisations that hold corporates and governments to account.

  2. Those that do commercial work to generate profit to support their mission work. These organisations typically have a subsidiary that’s able to operate semi-independently and (if successful!) creates a sustainable revenue stream to support work in the main non-profit/charity that looks more like #1 above. Having an independent subsidiary provides some level of reputational protection when the commercial arm does work that the non-profit wouldn’t view as contributing to the mission.

  3. Those that advance their mission through commercial work. These are “insider” organisations who view the commercial work they do as advancing their mission, because their mission includes working with and changing (even a little bit) the behaviour of private and public sector organisations. Being paid helps them be taken seriously, avoids subsidising resource-rich organisations, and increases the diversity of their revenue. But it also creates questions about whether they’re taking on a piece of work because it advances their mission, or just because they need the money.

This third option is the one we’ve generally pursued at ODI, but we also sometimes talk about our commercial work as if we’re doing #2: subsidising (separate) mission work. Both models are absolutely fine, ethical, models that enable organisation to satisfy their mission, but oscillating between the two explanations can be confusing for the people who fund us, those we work with, and the team. So over the next few months we’re going to try to have this conversation explicitly and review our content and reporting so that the implicit messages we send about what’s important to us reflects our explicit position.

This is a difficult conversation to have because it gets muddled up with the ethics of working for or accepting money from particular organisations or types of organisation; with the need to sustain the organisation and the responsibility to our team; with concerns about funder influence on organisational shape and strategy; and even personal impact on people who feel the work they do is under-appreciated or even disapproved of by others in the organisation. So I think talking about it is likely to surface other tensions, but is really necessary to give clarity and consistency.

Other internal work

  • We selected a great new Head of Consultancy who will start in May. Yay!

  • We’ve started identifying topics and keynotes for this year’s ODI Summit. Let me know if you have anyone or anything on your wishlist.

  • Our 2020 Annual Report is out!

I wanted to share something else here. Normally, we have whole-team 15 minute standups every Monday, Wednesday and Friday. On Mondays we share our main goal for the week; on Wednesdays, give a shout out to people in the team; and on Fridays share a success or surprise from the week. This last week we mixed it up a bit and had an LGBTQ+ shout-out stand-up on Wednesday (we had a one for Black people we admire in November last year; unfortunately I missed it). It was really lovely hearing so much admiration and appreciation from the team for LGBTQ+ people in their personal lives or in the public eye. I was left with a real sense of warmth of common humanity. Highly recommended as an exercise if you want to try it in your team.

Project work

This month I have contributed to:

  • Our public policy work – which I’ve been helping Milly scope and shape – is ramping up now. There are lots of opportunities for the UK to be internationally influential this year, with the G7, D10 and COP26, so we’re working out how to make the most of them.
  • I helped us find a good partner to work with to help us build a register of data institutions, using Wikipedia as an initial source of data.
  • Reviewing work by the team, including:
    • our response to the Ofcom consultation, in particular advocating reviewing the cost and governance of the Postcode Address File, as advocated for by Anna at the Centre for Public Data
    • our response to the UK government’s public procurement consultation, where we’re emphasising how requirements to open and share data needs to be embedded into ITTs and contracts, as well as the stuff you’d expect about the need for transparency around procurement
    • chapters towards a report we’re doing for the World Bank on data institutions in low and middle income countries
    • a paper we’d advised on around developing a city-level data ecosystem
    • a blog on how we think about and go about creating data strategies
    • a guidebook we’re creating on assessing, building and demonstrating trust and trustworthiness
    • a playbook we’re creating on designing initiatives to improve data infrastructure and data access
    • a workbook we’re creating on developing business models in data institutions
    • thinking around how OpenActive might develop into a data institution

Interesting conversations

This month I’ve chatted with:

  • a guy from Sweden about how to increase their publication and use of open data
  • people in GDS about government’s (open) data plans
  • people in BEIS about evidence around the value of data
  • Lloyd’s Register Foundation about role of data in supporting evidence-based decision making
  • various people about the future of CDEI (I was really surprised to see the advert for a new Executive Director given that this still seems to be under review)
  • NHS England/Improvement about publishing open data about vaccinations
  • Microsoft about how business feels about the EU Data Governance Act proposals
  • the McGovern Foundation about our continuing work around data institutions
  • Rudi Borrmann at OGP, about their work on OGP Local and links with our own work around open cities
  • Consumers International, in particular exploring how consumer advocacy organisations might take on more of a role of data institutions, for example by stewarding data about consumer experience with products
  • Christina Colclough about labour unions as data institutions
  • Simon Manby and Lisa Moretti who are working with the Office of the Public Guardian to modernise lasting powers of attorney – this was a fascinating conversation, bringing into sharp focus difficult questions about digital identity and the role of public bodies as data institutions – do take a look at Simon and Lisa’s work here, it’s excellent

I was interviewed four times, by researchers looking at:

  • open analytics and data for research in the health sector
  • health data sharing arrangements
  • ethical and equity issues around the Ofqual qualifications algorithm, in particular exploring how it affected students outside the UK and what the future might hold for the use of data to inform grades (I really enjoyed and learned a lot during this one!)
  • the UK’s AI research and innovation strategy

And I saw presentations by:

  • Roza Vasileva, who is wrapping up her PhD looking at smart cities and participation, particularly in Tanzania and Kenya (ODI has been a kind of host organisation for Roza during her PhD)
  • the University of West England, whom we commissioned to evaluate our R&D programme and had some good recommendations about how we can improve our knowledge and process management, and the evaluation of our work programmes


There’s not much additional to report here because over the last month we’ve really just been concentrating on getting the concept notes that we’ve been working on (and I wrote about last time) into a state where they can be shared for input from the GPAI Steering Group and other organisations.

As a reminder, the topics we have been working on within the Data Governance Working Group are:

  • data justice
  • data trusts (etc)
  • balancing innovation and data protection in legal regimes
  • handling co-creation rights
  • international rules on text and data mining
  • dataset documentation and management
  • privacy enhancing technologies

If you have an interest in being connected with the people leading the shaping of our work in any of these areas, please let me know and I’ll connect you to them. We want to narrow this set down over the next month or so, and external interest and overlaps are factors in working out how to do that.

Other work

I attended a couple of oversight boards, namely:

  • OpenSafely: This was the first meeting so a lot of the time was spent going over how OpenSafely works. I did learn, though, about an elegant expectation mechanism that they’ve put into place. This enables people defining studies to specify their expectations about the distribution of values for different variables, and serves two purposes. First, it is used to generate warnings when those expectations aren’t met, which might indicate something is going wrong with the analytics code (so helps with testing and debugging). Second, it’s used to generate dummy data that can be used to test the analytics code. 

  • Data Ethics of Location Data: The first meeting of a project being run by the Geospatial Commission, with the support of Sciencewise, who have commissioned a public consultation exercise around the ethics of collecting and using location data. There are still some questions about scope (there are lots of questions to answer, and while “location data” really means all geospatial data, most people seemed most interested in talking about the collection and use of data about people’s locations), but I was really impressed with the design of the project overall and am looking forward to seeing how it unfolds over the next 8 months or so.

I don’t pay much attention to being on lists, but I was one of the DataIQ 100; it was kinda fun virtually seeing friends at the award ceremony.

I spoke at a Westminster eForum conference on geospatial data. I spoke about the unique characteristics of data, the importance of data institutions and how essential it is to ensure their business models incentivise them to steward data on behalf of wider society rather than their customers (this plainly not being the position that Ordnance Survey has been in for many years). I called OpenStreetMap the best geospatial data institution, and a great example of global Britain, because it really is, and it so frustrates me that government data policy effectively acts as if it doesn’t exist.

I also attended SocSciFooCamp for the first time. It was a bit overwhelming and intimidating, and in many ways I was glad that it was entirely virtual (even if the PST timezone meant three intense evenings). The organisers created a Discord server and there was a lot of chat both pre-event and during the event, which was nice in terms of building a sense of community. I gave a lightning talk on how we should be focusing more on helping existing institutions take on the role of being data institutions, rather than solely new data institutions (which was well received). I also ran a session looking at how experimentation with new forms of data institution might go wrong, which I wasn’t so happy with (but I know some people did enjoy). I enjoyed listening to Cory Doctorow talk about commercial compatibility (about which more below); Nancy Potok and Amy O’Hara discuss plans for the new US National Secure Data Service; and a session that talked about care and the experience of caregivers.

I also attended (but didn’t speak at) an “AI Dialogue” which focused on risk-based regulation around AI. A lot of the conversation made me wonder why / how there are specific issues about risk assessment or regulation around AI as opposed to other technologies. Is it because it’s a general purpose technology and therefore has potentially wide-ranging effects? Is it because it’s developing faster than other technologies have? Is it because the potential impact is particularly harmful? The big questions centred around how to work out which areas of AI development were the ones that were really problematic. One observation I found particularly interesting was a discussion of the US’s comfort, with (and active pursuit of) ex post regulation, ie waiting until something goes wrong then litigating the hell out of it, compared to Europe’s preference for ex ante regulation.

Thoughts I had

  • For the talk about geospatial data, I wanted to explain why restrictive licensing around onward use is so problematic. We’ve previously used the phrase “poisoning the well” or “digital cholera” for this, but I wanted to incorporate the fact that we get most value from data when it is combined with other data. So I used the analogy of a tree, with data being joined together, and the results of that combination being used with yet more data, in a branching structure like the roots of a tree, growing and combining to create a wonderful canopy. A licence that doesn’t restricts onward combination of data means you don’t get a strong root system, and the tree is stunted. A licence that virally propagates poisons the whole tree, and if you have many of them from different sources they are like poisons that interact and kill the tree.

  • Another analogy that’s been fun to explore with Jack this past month, is around the different models the ODI could use for our deeper interactions with some data institutions. We think there are advantages (for us and for them) to actually running some data institutions, but the question is how (and why). We could run a lab, like the one that created Dolly the sheep, focusing on experimenting with different data institution models but not actually trying to create anything long term viable. We could run a zoo, where we keep the data institutions long term as examples to show what’s possible. We could run a breeding programme, where the focus is on getting data institutions into a state where they can be released into the wild. Or we could run a sanctuary, where we help ailing data institutions to die gracefully and respectfully (influenced by Cassie Robinson’s work on Stewarding Loss here). Or some mix of all four. We’re talking this choice to the Board during March; I’m interested in any opinions.

  • I saw a bunch of comments on Twitter objecting to the term “unintended consequences” when it comes to the use of data and tech. The words we use matter and it’s useful to examine them. I understand the objection to “unintended consequences” when it’s used to describe consequences that arise from people, purposefully or through lack of thought, ignoring adverse consequences of tech and the people who point them out. But I do think that there are sometimes consequences to new technologies that are unpredictable, simply because we live in a hugely complex and interdependent world, where even with the best consideration, will and preventative action in the world, shit happens. Unknown unknowns. I’m going to try using “ignored consequences” and “unpredictable consequences” for these two types of outcomes in future.

  • At SocSciFooCamp, Cory Doctorow talked about a really interesting paper from EFF advocating removing legal barriers to apps scraping websites in order to eg port data to another service or provide services layered on existing ones, to drive more official mechanisms for data portability and interoperability. The paper advocates turning legal battles into technological ones and characterises (some) privacy concerns about things like these initiatives enabling fraud or revealing data about third parties as chaff thrown by big tech to maintain their monopoly strangleholds. It’s an interesting paper, with a lot I agree with and some that concerns me, namely (a) that while it may improve privacy over the long term, there could be a lot of damage along the way; (b) that technological battles just seem like a lot of wasted effort (albeit fun for hackers); and (c) that placing existing monopolies into the position of being fundamental infrastructure for other services will only embed them and their power. I do need to write this up properly.

  • I’ve been thinking a bit about what I would do in some of the roles that have been advertised recently, or are likely to be advertised shortly, including the Executive Director of CDEI, the Innovation Commissioner, and a putative Chief Data Officer within the Central Digital and Data Office. I find this “if I had that job” exercise useful because it forces a level of practicality and understanding about the enormity of the challenges, compared to just thinking about what I’d like to see someone else do in the role. This recently published introduction to levers for digital service groups has some ideas but seems premised on the notion that the early GDS approach was a success (which in many ways it was, of course, but also wasn’t in others). These are all roles that will need a combination of vision, systems thinking, and canny use of the limited set of levers that they will actually have at their disposal.


Work life

  • Email: I’m really struggling with email (as usual). I get time to triage, which is good, and I’ve introduced urgent and diary labels to help give more nuanced to that. I am able to keep on top of those that require urgent action and that require diary appointments. But things that require more thought I’m just labelling as “action” and they’re frankly neglected. I book time for actioning them but it gets overbooked or used for other things or simply isn’t long enough. Sometimes I know I avoid them because they require thought or I simply don’t have a strong idea about how to answer them. I need to be stricter about getting through them, or recognising when I’m choosing not to reply.

  • Booking time: I’ve managed to get (most of!) the ODI team to book time into my calendar when they want me to do things, in particular things like reviewing documents. It helps keep the timing of those realistic, and it ensures that my days are usually balanced between meetings and heads down time. But there’s often a few meetings that I’d really like to take earlier than my diary permits. I’m thinking about holding “last minute meeting” time in my calendar each week explicitly to enable that to happen.

Home life

I had two days off during half term, the highlight of which was watching Hamilton with my kids. I had tickets to go and see it live last year, for my birthday in May, but of course all the performances were cancelled. Perhaps I’ll appreciate the live performance all the more for watching it (on Disney+). King George’s Songs have been echoing in my head ever since; they’re such an earworm.

My eldest, who is studying Government and Politics A level, was given some extra work to do, namely a course from Hong Kong University called Europe without borders? We decided to do it together, to make it more fun, and it was! Really well put together, explaining the EU’s governance structures and evolution over time, its internal tensions and some of what the future might hold.

Video games:

  • Drake Hollow – nice exploration/base building game; a little too much combat for me
  • Slay the Spire – I’m watching my eldest play and occasionally throwing in suggestions about what would have been a good thing to do, just after she’s taken her action
  • Spiritfarer – just picked this up again; the kids were playing it but my youngest got tired of it so I’m playing in their stead
  • Sudoku – more specifically killer sudoku, where you are told the total value within groups of contiguous cells – has been an obsession for me over the past year. I was doing the two in the Guardian each day (not using annotations on the easy one to give some extra challenge). Then at Xmas my parents recommended a great killer sudoku app, with some really challenging puzzles that would take me hours to complete. That same company makes a number of other sudoku variant apps, and I’ve worked through sandwich sudoku, chess sudoku and have almost completed thermo sudoku. Earlier this month I thought I was spending too much time on it and could do something more productive with my time, but when I stopped I realised that the time I was spending doing these puzzles was not time I could spend more productively or creatively. It’s time when my brain needs a break (or a different form of stimulation).

Other games I’ve been playing:

  • Terraforming Mars – I’ve been playing the solo player variant jointly with my eldest (as opposed to competing with her). You can never be confident of winning when playing solo and if it comes together at all it’s always in the very last round.
  • Quacks of Quedlinburg – a rare competitive game for us but its gameplay is such that you feel you’re only competing with yourself and lady luck. Highly recommended.
  • Horrified – hadn’t played this for a while, but it’s a fun, fast paced, thematically coherent and challenging co-op game.

We’re still playing Masks. I’ve also been reading and fantasising about playing Good Society (especially after watching Bridgerton), as I love the idea of role playing that doesn’t centre on combat, but on character relationships and desires. If it’s something you’d also be interested in let me know and maybe we’ll be able to find enough people to play…

TV we’ve been watching:

  • Bridgerton – binged on this with my eldest; I particularly enjoyed that it didn’t stop at marriage
  • The Queen’s Gambit – so good, even if I can’t shake the feeling that Thomas Brodie-Sangster is 15 playing dress-up (he’s 30)
  • It’s a Sin – wonderful, heartbreaking
  • The Expanse – Alex! No!
  • WandaVision – a bit disappointed with the episode 7 reveal, but such a good show
  • Star Trek: Lower Decks – cartoon set in the Star Trek universe! Excellent
  • Law & Order – part way through series 3 and we’ve got one female returning character (though she’s not a lead). Every episode, DA Schiff gives a “you’ll never win this, Ben” speech and my youngest and I crack up at the predictability.

A few films, mostly with my wider family:

  • Roma – honestly people found it tedious and unpleasant so we stopped part way through
  • Set It Up – went for a romcom for Valentine’s day; it went down ok here but the wider family didn’t enjoy it
  • Plunkett and Macleane – remember really enjoying this when it came out; the modern/historical mix seemed really innovative then, but less so now. Still adore Alan Cumming and Craig Armstrong’s music though

Mental health

There’s a lot of noise at the moment about “getting back to normal” and I’ve been thinking about this in particular in relation to my youngest child, and myself. There are some people who are thrilled with the idea of a return to normal office/school life, and obviously I really get that, particularly for those with difficult home situations, those living alone, those who don’t have necessary equipment, those who aren’t able to work during lockdown, those who have had to look after small children, and those who just love being around people and learn and get energy from those interactions.

But there are also some of us for whom lockdown has not been all bad, and might even, on balance, have been better. I love being able to structure my day so I can get on with work first thing, because that’s when my energy is high. I love having proper heads down time where I can manage interruptions and distractions in ways that are just impossible in an open plan office. I love not having two hours of travel each day, pressed up against other commuters. I love not wearing shoes all day. I love that calls that span continents feel the same as calls with people just down the road, and not having weird hybrid meetings that exclude those on the computer screen. I love having my own space. I love being able to go for a long walk each day, especially now it’s sunnier. I love parallel text chats in video conferences. I love seeing my family during the day and not being away from them for days on end due to work trips. I love not flying.

That’s not to say there aren’t things I miss from the before times too. I miss hugging my friends and colleagues, probably most of all. I miss catching up with John most Friday evenings, a lot. When I’m speaking at conferences, I miss seeing people’s reactions, hearing their laughter and their applause, and I miss being grabbed for a question or chat, and bumping into friends for a gossip.

I do recognise my privilege in being able to see benefits in how I work now, and that it’s very different for people at different life and work stages. I guess I’d just like us not to fall into the easy assumption that everyone is over the moon about the idea of a “return to normal”. For some of us it’s actually a cause of anxiety.

That goes particularly for my youngest, who has found schooling from home much more suitable to their needs than the school environment. I wrote last month about their autism. They seem to learn better, been more engaged and enjoy learning more when they are able to manage their time, follow their passions, and are able to doodle and stim during lessons. Being home has given them space for that. I feel really quite sad for children like my youngest who have been shown that there is another way to learn (just as we adults have been shown there is another way to work), only to have that taken away from them again. They will be moving schools anyway after GCSEs, so we’ll be looking for sixth form colleges that can accommodate more remote learning, if such a thing exists.