Following on from my February month notes, here are some for March.
InnovateUK R&D programme
March is always a horribly busy month at ODI, because we tend to have a lot of projects that end with the government financial year, on 31st March, and no matter how much we try to plan not to have everything completing at the very last minute, somehow they always do.
This March has been particularly hectic because it’s seen the end of the InnovateUK-funded R&D programme of work we’ve been working on over the last four years. If you’d like to learn more about that, take a look at the highlights of the programme as a whole.
So I have spent a lot of time reviewing and signing off other people’s work.
Tensions between reviewers and the people whose work is being reviewed have been a common challenge at ODI. The senior internal review process we try to adopt now, particularly for reports and other substantial publications, is roughly:
- An early discussion on the target audience and scope of an output (30-60mins).
- A mid-point review of the structure / flow based on a draft with headings and bullets that indicate rough content; usually this is best as a combination of reading/commenting (30mins) and discussion (30mins).
- A final draft detail review that focuses on content and language (I get through roughly 20 pages/hour if the document is in reasonable shape).
- Sign off on the copy-edited and formatted final output (30mins).
This should be complemented with other internal and external reviews too. But I will say it doesn’t always work out like that. Some of the early stages sometimes get skipped, or there are direction changes between steps, or there isn’t much time between them.
I find this tricky because:
When I tap into them, I have pretty high standards and strong opinions on everything from content to structure to grammar, so am seldom 100% happy with anyone’s work (including my own). I also feel a lot of responsibility for ensuring ODI has a reputation for quality.
I rarely feel totally confident about something unless I thoroughly comprehend it; I need to probe and understand the basis for statements and conclusions. But in reality, aside from academic research papers and Wikipedia articles, documents seldom need to or should contain the level of justification that makes me feel completely comfortable with them, so I end up asking “how do we know?” questions that take time to answer.
I am super aware of the knock-on impacts of the feedback that I provide, both on those whose work I’m reviewing and on project planning and delivery; tight deadlines mean that sometimes I hold back and let things go, and sometimes I have to adopt a more directive leadership style than I’m usually comfortable with. It means there is a limit to what changes are feasible.
I have limited time and rely on people to tell me what things need my attention, which means that I have blind spots – a lot of unknown unknowns – which make me itchy, but I also think it’s important to (a) trust the team and (b) preserve my downtime.
The upshot is that amongst all the work that’s just been published, there’s some that I feel really pleased with and confident about, and other pieces where I wish we’d had more time and more eyes. BUT all the pieces of work that we’ve progressed this year will be directly useful to ODI’s future programmes of work: there will be time to iterate and improve them, in the same way we were able to gradually refine the Data Ethics canvas over a number of years. We’d love feedback on them:
The Data Landscape Playbook will be used within our Data Ecosystems and Innovation programme, as a methodology for understanding current data ecosystems and thinking through the logic model of data access initiatives.
The Trustworthy Data Stewardship Guidebook will inform our Data Assurance programme and has already helped us think about the kind of support that organisations actually need to assess, build and demonstrate trust and trustworthiness.
The Sustainable Data Access Workbook will help give some structure to the conversations we have with all kinds of organisations, but especially data institutions, about how to adopt business and revenue models that support (rather than restrict) data sharing.
The other thing I know is, in the words of the Agile Prime Directive, is that everyone on the ODI team – from the researchers, delivery managers, and project leads, through to the designers, editors, and marketers – have done “the best job they could, given what they knew at the time, their skills and abilities, the resources available, and the situation at hand”. I’m really proud of them, and grateful for their work, and all my reflections are about how we can create processes and structure that make things work even better in future pieces of work.
This month has seen the “end of the beginning” for the new ODI programmes of work. The programme leads completed their strategies and plans for 2021, presented these to the ODI Board and received some final steers, and the new programmes officially started on 1st April.
I wrote last month about structuring questions-options-and-recommendations papers for the Board, in the hope that they would help structure meaningful discussion. The good news is that this worked well, and we got very positive feedback from the Board about the way the questions were pitched and structured. I have found it extremely personally satisfying being able to help make that interaction between programme leads and the Board work well.
I also wrote last month about my unease around the setting of KPIs for the programmes. We managed to secure a few days of ex-ODIer Phil Horgan, who is a monitoring, evaluation and learning expert. He gave some useful feedback on the logic models we’d put together and some suggestions about appropriate KPIs against them. When we’re back after Easter we’re aiming to finalise those.
The final challenge is reorienting the team, most of whom have been focused on the InnovateUK R&D work over the past few years, towards the projects the programme leads are aiming to take forward, and turn the grand strategic plans into concrete implementation. But that moves outside my area of responsibility.
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. During March, I took this conversation to the Senior Leadership Team at ODI, and work on it will be embedded into two ongoing pieces of work:
- Shaping our narrative and key messages about the ODI, as part of our editorial guidelines.
- Designing the process for deciding what work we do.
We’ve been round this loop a few times before at ODI, but the new programme structure, focus on SDGs, and new leadership in key areas present a good opportunity to get something solid in place that can underpin our work for the next few years.
Diversity, equity and inclusion
Looking back at this month, a focus on diversity, equity and inclusion (DEI) has been a constant theme across all aspects of my work and life. In terms of internal work, there have been a couple of things. First, the ODI Senior Leadership Team had an afternoon of diversity and inclusion training from EW Group, which I was also included in. And second, arising out of that, I have been working on designing a diversity census for the ODI.
It is interesting, and challenging, for me to reflect on my attitude and approach to DEI, as a person and a leader. I am a woman in the tech industry, and bisexual (but in a relationship with a guy since my early 20s, so I look hetero) but I have seldom personally felt like I’ve been set back or had a markedly different experience from other people around me on account of those characteristics. On the few occasions I have encountered problems or barriers, correctly or not I ascribe it to my youth (when I was young), personality (eg being an introvert), values (eg money is not my primary motivation, so it’s reasonable when resources are tight for me to be paid less than the men around me who are motivated by it) and luck. And when positive things happen (such as being asked to speak or be on an Advisory Board), it offends me to think that these are because I’m a woman rather than my work.
In addition, I never feel happy professing expertise in something that I haven’t researched, studied and spent time thinking about. (Just this month, I was asked to speak about gender bias in AI and said I couldn’t because it isn’t an area I’ve specifically looked at.) I also feel it is particularly problematic to do so when there are scholars, practitioners and organisations from minoritised communities who do have that expertise and whose voices need to be heard.
Consequently, I always feel very uncomfortable when I’m asked to talk about being a woman in tech, let along other aspects of diversity. That’s spread into being reluctant to take a leadership role around DEI within ODI, or to push ODI to do so within the wider data community.
But here’s the thing: stepping back to make space for others looks a lot like stepping back because you can’t be bothered, or don’t believe in the cause. Stepping back to make space doesn’t support those who are stepping forward, it leaves them isolated. I am coming to the realisation that I and we need to be more like the first follower. This is a hard realisation because I am filled with regret and guilt for the times I have not been that, and knowledge that I may well continue to fail people in the future. It’s also empowering because helping others achieve their goals is one of my favourite things to do. And it’s challenging because amplifying others can turn into exploiting them, and supporting them can result in unintentionally taking over or drowning them out, neither of which are things I want to do.
Practically, what that’s meant this month is that I’ve gotten my teeth into taking forward an idea from the Diversity, Equity and Inclusion Group within ODI: carrying out a diversity census of the team and Board, to help prioritise our DEI activities and create a baseline against which we can assess the results of our efforts. This was something the DEI Group first looked at in September, but – because of a lack of time and, I think, a feeling there wasn’t high level support – weren’t able to progress.
I’ve spent several very happy hours documenting:
- why a census is necessary
- the principles that the process and the census have to meet
- the process and timeline for completing it
- the questions and options for the census
- how it will technically be administered and analysed
- all the privacy considerations
It’s been fascinating exploring which questions to include (what kinds of diversity do we care about?), finding the standard ways to ask them (thank you GSS and ONS for publishing harmonisation guidelines!), and working out how to adapt them to the ODI team. I’ve also enjoyed getting practical around privacy, ethics, and participation – things I talk about all the time – in a concrete data collection exercise.
Once it’s all in shape, we’re planning to publish reflections and the materials for others to use and learn from as well as actually running the census to understand ourselves better.
This month I have contributed to a couple of roundtables that have been really expertly led and put together by our Head of Public Policy Milly (who is simply brilliant). These were on:
Inclusive Data which we ran in partnership with the Ada Lovelace Institute and the Centre for Public Data where we hoped to support organisations responding to the ONS consultation on inclusive data, and to learn from them about things we should include in our own response. You should read the report from the roundtable, particularly the case study about data collection by Gypsy, Romani and Traveller communities, and listen to the provocations if you have the time. It was such a rich discussion and I loved how Milly framed it and encouraged us to think of inclusive practice as “rich, sophisticated and rewarding” rather than “difficult, complex and challenging”.
Innovation and the Data Economy: Opportunities for UK Regulators which we ran in partnership with the Better Regulation Executive and have yet to write up. It’s clear to us that regulators have a really important role to play in creating open, trustworthy data ecosystems for their regulated sectors. The question is how to help them to do that better. We explored areas that are relevant to our new programmes of work, such as how to stimulate innovation, develop open standards, and improve data governance across regulated sectors.
Milly has also been structuring and leading some work around embedding some of the key things we think are important into discussions around the G7. Ideas like:
- open data standards, allowing data interoperability and building on the policy priority of digital technical standards
- open models for cooperation and innovation, recognising the G7’s ambition to place the needs of open societies at the heart of the technology debate
- data institutions for trusted data sharing, crucial for data free flows with trust, cross-sector collaboration, pandemic resilience and climate action
are important across all the G7 themes. We’ve been working with Gavin Freeguard and Mark Boyd on this; we’re having more traction than expected; and it’s a team that likes to have fun! Milly’s leadership has again been superb.
I also helped shape and frame our work on the impact of the pandemic on teachers’ lives, which was led by the Allegory team, researched by Dr Miranda Voss and supported on the data analysis side by Mime Consulting using data opened by the NASUWT. This is one of what we call internally our “Flagship Stories” – where we make points about data and data infrastructure through a story that will reach normal people. These are really difficult to pull off, because it’s easy to lose the data message in the consumer story, or to be so pernickety on the data front that you lose the story. The big challenge with this one was getting to a state where we could say anything interesting at all on the back of a survey with a low response rate.
I also contributed to our consultation response on Transforming public procurement which enabled us to resurface work that we did years ago around how to embed the creation of (open) data infrastructure into public procurement ITTs and contracts.
This month I’ve chatted with:
- Advisers at No.10 and HMT about government’s data ambitions.
- Art UK, which is a brilliant example of a civil society data institution, holding a catalogue of art held in private collections across the UK and with an interesting and ingenious business model that includes helping those collections sell merchandise based on their artworks. I really hope that we can amplify their work more and work more with them.
- Catherine and Molly from Creative Commons.
- Various people at DCMS about the next steps with the National Data Strategy, CDEI and the G7.
- Martin and Swee Leng at Luminate.
- Ira at NHSX, about open standards and G7.
- OpenStreetMap about open addresses.
- The Smart Data team in BEIS.
- iSPIRT about data institutions.
- Simon McDougall at the ICO, and Tabitha Goldstaub for some good gossip and plotting!
- People at the Ada Lovelace Institute about their work on the location data ethics project (which I’m on the Oversight Group for)
- GOV.UK about content taxonomies, to follow up on previous conversations surfaced at the GOV.UK Advisory Board I’m part of.
- Rockefeller Foundation about our data institutions work.
- People at a big international company about their data strategy.
- Microsoft about the second year of our programme of work with them.
- Lord Tim Clement-Jones in a brilliantly energetic chat about our work and the UK AI and data landscape more generally.
- Natalie about next steps to support better data infrastructure in the justice sector
- Christina about unions as data institutions
- Andrew about good governance around data access and organisational data assets more generally. He introduced me to the notion of steward-owned companies which is really interesting.
- Lorrayne and Bertrand from the Internet & Jurisdiction Secretariat about their upcoming report around data sovereignty and the free flow of data.
I took part in a roundtable on best practices and guidance on sharing the value of NHS data, in support of some work commissioned by NHSX and being delivered by Imperial.
I also presented to the Legal Services Board about the role of regulators in data ecosystems.
I was interviewed twice by researchers doing research for DCMS on:
- how organisations get private value from data
- the economic value of the wider sharing of data
and about an ESRC initiative to create data institution(s) to support researcher access to new and emerging forms of data.
Making progress around GPAI has been a little difficult this month, simply because it’s a young organisation that has a lot of ambition but not yet either a clear direction or focus, nor a tested process for identifying one.
The Working Groups, including the Data Governance Working Group that I co-chair, were up and running last summer. The Steering Committee, which is supposed to provide direction (and brings together government representatives with elected non-government members) was only elected at the end of last year.
The Working Groups have spent the last couple of months putting together concept notes for potential projects to run over the next 18 months, and the hope was that the Steering Committee would be able to provide a steer around which of these to prioritise. But they are still forming as a committee, and the feedback we’ve been given is a combination of things we already know (eg that we need to prioritise), that are difficult to achieve (eg deliver something impactful, quickly), and that are contentious (eg don’t make recommendations to governments).
I am sure these are teething problems and that discussions over the next month will make things clearer, but it results in a challenge for Working Group co-chairs like me who have to keep Working Group members, especially those who have donated so much of their time and thinking to the creation of concept notes, active and engaged.
The approach we’ve taken is to anticipate the kinds of things the Steering Committee is likely to ask us to do, and started work to:
Prioritise and merge the concept notes for projects that we wanted to take forward to narrow down to 2-3 we can take to the Steering Committee for discussion, and hopefully progress.
Create a list of data governance activities that we can embed into SDG-focused projects led by other Working Groups, such as around drug discovery, pandemic response, climate change or the future of work. These include practical implementation of some of the areas we think are interesting from a data governance perspective, so should serve the dual purpose of supporting those projects and advancing understanding around data governance.
Bring together the concept notes for the larger set of project ideas into a publishable “applied research and development agenda”, so that the work is surfaced outside GPAI and can hopefully prompt governments and others to fund projects in those areas.
Through all this, I have to say that it is just a joy to work with my co-chair, Maja Bogataj Jančič, who is funny and frank and so kind and generous, and with the superb support of Ed Teather who is brilliantly diligent (and has good taste in board games). We also have a lot of truly awesome, thoughtful, hard working experts in the Working Group itself, and a new cohort (with new energy) has just joined as new countries have signed up to GPAI. So thankfully that side of the co-chairing role is great!
I attended the GOV.UK Advisory Group this month, where we discussed their priorities and roadmap. There are some people (in the advisory group) who think of GOV.UK as a destination or one-stop-shop for citizens and businesses who need to interact with or get information from government. And there are people like me who think of it much more as a platform of official information and transactions that others should be able to reuse, re-present and interface with. It reminds me of the debates I used to have with John Sheridan when we were working on legislation.gov.uk, trying to work out what functionality we should provide as a public service and where we should leave space for others to build additional tools on top of it.
Conference season hit with a vengeance. I spoke at:
- The Public Sector Innovation Conference, about health data innovation (watch here)
- Northernlands 3 about how we should consider individual and collective, and “donor” and “victim” interests, rights and participation in the design of data systems
- Everywoman Tech Summit about developing data strategies that look outside business boundaries
- The Turing AI UK Conference talking about “Navigating the AI Council’s AI Roadmap”, particularly focused on data obviously
- An OECD event on Addressing the Gender Gap in AI Data
- The PARIS21 Annual Meeting about data literacy. (They’ve put together a great briefing paper on the topic of data literacy that it’s worth taking a look at)
I was also (very briefly) on BBC Radio Wales, talking excitedly about the census! My first radio gig in a while.
Thoughts I had
Aside from DEI, which I already touched on above, another theme for the month has been about the role of organisations like community groups and unions in collecting and reporting data about their members. We talked about this at the Inclusive Data Roundtable we ran, demonstrated a form of how it can work through our work with the NASUWT around teacher experience in the pandemic, and it’s been a theme in several of the conversations I had this month.
The idea of “counterdata” is one I was introduced to through the amazing and highly recommended book Data Feminism by Catherine D’Ignazio and Lauren F. Klein, as a way of providing an alternative narrative to that provided by those with power and privilege who so often shape data collection.
We can see some of this play out in the report of the Commission on Race and Ethnic Disparities, which makes great play of being informed by data collected by the Cabinet Office’s Race and Disparity Unit. I was struck by this sentence in the report (my emphasis):
We suggest that pessimistic narratives about race have also been reinforced by a rise of identity politics, as old class divisions have lost traction. Well organised single-issue identity lobby groups also help to raise the volume. These organisations can do good work protecting the vulnerable, but they also tend to have a pessimism bias in their narratives to draw attention to their cause. And they tend to stress the ‘lived experience’ of the groups they seek to protect with less emphasis on objective data. It is not surprising therefore that mainstream public debate about race sensitises minorities to discrimination, but does less to highlight minority self-reliance and resilience.
This sentence highlights the common assumption that data is objective, the implication that it is superior to qualitative information, and the idea that negative experiences felt by some people can be balanced against more positive experiences felt by others (ie that what matters is the average). All these are problematic. Data does not simply emerge as a ground truth, it is collected and shaped by those who collect it, especially when it is about people. Data is only one part of a richer set of material that people have to use to fully understand reality; it is necessarily simplified and restricted. And I simply do not agree that harms in one place can be balanced by benefits elsewhere: they should be reduced regardless, because they are harmful.
Back to the point I wanted to make: I do think that unions, consumer advocacy groups and other civil society organisations have an important role to play in gathering data to represent the experience of their communities, especially in a context where authorities place more weight on “objective data” than they do on other forms of evidence. We also have to prepare for authorities to ignore and undermine that data, critiquing its methodology and biases while failing to provide the same critical lens to its own. And we have to avoid falling into the same traps of thinking that data is the only form of truth that it is worth surfacing.
- I have made no progress on getting through my email backlog and am struggling to keep up with following ups from meetings.
- I have found it very hard to be in a good routine around reflecting on my day and planning my priorities for the next one; I need to be firmer about the end of my work day and using that daily review as my ending-work routine.
- I am not using my standing desk enough!
On the positive side, getting down to work at 8am is still going well, but as we come into April I’m not going to have as much reviewing to do, so planning how to use that time fruitfully is going to become more important.
My kids returned to going into school near the beginning of the month, and started their Easter holiday at the end. My partner has received the first dose of the coronavirus vaccine, but I haven’t. We’re still mostly in lockdown.
I’ve started growing some seedlings, including some sunflowers from seeds made available by one of our neighbours. Growing plants from seeds was something I did last year, during the first lockdown, and really benefited from later in the summer when the flowers and fruit came out. I’m not sure they’re growing as well this year, perhaps because I’ve got them at the back of the house where there’s less sunshine.
There’s been quite a lot of stress and excitement in the house because one of the games my partner has been working on for years – Outriders – has just been released and is doing well (with some teething problems with servers). It’s not the kind of game that I enjoy, but if you’re into third-person shooters, you should check it out.
Spiritfarer – we really enjoyed playing this, but then got very sad when most of our passengers crossed over and we felt we were reaching the end of the game. So we actually stopped playing because we didn’t want it to end – there are some updates coming so it’s not as weird as it sounds to pause for now.
Haven – my eldest and I avoided this for a while despite recommendations around it, just because it felt like it would be a bit weird playing as a couple (who do get intimate with each other on occasion). But actually the way the dialogue system works means that you don’t really inhabit the individual characters, which makes it less awkward. We’ve really enjoyed it: exploring, gathering supplies, working out collaborative strategies for different enemies. The story is interesting and the dialogue natural. Yet to complete, but one I would recommend.
7th Sector – a puzzle game with a nice feel. I stopped playing though because I found the more physical / timing puzzles too unforgiving and frustrating.
Genesis Noir – adventure/puzzle game that is an indescribable mix of cosmology, civilisation development, love story and jazz, with a unique aesthetic. I did enjoy this.
Stardew Valley – I do feel slightly ashamed of this, but I got through all the Sudoku puzzles, and felt the need for something that I could play on my phone when I’m brain dead. But it is horribly addictive and I find it very hard to stop when I start playing, so it’s not something I can spend just 20 minutes playing at a time.
TV we’ve been watching:
- Lupin – very good
- Call My Agent! – we loved how this could go rapidly from comedy to drama, and how every character is both sympathetic and flawed; the cameos would mean more for people who watch more French cinema and TV, no doubt.
- Money Heist – just brilliant, I don’t think I’ve ever been quite so fearful about every twist and turn of a plot; we’ve just started the second series and know it’s going to turn out badly (not least because all the relationships we’re rooting for are doomed) but can’t help hoping for everything to be OK in the end.
- Law & Order – part way through series 5 now and two female leads (who just got swapped in for existing characters without any explanation)! We were very sad when Ben left, and still find D.A. Schiff absolutely hilarious in his pessimism.
Family film watching:
- Big Hero 6 – watched for about the fourth time I think, but still love it
- Mr Jones – flawed but not dreadful: could have done with being a bit shorter, and I didn’t believe in either the romantic relationship nor anyone’s support for Stalin’s regime
- White Tiger – really enjoyed this; pacy and thought-provoking; worth watching
I’m not going to get into all the details, but the return to school for my youngest has been even rougher than I anticipated and wrote about last month. Fortunately, the school has been incredibly supportive and flexible which has enabled my youngest to participate more than they would otherwise be able to. I have always tried to focus on what I see as important from education, namely fostering curiosity, a love of learning and a supportive peer group rather than academic achievement (the kids get more than enough pressure from school on that front). Right now, my sights for my youngest are even lower: I’m just prioritising them not being cripplingly miserable.
Completing the census for my non-binary child and some subsequent conversations on Twitter highlighted to me the frustration of identifying as non-binary in a world where binary gender is such a built-in assumption. I do understand how and why people can find it hard to think about gender as a social construct, and therefore one that is fluid and can change. I get oddly more annoyed about the lack of recognition of intersex people in official and medical records (perhaps it’s because it’s a data modelling and quality issue). The fact that whether an NHS number is odd or even indicates whether someone is male or female, for example, is an illustration of how data simplifies reality. It really feels like we should be able to do better.
But trans visibility day on TikTok lifted my spirits. I’ll leave you with some I particularly enjoyed: