So, a few things to talk about this week, but – to be honest – none of them are as exciting as Cate Blanchett’s dancing in the video for Sparks’ “The Girl is Crying in Her Latte”.
(The song is…OK. Not among their best, and it’s a bit of a grower…but hey, Cate Blanchett!)
Anyway, it’s been something of a tiring week, but probably in a good way. I’m definitely “tired, but not exhausted” – as one early books about Scrum used to say.
This is a welcome change. After the blip at the tail end of last week, working days are now looking largely sensible. On days I’m in the office, I’ve also got a shorter commute – which is hugely welcome. But these cumulative changes have definitely messed with my rhythm a bit: I’m not used to having this extra time, and I even accidentally went to bed an hour early on Tuesday because I’d ticked off all the ‘normal tasks’ and blindly assumed there wasn’t capacity for anything else.
I’m still grappling with the whole tasks/goals balance in the new role, and slightly trying to work out what the job really is. And isn’t. I’m aware that I personally took on a lot more than I really should have done at DIT, and got way too emotionally invested in far too many things, which led to getting pretty burnt out – and I don’t want to end up there again. So I’m still sifting through everything and deciding what’s a goal, or an opportunity to make an adjacent big difference, or a quest other people hope I might take on (for better or worse). Accordingly the goals list in my A3 sheet I talked about two weeks ago is getting really quite long. Nearly as long as the tasks list. Not all of that is going to be equally valuable – but I don’t yet have a handle on how to prioritise. But whatever – this all leads to good opportunities for user/stakeholder research into Me As A Product. I have a few thoughts on how to better design those sessions, but nothing solid yet.
The Vision Thing, Leaders and Leading
On Monday, a few of us presented our vision for Heads of Profession to the GDS People Board, and to a subset of our Deputy Directors. I was quite pleased with the structure and content of the story we told, and how we organised it as a set of hypotheses and opportunities. Feedback from the first session was pretty good, with good (if perhaps slighlty premature) questions about metrics. The second session was a little beige – something didn’t quite chime with them. Still need to work out why – which ties into the better stakeholder conversations in the previous paragraphs.
I had the first one of those in on Tuesday afternoon, and thoroughly enjoyed it. Got some useful confirmation on the particular issues I thought I could see in that DD’s area, and positive feedback on how I was proposing to help. I think we are going to be a good team, and I’m going to bring some useful capacity to help us improve things together.
One thing I’m pondering on the back of that discussion is how much it’s the responsibility of Product people to “lead”.
Personally, I tend to think that in high-functioning multidisciplinary teams, the vision and strategy should come from a combination of design, technology and product working together – and it should be something continually validated through user research, analytics, economics, the evolution of the roadmap to emerging circumstances etc. Product people definitely have a key responsibility for making sure the business value side is thoroughly understood – and teasing out the riskiest assumptions in the strategy/approach, so they can be dealt with ASAP. But the bulk of this work should still be done together within the core team. Of course, not all teams are equal, and there are sometimes capability gaps – so the product and delivery people are still usually on the hook for making sure there’s something sensible to go forward with, in a sensible timescale. But if all the ideas have come from the product person, that’s something of a wasted opportunity. That kind of CEO-of-the-product leadership should really be the absolute last resort.
How that plays out in practice (and how you might formalise that as a set of principles for teams) is something I want to work through with the other Heads of Profession. We’re trying to model the behaviours we want to see in our teams – and if we can’t sort this out between us, how should we expect our people to know what good looks like?
Consistent, not uniform
Another thing that’s emerging from these conversations and thinking is that I suspect many of our Heads of Profession actually have quite different jobs. We have things in common, but we aren’t identical. I’m currently wondering: where are we separately strategic, operational, coordinating, managing, inspiring? Each profession has its own balance between those, and there are different pressures in each directorate. So we need the time to think about this carefully as we iterate this vision – because we want to make the most of the areas where HoPs overlap, but shouldn’t force everyone into the same box.
Flashbacks
Elsewhere I found myself helping our Business and People Services team think a bit about how to categorise different types of work. They’ve got a lot of things that “just have to happen”, which are definitely prone to political and organisational ‘weather’. When will an unexpected request from on high arrive, what will a role be greenlit and suddenly have to be advertised, how will you deal with a key person being on leave just as a monthly return to Cabinet Office is due? All that makes it hard to make time for important projects to improve things. And without visibility into that day-to-day, it’s hard for management to have confidence or shape what’s going on – they just see projects running late.
As we talked it through, I realised this was remarkably close to the SRE ideas of ‘service level objectives’ and ‘error budgets’. Were teams over-responding to BAU tasks as though they were “incidents”? Did they have a clear handle on management’s priorities in the most valuable BAU? And if a given project was critical to improvement, is it unnecessarily getting pushed aside by hitting an over-ambitious service level on process x?
It was a really interesting conversation, and I liked drawing up a fictional set of SLOs to help illustrate the sort of choices that might lead to – in teams and with senior management. An unexpected place to be using my creative side, but I think it’s going to lead to some good benefits elsewhere.
(I did have moments where I thought “where are Emily and Jess when you need them”, of course. 🙂 )
“Measuring” Buy-In
There’s been a lot of change happening at GDS recently. GOV.UK has a new strategy, the Government as a Platform programme is hoping to be a more unified and self-reinforcing suite of tools, Digital Identity has succeeded in replacing Verify and might be able to be less heads-down.
With all that rapid change, in a distributed organisation of way more than 500 people, that works in a hybrid way, there’s a real chance of teams and leadership becoming detached from each other. So I’m lightly pondering what leading and lagging indicators might be of scenarios around that detachment such as:
- This team simply doesn’t know the strategy
- This teams know the strategy but doesn’t understand why it is the strategy
- This team knows the strategy, but doesn’t believe your reasoning
- This team knows the strategy, but can’t free up enough mental/delivery capacity to work out how to make any progress on it in their own work, so have become passive
- The team knows the strategy is wrong, understands the goal and believes in you, but needs help and context for coming up with a better idea
- The team came up with a better idea, and didn’t notice you’ve changed your original plans because of their feedback
I don’t yet know if any of those apply anywhere, but it’s a first rough framework to have in my mind when talking to teams – and working out where I can help everyone collaborate and do their best work.
Funnels and Systems and Flows for our Platforms
As I mentioned above, the Digital Services Platform (formerly GaaP), is trying to be a more coherent and self-reinforcing offer. It’s also focused on growth – getting more people to use these products to make their own services quicker and easier to use. This blogpost from Lenny’s Newsletter on Duolingo’s various bets to improve usage has been doing the rounds, and I was really impressed by a few things.

Firstly I really liked how they’d managed to divide up the flows of different user groups and types, and how they measure the rates of dropoffs between them. Obviously with our services it’s more about the adoption cycle – what’s it going to take for someone to use this the first time, to then get it into production, to then use it on multiple services, and finally to try more complex products. But this was the sort of diagram I could half-feel in my head when thinking about how to explain that, had it not been for my aphantasia stopping me.
Secondly I loved how they’d then run the model over time to see which changes to which metrics would have the biggest effect on overall ROI. They were able to prove that keeping current users satisfied was by far the most effective tactic for sustainability in their first phase. Once that was sorted they could move onto other metrics/paths of the network.
(That they were talking about gamification as particular tactics to achieve that is of course much less relevant – analysing the network, and working out what ideas are relvant in your context, then being able to measure their impact is far more important. We need to do that at an individual product level – but also at the pan-product level. For example, how likely is a developer or designer that actively uses Notify to also try Pay? Do designers using the prototype kit understand the problems the rest of our tools are there to solve, and have we made it sufficiently easy for them to build the concepts into their fledgeling service ideas?)
Miscellany
- Weirdly, no new gig or theatre tickets bought this week. I think.
- Had a very, very Zen piano lesson. “You’re thinking about how to get from A to B, when actually you should be thinking about how to play it”. But I’m making good progress on Schubert, and it’s reassuring to see that Debussy’s Sarabande is as hard as I think it is.
- Have been enjoying reading a few small Product Craft books this weekend – little essays really – by both Jeff Gothelf (Agile vs Lean vs Design Thinking) and Josh Seiden (Outcomes over Outputs). Lots of exciting ideas came out of that, but I realise that I don’t currently know where to take that excitement at GDS. At DIT I know exactly who to go to, and we’d go “oooh” at each other for a bit while we worked out how/when/where to apply it. But I don’t yet know how to share these ideas within GDS, how to get buy-in, and to understand the landscape to create some interventions. Probably another place to try some experiments – because I definitely need to come across as “here to help” rather than a troublemaker.
- Hugely enjoyed seeing some old friends at Product in the Aether #42 on Tuesday evening, and was pleased to be able to blithely pass on Tom Wujek’s “mental models” exercise to a new set of people. (I suspect I’m going to need to use this myself quite soon, so it should turn up in a new blog post in time.)
