So, I’m currently working on the Probation Reform Programme a huge piece of work to transform the way in which we manage people on probation and ultimately try and rehabilitate them so they don’t re-offend.
In the first year I was working as a contractor doing user research, and in the last two years I’ve worked as a Service Owner. The role, as I’ve interpreted it, has been a mix of product leadership and service management (procurement and contract management, and getting the infrastructure in place for teams to deliver value). There are a bunch of learnings I want to share about this last few years, and the first is about prioritising.
In the commercial sector priorities for determining product decisions are driven by the requirement to make a surplus. Even the unicorns (in fact especially the unicorns!) funded by VC have the goal, ultimately, of returning a healthy return on the investment they receive. Determining the product -market fit and the value to customers is hard, but clear. Will enough people pay to hire your product or service over another? This is how you determine the value your product has. All subsequent product prioritisation decisions then stem from this – segmenting your customer base, growing market share, creating a defensible competitive advantage etc. to extend and drive that value higher, and into the future. Frameworks like RICE help product managers figure this out, and user needs and data on usage and behaviour give you a steer as to the right direction to test and iterate.
Building products in government ought to be simpler; there are very few competitors for your audience and you usually have a monopoly*.
Yet, because there is no clear way to determine value because your users rarely have a choice but to use your service, product decisions are complicated.
Here are three things that have helped me to determine how to prioritise work at MoJ.
Focus on the thing that meets the policy intent / or desired outcome most cost effectively. In the probation programme the ultimate goal is to reduce re-offending. In the area of Interventions, one of the decisions I made early on was that unpaid work (sometimes known as community payback) was not going to help in achieving that outcome as much as a rehabilitative intervention such as an accredited programme. That value decision seemed very clear to me, despite unpaid work having a huge backlog of projects that needed managing, reducing that backlog wouldn’t have helped achieve the outcome. It’s not always as clear as this example, but developing hypotheses to test should provide confidence in which direction to take.
Go to where there’s clarity and agreement on policy. One early discovery we did was around resettlement. When people leave prison the first 48 hours are vital to get them on the right path. We *knew* this from existing research and we felt that digital ought to focus on this problem. However, the policy about how to support people leaving prison was in its infancy, and there was a lot of uncertainty about the direction of policy. So despite knowing it was a key area, we stopped the work at end of the discovery period because we knew that the policy wouldn’t be able to support decisions the digital team would need to make: it was too soon. Whereas on the first project I was involved in – delivering value to court – we chose to focus on an area the organisation wasn’t that interested in – admin! Admin tasks rarely make it into a target operating model, or policy guidelines, and this means it’s often easier to deliver work quickly and effectively into the hands of users and show the business what digital can do. Having proven the value of digital and won trust of stakeholders, that team was able to make greater inroads into areas of court that were the preserve of policy.
Ensure the digital levers at your disposal are powerful enough to make a difference. Software has the potential to change behaviour radically, and we see this in everyday services we now use to travel, book events and connect with other people. Yet sometimes software can’t effect the change needed to make the cost of building a digital service worthwhile. For example, in managing unpaid work the cost of providing devices for supervisors to use in the field (to manage attendance records etc) is significant. A tablet or smartphone is also open to be stolen, or broken, and wifi and data signals can be patchy. Paper is robust and reliable in this situation! Similarly, the resettlement team (above) found in their discovery that the ability to impact significantly on the lives of people leaving prison meant getting buy-in from those managing benefits journeys (DWP) and also commercial contracts with accommodation providers. Any effect they could have was dependent on others (departments, agencies) and whilst that that in itself didn’t prevent the team progressing, it was a significant factor in deciding not to prioritise the work into Alpha.
* some platforms provided by government have competitors – Notify, for example, but citizen facing services are (all?) monopolies.
Update: As Keelan points out there are GOV services vying with commercial services to meet users’ needs:
It seems ridiculous that I struggle to remember what I do when I reflect back on my week. And yet I do struggle. A terrible memory may be a contributing factor, but I also want to believe that it’s the nature of my job, too. My job as Service Owner at MoJ involves a lot of time spent doing lots of small discreet tasks that rarely add up to anything that significant – they have my attention for 30 or 60 mins, sometimes my partial attention as I try to complete work in parallel to talking. Then, gone. It can be a challenge to remember what a meeting is for, and why I’m there, as I bump from one to the next. So trying to couple these events together for the purpose of a weeknote narrative is hard. They’d make more sense as (really dull) tweets or instagram posts. Maybe that’s what I’ll try next week – aggregating moments from the week and see if there’s a thread.
So my week as a partially remembered list of things looks like this:
Benefits realisation feels fundamental to any goal but the amount of effort that goes into actually thinking about this is far outweighed by any actual benefit realised. Fact. OK, probably not, but it seems like it this week. Also, saving time and better data quality are not benefits to realise, they’re the things that could help you realise some value.
A meeting IRL with the management team and in Sheffield, the low light of which was discovering we have over 16 boards in our organisational area. That’s a lot of attention to manage, or avoid managing so you can get stuff done.
Engaging with GDS expert services to try and help deliver technical dev capability on a team. Didn’t know this support existed. Odd. Happy!
Engaging with a bunch of performance analysts to ask them to come and talk about their work, to try and help our teams understand that role and how we could benefit from using it. Baffles me how we don’t bake performance measures into everything we do and make it simpler and easier for all users to see *how* our services are doing. From my time at the Co-op it’s the single most powerful thing you can do to engage stakeholders in the development of a service.
Building regs require us to use Intumescent paint, which is paint that protects steel from extreme heat in the event of a fire, giving (hopefully) time to escape. Yet, the only proof of you using this paint for building regs seems to be a note you bought the paint. Any why protect the steel and not the wood when the entire floor is wood up 3 stories?
My chronic achilles pain is back. And I realise I take an almost obsessive interest in other peoples’ injuries, as this pain has been such a big part if my existence for the last 2 years. I offer up unsolicited and probably unhelpful advice on anything bad happening from the waist down; hips, knees, calfs and most especially achilles and feet. These two wonderfulbooks are all I have to bolster my ‘expert’ claims. Call me if you have a niggle.
Octopus energy have told me that the two year fixed term price on our contract is ending and that the new tariff options will mean our household bill rising from just under £2,600 a year to over £6,889 a 265% increase! Gas rises from 2.89p per kWh to 9.65p per kWh, and Electricity from 16.51p per kWh to 35.14p per kWh. I’m still processing this but I do know that it is this kind of extreme change that will wake me into doing something radical with our energy consumption – insulating the hella out of the place, and looking at alternative energy (I think this may be a positive spin on what could just be an example of shock doctrine!)
I bought 7 pairs of shoes to try on because I didn’t want to have to travel to Manchester or, in fact anywhere, to have to try them. Unduly interested in the model around retail and returns. In Germany, apparently, there are far far more returns than in the UK and retailers and suppliers build this into their business models (fast, clear, hassle free process). I feel close to the German shopper.
This last week was marked by the first meeting of all the Folksy team IRL since 2019. The team have grown to four full-time and two part-time staff and this week most of them met up to discuss objectives, priorities and a product strategy. I joined them for the day on Wednesday to facilitate a discussion on ‘jobs to be done’, to help determine the key jobs people hire Folksy for, and where we compare against competitors for meeting key needs.
I love doing this kind of design research and product strategy, and I’ve missed it. Organisational strategy in GOV is often woolly and opaque, and product strategy is whatever gets you to simpler, clearer, faster (rather than market and surplus).
So, this week was energising! We have a strategy for growth based on user retention, and doubling down on quality and improved delivery times (and clarity on when an order will be delivered) 💥
Other things of note this week:
Mitigation to cyber security threats. So many unknowns around this and the advice is often very patronising – so I spent a chunk of time looking into what the mitigations are. Things like pen tests give a false sense of risk when you have a rapidly changing suite of products, and evolving Dev Ops and teams.
A few wobbles at work with regard to how to do transformation in gov led to me re-watching and sharing this presentation by Tom Loosemore at Code of America. Eight years on and the ambition and drive of those early GDS pioneers still seems so relevant and he’s right; it’s not complicated, it’s just hard.
A week mostly spent trying to figure out how to stop teams needing to use arcane processes (support, data protection and information assurance things) that are from a different IT based era. And also working out how to define our spend in relation to ‘capital’ (capital, resource) that I found oddly engaging (rules, assumptions etc). Also met some people from one of the product teams I’m responsible for IRL which felt far more exciting than it had any right to, especially as I spent most of the day in meetings via a screen.
We’ve tanked out our cellar during lockdown joining a swathe of other bored middle class families in making more of the space we have (when it was actually already enough but anyway). It’s been a long slog and that’s just as an interested observer.
One issue we’re now trying to fix is the heating in the house which went awry when the cellar zone was added and more rads plumbed in to a plastic veins passing hot water. We lost heating from a bunch of radiators and it’s been a complex job trying to figure out why, especially now the walls are plastered and the plastic piping hidden.
So this last week I’ve learnt about heating: pressure, pumps, balancing and some praying and hoping. There is undoubtedly some bad design work in the way the system has been reconfigured but with a new pump, and some ongoing balancing the previously lukewarm radiators are now (mostly) hotting up. Win! In trying to fix this heating my head was turned by the prospect of smart thermostats that will make things better. Trying to make this very dumb heating system smarter feels a bit like ‘lipstick on legacy’ that I spend my working hours avoiding. However, I haven’t really bought it for any smarts, rather I’ve bought it because I just wanted a thermostat that was easier to use.
The management team at Probation Digital decided to embrace OKRs this last few weeks as a means to bridge the gap between some log term goals and sprint goals. Super pleased we’ve decided to do this after promoting the idea last year. So this week I’ve been reading some of the books and blog posts from previous attempts at this, and also some more from Neil. Rather than the format Christina Wodtke advocates (more of a qualitative objective to get teams fired up and enthusiastic) we’ve gone with an hypothesis based approach, which Gav (one of the Delivery Managers at MoJ) has used before and taken from the book Super, Safer, Happier (though there’s not a heap load in the book about OKRs, it is a fantastic resource for doing agile in large organisations).
Conscious that we want teams’ to own these, and want to find a format that works for us, this week has mainly been about drafting things and learning what feels right, and why.
One of the things that’s helped me to figure out where to focus is the Eisenhower Matrix – which Wodtke mentions:
The stuff that is important and urgent is the *stuff you’ll do anyway* whereas important and non-urgent are things ought to lead you closer to your goals. I see that a lot at work; adding questions to a risk assessment, providing a route to a different type of referral or case r etc. help operational work to be done, but won’t transform it. So OKRs can help us look up and focus on the non-urgent but probably more transformative work.
A far more trivial thing that I’m excited about is to write and sketch more, of which this is a first attempt, and I’m trying paperlike on the iPad to see if that makes it feels more natural. Seems to divide opinion.