It’s hard staring new things. This week I had to suss out quite a few things: how to make sure that the future direction of admissions was going to Do Things The Right Way (without making people too nervous). I had our first “technical direction meeting” which I’d been dreading. I realised about halfway through the week that there wasn’t really a way to not feel like you’re in the dream with the exam you haven’t revised for. I’m in a new system, things are not where I expect them to be (I can’t just point to NCSC guidance or platform assurance), chains of evidence are different (user research often means “testing” here rather than “discovery/research questions”) and the way that the team works together isn’t quite what I’d expect, but a lot of that is because we have a lot of roles I’m not used to incorporating (I have only worked with business analysts once or twice and for my money they are often service designers with ties).
This tweet from Cantlin did the rounds.
https://twitter.com/cantlin/status/1402590276900855813?s=21
And I have a lot of sympathy for it in a lot of ways. The phases run a spectrum from “stage gates” through to being reflection points for a mature team to look at what they are doing and invite proper design crit. Plenty of places avoid a serious look at their work, plenty fudge assessments, plenty treat the progression as linear and something that means you don’t bin experiments or that you can’t stop once you’ve started. They also treat a service as one thing, even when it has many parts.
Where I am now, I have the opportunity to think about that a lot. Admissions has about 4 moving parts. All of them right now are what I’d call a working beta, but several are about to re-enter discovery/alpha. They are microservices that talk well to each other and integrate commercial software that it didn’t make sense to replace. But right now, I’ve got two big strands of work coming up. We are working on a data-caching layer to make sure we don’t futz up the Cambridge people database that we don’t control and that doesn’t love millions of API calls. We are also starting testing on a user facing design prototype. These pieces of work are, broadly, separate. They both help the same service, but they are spinning in their own iteration circles for now.
So, putting the whole service as an alpha or whatever is misleading. We are trying to replace old bits with new bits gradually. Trigger’s broom. The new bits will work better and be a bit more stable than the pandemic driven bodge we have now, but we are not working in a greenfield project. We are working on a core university process with 20k applications each year and a hefty complaints procedure for getting it wrong. At the end of this big push to get things over the line, I will still have stories to improve the service and I am pretty sure that some of the microservices will be in an “overburdened alpha” state just because the effort will go into user facing services (both applicants and administrators) and the backend will have minimal gold plating. I think anyway.
I always think of Sanjay’s blog post here. The iteration spirals need to happen, but sometimes it is bits within a service as well as the service itself.
In my case right now it is also important because we get to write our own manual. We are using the NHS service standard as a starting point, but this sort of stuff on disco to live is up for grabs and I wonder what we can do better in an organisation that isn’t used to proper agile stuff, especially mixed with UCD to make a comprehensible and actionable framework.
So the meeting itself happened. We mostly discussed things like security, prioritisation and the rest, but I did manage to put in a spike that I’d been discussing with the user researcher and content designer about renaming the service (it is currently a “portal”). Again, it’s weird when you have to talk about the basics like “the name of the service should be a verby statement that explains what the user is trying to do”, but not to have the backup of an assessment team saying JFDI. There are also complications: apply to Cambridge wouldn’t quite work. As an undergrad, a postgrad, a researcher, staff? There are a lot of ways to apply for a lot of things. Anyway. We have a spike, we will think about it.
The whole meeting I felt like I was being my own therapist. I felt like I didn’t know stuff (I suggested a DPIA strand of work that had already been done) or looked for platform assurance (how sensitive is the data you can keep on name of platform (as we don’t work with official → secret that’s a harder question. Everything is in impact levels like it’s 2013 again)) and had to keep reminding myself that it was day 8 and really, the project history is the hardest bit to get.
I think the thing I’m going to find very hard is that I am meant to be setting up a product management function, but for now I am the only one. My community of practice is trying to talk to a neighbour with a greyhound on the dogwalk who is a director of product for some big fancy company with money and people with “global” in their title. We’ve been talking about doing more work across the university, the press and the exam board as we all have big digital teams and I think that might be an interesting way of addressing it.
Bruce has been a bit ill this week, and still isn’t quite over it. He has got back into his summer routine of lying in the sun until he is far too hot and then passing out in the coldest bit of the house.
No real progress this week. Still getting time sorted with the new job and my energy levels are through the floor. I’m not sure what to do next: the chapter on what “neutrality” means for civic tech and how you have to talk like government and smell like government to be listened to by government (and what that looks like to civil society and the compromises it entails). There’s a bit more I need to read and ever more interviews to do to support the write up. Supervisor is sending me a comp copy of Gillian Tett’s book that he was offered so I might give myself a bit of a summer holiday from PhD guilt and just read a nice (comparatively) easy book instead.