The Wind Down
A four-day week spent closing open loops. Took yesterday off to pack for my holiday which I’m really, really looking forward to – especially because it’s supposed to drop to 1°C next week!
Four Things That Happened
We did more things than this
John Cutler spared us an hour to chat product, and it was really beneficial. We talked quite a bit about valuing non-feature work, how it’s important and continuous improvement needs to happen under the hood too. That sort of work is our main focus on Platform Health, really, and selling it to ‘the business’ isn’t too hard.
But how this work is described by senior teams can irk me. Platform work is often referred to as ‘keeping the lights on’, but it’s really much more than that: it’s chopping wood, filling buckets, building aqueducts…
John mentioned Cost of Delay and it’s a useful prism to view all platform work through. Does the value opened up by doing this story diminish over time? That sort of thinking allows you to take on anything urgent and create value in the short term while also keeping an eye on the long term.
Had check-in with the programme team on Wednesday, or at least we were supposed to have check-in with them. The majority of them weren’t there, which was a let down as we had some important things to tell them. Really frustrating when you’ve had to work through lunch to get the presentation together too, because there’s so little time on Tuesdays and Wednesdays what with all the meetings. And putting the slides together on Monday would miss 20 per cent out of the last two weeks.
Despite that, our new Head of Product, Leanne, was there in force and tried her best to accommodate us. She was pleased with our progress and some of the optimisations we’ve brought to the publishing pipeline, wanting to push us to show it off. I’m really keen to get the team showing off what they’ve done because, as I always say, it’s not sexy work but it’s damn-near as important as building features.
The excellent Emily gave our team a run-down of SLIs, SLOs and SLAs – henceforth known as SLXs – and how they can create psychological safety in our work. We’ve had a go at creating one already (which I covered in S02E03) and this was included in her slides, which meant to work was framed much better than I had introduced it to them. As Emily spoke about the noise of alerting and how to make alerts more meaningful, there was lots of nodding amongst the group. It certainly hit home.
We’re going to be looking at building SLOs for the other jobs of GOV.UK soon, once some important Brexity business is out of the way. In lieu of an actual stress-test of the pipeline, making our current data more meaningful will provide a decent, actionable contingency for the months ahead.
Darren Barnes from ONS came to talk about and demonstrate the amazing GSS Data project. Using linked data principles and technologies such as concept schemes, triple stores, URIs and all that other stuff, they’ve pulled together a neat open data portal that allows you to find statistics about stuff. It was super impressive, and it was nice to hear that they held Mike’s work on ESD and LG Inform in high regard – as an exemplar of what could be done.
What I Learned
That everything I learned about linked data at Porism could be useful during my time at GDS. That was quite a surprise. Darren was throwing out technical terms left, right and centre – and I understood everything.
What I Could Do Better
Compose myself during meetings. I get very excited about things and that’s not always good, especially if your body language communicates a commitment to a project prematurely.
- Prioritizing “Non-Feature” Work and Continuous Improvement
- Emerging Tech Trends for 2018
- Analyzing Experiment Outcomes: Beyond Average Treatment Effect
- How The Wall Street Journal is preparing its journalists to detect deepfakes
If you’d like to receive articles like this in your inbox every week, sign up to my mailing list.