February 22 2019


We had some playback of last week’s user research in our Hedge End office. I was remote the whole day — tricky but surprisingly good. The main issue was motion sickness as Steve (product owner) did his best Blair Witch “walking around with laptop facing outwards.”

For me the primary insight from our labs day is this:

In a context where you roughly know the landscape, you can muddle through — if I’m at home and I’ve lost or run low on my meds I generally have a couple of points of contact:

  • My GP surgery (and associated out of hours service)
  • The pharmacy I regularly sort out my meds with — be that in person or via delivery

While I might not know the official procedure, there are enough dangling service hooks for me to bother a few people, and then get onto a service track that can help me.

If the context is for some reason different (for example, I’m on that wilderness retreat I’ve been dreaming about ever since starting work on 111 online), then things break down. When probed, none of the people we interviewed knew what to do if they needed an emergency prescription whilst (for example) away from home on a bank holiday.

There didn’t seem to be a general understanding of what you’re actually supposed to do in these situations. In a normal context, people essentially muddle through. In an abnormal context people are stumped.

I think this validates our service proposition, the primary need we’re aiming to meet:

“I don’t have my regular meds with me, and I need to take some soon — help, what do I do?”

Later I ran through the service maps and user flow for what could be our MVP. It felt slightly weird to be a disembodied voice and a mouse cursor on a big screen somewhere, but it seemed to go okay. The team had excellent comments, questions and challenges. Sometimes I hate them all.

We also had a session discussing how what we’re designing could be built. This has been a weak spot of mine within this team: the main build team are based in a different location, and that’s something I’ve not been used to. So in order to improve we’ve been starting to try to have better sessions on how we get from epics (call it whatever, the big thing, the feature) into a set of smaller tasks. Can pieces of an epic be releasable small features by themselves? How many iterations do we think small pieces of production code might need in order to achieve the feature we’re after?

Busy day.

Tuesday onwards

Mainly off doing personal stuff.

For this part of the week I’ve been just dipping in remotely. It’s interesting how new(ish) tooling has made remote working a lot easier. Slack and Realtime Board are two simple tools that really change the experience from the old phone / email / dead silence routine.

But it’s still easy to feel out of the loop — knowing that there are people sat together having those chats where a lot of the thinking is thrashed out.

It definitely feels like full-on remote is doable, and co-location is doable. The range of working going on between those two poles is where a lot of conscious communication effort is required. Ho hum, I have no magic solutions.