Several weeks after working on the Testing programme I noted that all the designers from consultancies were only trained to use Sketch. There was a design approach with one goal in mind: “use a tool to draw ideas quickly, get them signed off quickly, and code the solutions quickly.”
This was efficient as a process gets, until the team needed to start testing designs every week – i.e. to check they worked well for users before national release.
We did conduct a few rounds of testing using Sketch prototypes, it can be o.k. for some projects (e.g. you can make the prototype clickable), however it’s certainly not ideal for usability testing because:
- you can’t see what a design looks like on different devices (e.g. a mobile view),
- users cannot interact with form elements,
- errors / validation don’t work,
- users are exposed to button prompts every time they click the incorrect button, leading them to the answer each time.
These factors make for a sketchy usability testing session, whereby the participant is unable to experience the service as they would if they were at home on their own. It gives the research and team less confidence in the findings.
As a core research requirement for usability I always prefer to give the participant a link to the prototype that will work on any device they are using, at any resolution, with any additional software tools they have installed.
I made the case to the consultancy designers that they needed to create these prototypes in Html, using the Gov.uk prototype kit; however as they had not done this before as most hadn’t worked as interaction designers on X-gov projects.
If you know what good looks like – don’t stand still – move things forward
Honestly, I was really frustrated at the unavailability of a proper prototype. I decided to take two days out of my role as a researcher and build the first end-to-end journey using the Gov.uk prototype kit. We tested with this version for two weeks, while NHS Digital quickly redeployed interaction designers from other national programmes to continue the design work and train consultancy staff in best practices.
In the end, the outcome was positive and the researchers have had good quality prototypes to work with months now; although at the time it was slightly awkward to need to step over an imaginary UX “role boundary” just to help speed up an inevitable change. Obviously it caused some challenges and debate within the team at the time, but from that point onwards the many benefits of being able to design with version controls, test on any device, conduct unmoderated testing, have hopefully far outweighed the initial stir within the team.
Map of the Testing world
Another new concept I introduced within the team was the end-to-end interaction design service flow. From the prototype I created, I took screenshots of every page and link these all up as a flow in Miro (Yes, it took a while).
[image goes here]
The primary reason at the time was to help capture notes from multiple user research sessions in the same place, so that we could identify which screens were most problematic by ‘virtually affinity sorting’ the findings.
This approach was then adopted by the service designers and interaction designers, who continued to maintain up to date maps of the current, the past and the future design iterations week on week.
The resource has been great for user researchers, designers, product, clinical and governance colleagues, but more generally anyone who recently joins the team they can instantly view what the end-to-end service looks like all in one place.