Accurx’s biggest ever product redesign
Compose Refresh: where it came from, why we did it and how it went
A few weeks ago, we hit the big red GO button* on Compose Refresh, the biggest product redesign Accurx has ever done. Months of design and development work, plenty of internal debates about designs and features, and testing with countless users had led up to this point.
But why did we do it? Our backlog of desired features grows by the day, so why did we prioritise a redesign?
*Side note: I’ve been informed by one of our product engineers that it’s not actually a big red button. “It’s a bit more complicated than that, Sam.” 🙄
Where it all began
In Accurx’s early days, one of the first features we released allowed a GP to send a text message to their patients. It looked something like this:
While it was never going to win any awards for design, it was simple and let people get stuff done quickly. But naturally, as time went on, we learnt lots of things about our users and their needs. The result? Lots of other features!
Back then, adding a new feature meant getting it into the product as quickly as possible so people could use it, rather than meticulously planning where it should go alongside our other features. As you can imagine, this approach only works in the short term. Before long, these as-and-when updates start to look messy and can get lost when navigating around Accurx.
Compose Refresh: a clean slate
To solve all these things, we embarked on Compose Refresh, our big spring-cleaning-like redesign of the compose screen.
At the time, we had all these great features that we’d built based on years of user research. And now, finally, we could step back and methodically plan how we can make these easier for people to use.
At Accurx, one of our core values is Kaizen (which means continuous improvement) and this was the perfect place to put that into action.
The three aims of Compose Refresh
- Re-introduce features that some users might not know about. 📍
- Plan ahead for where new features could live in the future. 🔮
- Spruce up the Accurx interface to help it stand out. 🎨
And so began the months of design, backend engineering work, internal reviews, redesigns, user research, more internal reviews, usability testing, more redesigns… but then finally the day came. We had something we were happy to roll out!
Time to roll out
We know how busy GP practices are, so our top priority was rolling out the new product design in a way that didn’t disrupt their workflow.
We decided to take a staged rollout approach, so we turned on the new design for 10% of GP practices. And then we waited. A few hours went by. Nothing was on fire. Our support team hadn’t blown up our Slack channel with live issues. No news is good news… right?
So we gradually ramped up over the next few days. 25% of GP practices… tumbleweed. 50%… a couple of comments came through to our support team, but nothing that indicated we needed to make major changes or pause our rollout plans.
With confidence on our side, the following Tuesday morning (we don’t do big rollouts on Mondays), we gathered round that big red ‘GO’ button to roll out to 100% of GP practices across the UK… and pressed it.
So, how’d it go?
After going live with our full rollout, feedback came in thick and fast. It wasn’t just the content of it that told a story but the volume also. The feedback came in suddenly, from lots of different sources. Some of it was positive, and helped us to find a couple of pesky bugs 🦟
Some of it was less positive, but constructive, and gave us food for thought. 🤔
For many of us in the team, it was our first time dealing with this quantity of feedback. Not only that but, from regularly spending time with our users, we were acutely aware that the change of Compose Refresh could have a real, stressful impact on them.
When it came to the negative feedback, it was hard not to feel disappointed. Needless to say, the last thing we wanted was to inadvertently make life harder for users through the changes we’d introduced. 😣
What we learned
Truth be told, we had a tricky couple of weeks after the rollout, responding and actioning people’s feedback. But now that we’ve had a chance to reflect, there’s some really important things that we’ve learned. Here are the main three:
- Take a breath and remember the mission.
The temptation to drop everything and make immediate changes in response to feedback was strong, and we definitely made a couple of snap decisions which, in hindsight, weren’t too clever. Fixing a bug that affected people with a slow computer (in other words, most of them)? Good choice. Making a user interface change based on a totally untested assumption about user behaviour? Not so much.
We decided to stop being so reactive and looked at the feedback we were getting as a whole, and focus on our mission and what we initially set out to achieve. After doing this, we were able to prioritise issues and identify some root causes, rather than making haphazard changes that didn’t fix the real problems.
2. Make informed decisions as a team.
All that snap decision-making was really a symptom of a bigger problem — we hadn’t thought about a good way of prioritising or making decisions in advance. When we started looking at quantitative data and following up on feedback to discover if our assumptions were right, we were much more confident in our decisions, and still felt like we were being responsive without rushing.
Next time we launch something like this, we’ll make sure we give ourselves the tools to make informed decisions quickly. That includes doing things like tracking the right metrics, agreeing on prioritisation frameworks like a severity scale for usability issues, and drawing on strengths from everyone in our team to ensure we’re all on board.
3. With great scale, comes great responsibility.
Accurx is a product that’s used by over 130,000 people a day, and that scale is a double-edged sword. Firstly, it means that if you’re listening to everyone, you’re going to hear a really diverse range of opinions. And sometimes that means accepting that you can’t please everyone all of the time.
On the flip side, if you’re testing something with a really small group of users, you might identify issues that seem almost insignificant. But when you scale that to your entire user base, or a whole day’s worth of usage, sometimes those insignificant issues can be a lot more noticeable.
We experienced both sides of this sword and it has definitely highlighted the importance of thinking about which aspects of this level of scale are at play — and how they can mislead you.
What’s next?
After going through all the feedback we received, we were pleased (and more than a little relieved!) to find that the more negative responses were actually a really small proportion of our total user base. Some of the team who have visited GP practices since we launched have received some great comments about how much slicker the new redesign is.
We’ve still got a final few changes to make to Compose Refresh. We’ll only be happy to move onto the next challenge once we’ve addressed any difficulties people are having with the redesign. We’ve taken the time to think about these remaining changes, and look at the data to understand the reasons why we need to make them. Lesson learned. ✅
As a team, we’ve learnt so much from Compose Refresh. Having taken forward the three learnings above, we’re already planning how they can shape any future product redesign, whenever that may be! 👀
Interested in joining Accurx and being part of a future project like Compose Refresh? Check out our careers page for the latest job openings!