+15%
A 15% conversion uplift in completed applications.
£14m
What the 15% could be worth a year, if the lift held
+40%
Estimated by Zego’s Lead Product Designer: how much faster designers work
Industry
Insurance
Role
Design lead
Team
3 YLD designers, plus Zego’s team
Zego insures people who drive for a living, from Uber drivers to companies with hundreds of vans. It had grown fast, and its apps, journeys and parts had grown apart. I led two other YLD designers inside Zego’s own design team for six months. We built two things side by side. One was a north star, one app that works for every driver. The other was the design system to build it. The new quote flow converted 15% better in an early A/B test, and Zego planned the next three years around the north star.
Zego grew fast, and every new kind of driver added a path that could go nowhere
Zego’s growth looked great from the outside. It had deals with the big gig apps, and its cover grew from cars to scooters and vans. Inside, every new kind of vehicle or partner added another path to the quote flow, and some of those paths went nowhere. A fleet manager could get halfway through a quote and then be told to phone customer service. If you took the wrong path, you never spoke to the right team. That was frustrating for customers, and it cost Zego money.
The apps had grown apart too. Single drivers and fleets had separate apps, the app and the website looked different, and some drivers needed two apps for one policy. Zego’s teams kept rebuilding the same buttons and cards, and they drifted apart.
Cover for every kind of driver, and rules for every kind of cover
Zego’s customers range from drivers working for Uber, Deliveroo, Bolt and Ola, to fleets of 2 to 20 vehicles, and big fleets of 21 to 500 or more. Each drives something different, and each needs a different route to a price.
Every route had to follow legal, technical and business rules. Some fleets couldn’t get a quote online at all. And the team that knew all this best was too close to the product to see it whole.
I led two other YLD designers, working inside Zego’s own team
I was lead designer on the design system, its components and the UX of the prototype. James Blakey-Hevey and Manuel Ornelas helped build the tokens and components. We were three designers from YLD, including me. With Zego’s own designers, we were six. We sat with Zego’s designers and engineers every day, in shared Figma files. Their engineers could see each part taking shape and ask us about it straight away.
I ran four creative workshops, and a colleague and I presented the proof of concept that ended the first phase. We also asked Head of Design Benny Zuffolini the questions the team hadn’t had time to ask. Why do these flows split? What happens when a scooter driver becomes a van driver? Where does the experience break?
Every new product or partner multiplied the paths, and that kept coming back
We started with one-to-one interviews across the business, from customer service and data to engineering and leadership. Zego’s research team already had hundreds of findings, so we built on those.
My workshops turned what we’d heard into “How might we” questions, grouped into getting a quote, one app, fleets and Zego’s own tools. The team voted, and the ideas with the most votes became the north star.
We built the vision and the design system side by side, from week three
Most teams pick one or the other, the vision or the system. Zego needed both, and we had six months. So we started the design system in week three, next to the north star, and tested every part inside real screens.
Instead of adding another app for each new type of driver, the north star merges the driver app and the fleet app into one. It starts on the website and ends in that app, whoever you are and whatever you drive. A Driver Score can bring your price down when you drive well, and claims live in the same app.
Driver Score
Good
The new quote flow asks one question first, so nobody reaches a dead end
In the old flow, only single drivers could reach a quote form. Fleets were told to call. So the new flow opens with one question, how many vehicles need insurance, because the answer decides your route. Single drivers say what work they do and who they drive for. Then they see what drivers like them usually pay, before they’ve typed a single personal detail. Every answer after that sharpens the price, and you scan your licence instead of typing it out. You can go back and change any answer, so a wrong turn still gets you to the right place.
Fleets of 2 to 20 get their own online quote. Some fleets break the rules, like vans worth over £40k, and can’t be quoted online. Instead of a dead end, the flow says so and offers to get in touch, so Zego keeps the lead. Fleets of 21 or more go straight to booking a call with sales.
The system was built for the vision, and for the people building it
We grouped every part by what it does, like inputs, feedback and navigation, so names weren’t a matter of opinion. Every button came from one of three master buttons. Change a master, and all 171 button variants change with it.
Paying is where trust matters most, so the payment buttons are each phone’s own. Where the brand matters most, it’s all Zego’s own, from the Driver Score gauge to icons for braking and cornering. Everyday parts like inputs, dates and menus came in web, iOS and Android versions. We swapped older screens for the new components early on, to find what we’d missed. By the end, the system had 121 components and 891 variants.
Not every call went my way. I made the number plate component yellow, the colour of a real UK plate on the back of a car, with the official font, so it was easy to recognise. That started a debate about whether white or Zego’s brand colours would be better, and I lost it. Later on, it moved to plain brand colours.
The quote flow converted 15% better, and Zego planned three years around the north star
In early A/B tests with Zego’s app team, the new quote flow converted at 3.22% against 2.8% for the old one, 322 conversions against 280. That’s 15% better, in line with our hypothesis, from about 10,000 people in each group. It was an early result, short of the usual 95% confidence, so I treat the £14m a year as what it could be worth if the lift held, using that year’s revenue. The dead ends that sent fleet customers to customer service were gone.
The new onboarding was the part that shipped while we were there, and Zego planned the next three years around the rest of the north star. Its Lead Product Designer, Kyrsten McCartney, said “there’s enough work here to keep us sustained for years”.
Kyrsten estimated her team was moving 40% faster with the design system.
After the north star, we joined the app team to put the design system in place and test the north star’s ideas in real use. We handed it over with training and design-along sessions, then gave a talk together at the Design Systems London meetup. Zego then moved us onto another project area, and many of its product team adopted the system.
What I learnt
1. A vision only works with a plan to build it
The north star showed what was possible, and the design system made it buildable. That’s why we shipped it in pieces, starting with onboarding.
2. Build buttons with settings, not variants
If I built the buttons today, icons would be on and off toggles and the four button styles would be variable modes. The same buttons would need 24 variants instead of 171.
