FINN

For five years, I was Principal Product Designer at FINN, a German car subscription company that reached over 300 million € in annual recurring revenue and a valuation exceeding 1 billion € by June 2026.

I worked on the consumer shopping experience, from discovery and search to pricing and checkout, alongside retention and B2B work.

Working closely with product and engineering, I turned customer feedback, data and business priorities into design improvements and experiments. The examples below show what we tested, what changed and what we learnt.

Problem

The default car list reflected popular models and stock priorities, which didn't always match what customers wanted. Finding relevant options often meant using filters rather than browsing.

Result

We tested curated rows based on popular filter choices, alongside promotional content. Over five times as many users viewed the convertible page, previously the least-viewed car category, but overall conversion fell. We suspected the content encouraged browsing without helping customers choose, so we planned smaller follow-up tests.

Problem

Studio photography showed the cars clearly, but gave customers little sense of how they might fit into everyday life.

Result

We tested lifestyle imagery on a small selection of models and saw an increase in conversion. Wider rollout was paused until we could automate the production of suitable imagery across the full vehicle range.

"Tony is an exceptionally effective partner to product and engineering teams. He works comfortably with teams of any size, quickly understands complex challenges, and turns them into elegant, practical solutions - often at remarkable speed."

- Jens Kegelmann, VP Product, FINN

Problem

A single model could have over 40 available configurations, with differences in specification, colour and price. Comparing these options made it difficult for customers to find the right car.

Result

We tested a manufacturer-style configurator that matched customers’ chosen features to available stock. The first version had implementation issues and a substantial drop in conversion. Fixes across four versions improved performance, but conversion remained below the original version. Early customer feedback also suggested a preference for choosing an available car directly, so the test was stopped.

Problem

We wanted customers to be able to choose additional insurance cover when configuring their subscription, without making the journey harder to complete.

Result

We introduced insurance upgrades within checkout as an extra step. Revenue increased, with no negative impact on checkout conversion.

"I always appreciated that Tony never just accepted initial ideas. Instead, he would thoughtfully challenge proposed solutions and suggest great alternatives. He is also super quick to learn. As AI tools developed, he easily picked up Claude Design and Claude Code to improve our work."

- Iryna Lysenko, Design Lead, FINN

Problem

JobAuto lets employees subscribe to a car through salary sacrifice. Its calculator estimated their monthly cost and savings, but sat on an external platform, taking customers away from the subscription journey.

Result

We integrated the calculator into the subscription journey, allowing customers to estimate their monthly cost and savings without leaving the flow to checkout.

I used Claude to turn complex proposals into working prototypes, helping stakeholders compare options and understand how they worked.

Customers paid a separate 1,500 € service fee, either upfront or monthly. Paying upfront reduced the monthly subscription price rather than the service fee itself, making the saving difficult to understand. A clearer cost breakdown improved progression to checkout and conversion.

Data showed that customers chose the two payment options in roughly equal numbers, prompting a proposal to remove the upfront discount. I used Claude to prototype the options and compare total costs. The proposal modelled 5–9 million € in additional annualised gross revenue, depending on rollout scope. This is now in testing.

I also identified and pushed for improvements outside planned briefs, using design judgement to make the site clearer and easier to use. Improving the customer experience was part of my role, even where the benefit wasn’t directly reflected in conversion or revenue.

Problem

The homepage had so much copy that the make and model selectors were pushed down the page, making it harder for customers to start finding a car.

Result

I proposed a simpler layout with less copy, clearer vehicle selection and lifestyle imagery. The test increased progression from the homepage to the main call to action by approximately 9%, alongside increased use of the make and model filters.

Problem

Make and model were separate filters, with model selection disabled until customers chose a make. A long list of brands separated the two, making the next step easy to miss.

Result

I combined make and model into one filter, showing models directly beneath their brand. Following the redesign, filter adoption rose to four times its previous level.

Problem

Out-of-stock cars appeared as separate listings for every configuration, including previous model-year versions that were no longer in production and would not return to stock.

Result

I proposed grouping unavailable cars by model, with a single stock alert for each model. This was eventually implemented, simplifying stock reminders and freeing up space to promote available cars.

I also helped develop the brand across digital channels, creating skills to guide consistent AI design and turn simple photography ideas into detailed image prompts.