How the Lydia Design Team Gains Autonomy, Moves Faster, and Improves Conversion with Data-Driven Design

Insight/action/outcome: With Lydia’s loan calculator, they discovered 20% of users left before even seeing an offer. The team then redesigned the app to improve the UX. The result was a 14 percentage point increase in clickthrough rate.

I’ve always loved building things, from when I was a kid using Legos to discovering web development and design when I was 15 years old. Free courses on the internet were my starting point. I taught myself the basics to build my first web product. It was not beautiful or easy to use—in fact, it was a really bad product—but I was hooked.

I read books and articles to learn everything I could to create my own apps, like password manager Sesame. Five years ago I launched a podcast called Parlons Design, a space where I share the things I’ve learned with an audience of French product designers.

In the almost 10 years since I first learned about web development, I’ve found my approach to digital product design, which is to always try new things and iterate as quickly as possible. This approach is hard, though, without a way to measure success. Data is the only solution.

People think of design and analytics as mutually exclusive, but they’re very much intertwined.

People often think about design and analytics being in separate buckets, but they’re very much intertwined. You have to identify your objectives when you start a project, otherwise, it’s hard to know whether your solution is good or not. And if you want to grow, you similarly have to identify the markers of your success to then measure them. Without data, that process can become miserable. You could try user testing, but that’s time-consuming and very difficult in the beginning with a small audience. Even with a large enough user base, user testing often only catches the big pain points and misses the smaller details.

Every designer needs a tool where they can access data in real time to receive rapid feedback on design changes.

Amplitude Has the Granularity to Spot Friction and Opportunities

I was introduced to Amplitude during my six-month internship on the design team at French banking app startup Lydia. Every designer at Lydia has complete access to Amplitude, and I dove into the data from the moment I got there. I hadn’t used Amplitude before, but it was easy to learn, and Lydia also offered further training. I’d worked with analytics apps at previous internships, but nothing else I’d seen offered such an in-depth overview of data.

Every designer needs a tool where they can access data in real time to receive rapid feedback on design changes.

Lydia uses monitoring tools like Metabase to track operational activities like how many card payments are made every day. But these tools aren’t as useful for the product design team to understand in-app user behavior. For that, Amplitude was vital. The design team’s main long-term challenge is to identify the parameters, set targets, and measure the impact of particular features. We often want to revamp a product to better reach its target audience. Amplitude provides the granularity designers need to dive into a product’s events and determine what we need to change to achieve the outcome we want.

Amplitude helps the design team identify friction points as well as opportunities. I think the two are complementary. For example, Lydia offers a series of cards for different monthly banking plans suited to customer needs, including a free trial card. We monitor several steps in the card ordering process, and when we start to think about how we can improve this process, we look at our funnel. In a few minutes we can detect whether we’re losing users on a specific step. We then ask ourselves why this step causes friction, which we can answer by building a qualitative user test or, more often, funnels that will deliver data to describe the problem.

It’s the same for opportunities. By measuring who uses which features, we can determine the features key to making Lydia more successful. If only 5% of users discover a feature, but those users engage with that feature many times a week, that tells us it’s valuable and we should do more to promote it to other users. Similarly, if we notice that a feature has a high conversion rate with a specific user profile, we can target specific users for that feature or restructure the product to better match the expectations of our users.

How Amplitude Insights Translate to Design Solutions

While at Lydia, I witnessed a couple of instances where Amplitude insights improved product design. The first of these was our loan calculator. When we looked at the loan simulator flow, we found 20% of users left before they saw one of our offers. We asked pretty basic information of users: the loan amount, delivery method, and reimbursement period. So what about the experience was putting 20% of users off?

We identified three possible problems. One, the first screen dedicated solely to the loan amount was unnecessary due to other developments in the app. Two, the fields were ordered differently than what the user expected. And three, there were several dead ends that broke the user journey.

We decided to eliminate the first screen, reorder and reduce the number of fields from four to two (the loan amount and reimbursement duration), and reduce the dead ends by providing single-tap explanations for each item. The final step in our redesign was to set up a tracking plan which detailed every new event we would track to effectively compare our solutions to the “before” state. We provided this tracking plan to the data team, and then it was time to test our hypotheses in Amplitude.

Our results: for our main KPI, 94.4% of users clicked through to receive an offer—a 14-point improvement. We also learned something from Amplitude about these remaining users: A lot of them were relatively new customers, meaning we didn’t have enough information about them to make a loan offer. The iteration process never ends, and from this new analysis, Lydia can better target this feature for eligible users.

Another example relates to Lydia’s card customization page. Lydia allows customers to customize their credit card, a cool feature that not a lot of banks allow. People spend, on average, two minutes on the card customization page, which is quite a long time. But the design team also saw that many users dropped off when it came time to enter their desired customization.

It turns out that people spend a lot of time thinking about what they want to write on their card—after all, they’re going to see it for the next four or five years. But they’re not always sure what they want to say so they take some time to think about it, but they never return. We came up with two solutions to this issue. The first was to offer prompts to help users generate ideas. And for those customers who want the card but feel uninspired or ambivalent, we created an option to not customize their card at all.

Amplitude Gives Designers the Autonomy to Explore and Iterate Rapidly

When a company grows, it places increased demand on a data team. And of course, the demands extend far beyond product design. Queries come from across an organization, and the more people that can interpret data on their own using Amplitude, the more time a data team has to perform more complex analysis. Amplitude allows the design team at Lydia to be autonomous and data-driven, and move quickly based on insights about the product.

After seeing how Amplitude impacted our work at Lydia, I incorporated the platform into my own work as a freelance digital product designer. I had my own tracking system before, but it was complicated to make changes and improve events, so I couldn’t get new insights easily. It took a lot of configuration and a lot of analysis—too much time and effort for too few results. With Amplitude, I have an easy-to-use interface. I also leverage rich features such as cohorts and, within the funnel analysis, the conversion drivers, which is where I identify those friction and opportunity points.

Amplitude gives individuals and teams the autonomy to explore and make meaningful changes. When working by myself, I can dive into data more confidently and make the changes that matter. And when working as part of a team, we can get the answers we need without sending a data team yet another request. As a tracking tool, Amplitude is incredibly customizable and maneuverable. It’s easy to create dashboards and analyze specific data, allowing designers to receive insights quickly and iterate rapidly.

There aren’t a lot of companies where design teams have unfettered access to data. But gaining access to and exploring a platform like Amplitude is the only way to efficiently and effectively move forward and develop the products people like and want to use.