My role:

I joined Loycus as a UX designer to conduct a research process that enabled the company to pivot from a failing product. At that point, I was given the opportunity to start as a Product Manager to lead the development of the new product. Once launched, I assumed ownership first of the business area (B2B) and later expanded to oversee the apps (B2C).

Case Study: Increase user retention to 94%

Overview:

The Loycus system was initially based on the following process: A user connects their bank account, makes a purchase at one of our client's businesses, claims the purchase, we review their bank transactions, detect the purchase, and reward them accordingly.

When we launched the product, we ran a welcome campaign and a viralization campaign, which quickly boosted the number of active users on the app. However, we also noticed that user retention was not as high as expected. After a couple of days of high activity, users progressively decreased their engagement.

That's why I decided to prioritize improving retention metrics on the roadmap (along with other iterations of the MVP and pending small features). It’s important to note that we didn’t measure retention by app logins but by reward requests.

Discovery:

To better understand how to improve retention, I followed two approaches:

• Data analysis: Examining the behavior of users who were active.
• User interviews: Understanding why previously active users had stopped being active. This last point was carried out with the help of customer service (and overlooking privacy policies—StartUp style!).

Solution:

After analyzing the situation, we decided to work on each of the points. For points 2 and 3, proposals and developments were made in collaboration with the commercial team (2) and the technical team (3), but these are outside the scope of this case. Therefore, I will focus on points 1 and 4.

How can we prevent users from forgetting to request rewards?
The most obvious solution, and something I immediately knew we needed to work on, was using push notifications (activation of notifications is also outside the scope of this case study). But were we capable of implementing it at that time? The answer was no. The system was designed for users to request the reward on the same day as the purchase. Once the request was made, we accessed their banking history, and if we found the purchase, we granted the reward.

But this raised a different model: What if we reviewed all our users’ accounts daily and detected purchases from businesses in our database ourselves? This meant a change in the process (and scalability, as it implied a significant increase in resources). However, this solution would allow us to notify all our users of detected purchases.

Before proceeding with development, I proposed an experiment. We selected a segment of inactive users who had push notifications enabled. Each day, we checked their activity and sent notifications when we detected purchases eligible for rewards. The results were undoubtedly a success, so we decided to proceed with full development (what is summarized in one line here was actually riddled with challenges, which I’m skipping for this case study).

Improvement of the reward request flow Another issue we identified, both from user interviews and data analysis, was that our initial approach of using the homepage as a marketplace worked well for new users and discovery. However, when analyzing the flow of recurring users, we saw that they were accessing the app directly to request rewards. And we weren’t making it easy for them.

So, we decided to streamline this flow. After several proposals and tests, we introduced a floating button on the homepage. This button either directed users to their favorite businesses or displayed a list of businesses where they had previously requested rewards.

The best part was that, once the model change from the previous point was implemented, we could directly give users access to the businesses where we had detected purchases. (This was a game-changer!)

Outcomes:

Notifying users when we detected purchases in our businesses, combined with streamlining the process (along with addressing issues such as banking aggregation problems, onboarding new businesses, and running a notification activation campaign), allowed us to achieve over 90% recurrence throughout the year, with our peak reaching 94.4%.