Building a habit
Learning a language is all about consistency, even when using the most efficient vocabulary builder on the planet. This project aimed to expose users to Lingvist's core value loop quickly and set up a structure they would want to interact with every day.
ROLE:
Lead designer & product manager within a cross-functional squad for Lingvist, a language learning application.
The Problem
Maintaining retention for any goal-oriented app is tricky at the best of times.
More so when the goal that's trying to be reached is inherently hard, as in the case of language learning. To better set users up for success, we need to create an optimal cadence and a narrative to help them habituate.
THE PROCESS
Establishing the 'A-ha' moment
Using the Reforge methodology as a basis, we analysed existing data for users that have used the app regularly for at least a month to better understand when all users are most likely to be experiencing our core value, or the 'A-ha' moment.
User research on 'habituated' users
Taking the form of initial surveys followed by remote user interviews, we connected with users that had regularly used Lingvist over the course if the preceding 3 months and asked them about how they measured their success and what kept them motivated.
Initial wireframes and low-fidelity prototype
Using the data from both the quantitative and qualitative studies we conducted, early-stage designs were created and tested internally. Refinements were made according to the feedback given which ultimately led to the final design and a release candidate being built.
A/B test against the status quo
The new feature was initially released on one client (iOS) as part of a 50/50 split A/B test. The test was monitored closely over a course of 21 days with tweaks and minor iterations being made before the feature was greenlit for general release.
The Outcome
As a language learning app, Lingvist's core model is based around users answering a sequence of flash-cards that an intelligent algorithm sets up optimally through the use of spaced-repetition.
On the basis of our research findings, we reduced the number of cards to complete from 100 per day to 50 cards, four times a week, designing a new experience around this structure. This fifty card cadence had minimal impact on learning metrics whilst significantly increasing retention amongst users over the first week of use, and beyond.
This new experience also allowed us to inform users about their progress more incrementally whilst giving us a window to further motivate them through statistics, progress analysis, and streaks. As a bi-product, we were better able to manage our life-cycle messaging within the 50 card structure, reducing noise and ultimately, churn.