< Back

Exploration

Here is the flow for first-time users before we have implemented any changes. When users open the app for the first time, they have to pick the language they want to learn and their proficiency level. After that, they must register to proceed. Straight after registration, we present them the first up sale. If they dismiss the pop-up, they will begin their first lesson automatically. Since that moment, we gave them access to restricted content. Which means they can learn new words and phrases until they hit the paywall. After they reach the limit of free content, they can no longer continue learning, only reviewing what they already have learned.  

With the data team, we gathered some information about retention in the early funnel. We observe a significant drop around selecting the language and proficiency level, and the first three learning sessions start. Except quantitative data we have no proof of what is causing those problems. We have some hypotheses but no deep understanding of users' behaviour. Naturally, my first step was to investigate and collect qualitative data to understand our learners and their perception of the product.

Hypothesis:

✦ Users can’t listen to audio.

✦ Users don’t know how to start a lesson.

✦ Users don’t like our content.

We collected a good amount of feedback that was touching different areas of the product. To analyse the findings, we have run an affinity mapping session. This method allows us to understand users better and prioritize the next steps. Those sessions are most effective when run as a group, for two reasons. It prevents biased opinions and allows a team to move faster.

We learned that participants enjoy our learning methodology. Videos of native speakers and a variety of test types were very engaging for learners. That information is a big win as this experience is the heart of our product ;)

But more importantly, we found that our users are not always happy with the level of difficulty that was assigned to them. Participants weren't sure how much they could learn with the course they have picked. The benefits of the subscription weren't clear. We had noticed some issues with understanding the learning methodology and usability of the app. Loads of people expressed general confusion. They were asking what is coming next, how much they have to do. They weren't feeling ready to start the learning session. In summary, users expect the product to be simple so they can focus on the learning process.

Prioritisation

After defining major pain points, I came up with some assumptions. I have run a workshop with my team (project manager, developers, QA and data scientist). Together we came up with hypotheses and general solutions to the problems I have identified. Choose the most impactful and low effort solution for the problem. This exercise helped me to prioritise and narrow down my focus and, at the same time, allow me to include the team in the design process and keep everyone in the loop.

Key Hypothesis:

Users are starting lessons without any context and explanation of what will be happening next.

The interface is complex and overwhelming, and users don't know how to interact with it initially.

Solution

My next step was to start design exploration to find the best answer for users' needs. I have decided to create a unique experience for first-time users to lead them and introduce the product. The new onboarding experience presented a shortlist of tasks. While completing the tasks, users will explore Memrise learning methodology, the richness of content.

At this point, I explored the order and number of tasks. I paid attention to copy to make sure that each step is understandable and makes logical sense as a whole. I have tested my solutions internally and externally by using the guerrilla usability testing method.

When I established the correct content display for users, I explored visual directions for a suitable interface. My main goal was to create a well-performing interface that will give users a feeling of progress. 

To find the best performing UI. I built high-fidelity prototypes and test them with language learners. We have focused our attention on usability performance and the confidence of users.

To analyse prototype performance, we used an assumption matrix which speeds up the process by 100%.

We have interpreted sentimental feedback during a quick infinity mapping session which points out the main trends and users feelings.

Challenges

During the process, I consulted my designs with language specialists and developers. For the MVP version, I had to compromise UI, task order, and variety. I was comfortable continuing the work despite the obstacles. The past research proved that a simple interface and hand lead experience would improve users journey and progress in the early stages.

A/B testing

During the handover to developers, I took part in writing tickets and participated in refinement sessions. We ran A/B testing on both platforms (Android and IOS) for new users.

We observed the rise of the first three lessons completion and remaining neutral for conversion. Here is what we find out:

Looking at metrics, we noticed that we had moved the second and third learning session start significantly. That reassures the correctness of our solution. However, we haven't increased the first and fourth lesson start. That last metric might mean that there is no smooth transition from onboarding to regular experience. And that metric becomes our main focus for the next steps.

This project leads us to an evident focus for the upcoming work. We have started working on a new screen in the product that informs users about their everyday learning journey and provides hand leded experience. The screen provides users information about their progress, but most importantly, it allows them to start their daily learning with a little bit of context on what tasks they will be working on. 

The improvements we made were successful thanks to the initial testing of the current product and a deep understanding of users behaviour. It made me understand that I could follow the business's requirements and improve users' experience simultaneously. With confidence, I can say that the key to success was the narrow focus and iterative approach. If I try to improve different areas of the product in one go, it would take a massive amount of time and won't be possible to understand its impact. The following steps might be unclear, and what often happens in moments of confusion projects could be simply abundant.