An app that understands the user and adapts accordingly creates a happier experience. At Anghami, the leading app for streaming music in the Middle East, we observe the music a user listens to and personalize the homepage accordingly. This is great for users who listened to a hundred songs since Anghami has enough information to make recommendations they will love. But what about a user who just signed up?
The golden first experience will be subpar, and initial recommendations will be solely based on trending commercial songs. That’s why Anghami now asks new users the kind of music they like. User onboarding is the process of asking preliminary information so that the user finds the app useful. In the case of Anghami, users’ answers are used to introduce music they will likely enjoy. Instead of blindly displaying music that we think is great, onboarding makes it possible to introduce tailored music from the very first moment.
Anghami’s first onboarding iteration
In Anghami’s first iteration, onboarding starts by inviting the user to personalize Anghami. The purpose here is setting the right context and focusing on personalization. In the second screen, a list of popular genres is displayed, and the user chooses their favorite genres. The third screen shows popular artists from the chosen genres, and the user follows their favorite artists. Notice that the interface refers to genres in a friendly tone: “what kind of music.”
Injecting Analytics
It’s typical to inject analytics into every new feature so we can analyze performance. The analytics services we use at Anghami range from pure server performance to user behavior. Amplitude is the mobile analytics service we use to track user behavior in the app. In the case of onboarding, we added Amplitude events to each step: when a new user signs up, when they choose genres, and when they choose artists. Choosing artists marks the completion of onboarding.
How much value does onboarding add to our app?
Once Anghami’s onboarding became live, it was crucial to quantify the value of onboarding. This is typically done by looking at retention, which is the number of days users keep coming back. If users who completed onboarding were more likely to keep using Anghami, then that’s really good news.
Amplitude has a nifty retention feature where specific audiences can be defined, and Amplitude analyzes the retention for each segment. This helps me as a product designer to know which features we should highlight because users will likely find value in that. It also helps the marketing team because it pinpoints the audience we want to target.
I created a cohort on Amplitude for users who completed onboarding, and compared that cohort to all users in the retention tab. It turns out that users who chose their favorite genres and artists are 47% more likely to keep using the app in the long run.
Naturally, we’re obsessed about long term user retention at Anghami, since it’s a direct indicator that users had a memorable to-be-repeated experience with Anghami. And since users who complete onboarding are very likely to stay, we need to make sure that most users indeed go through the onboarding process.
Are users completing onboarding?
To analyze the performance of onboarding, I look at funnels in Amplitude, which is essential when introducing new features. Funnels show the percentage of users completing onboarding, and where exactly users decide to skip onboarding. A set of events are specified, and the funnel displays how many users dropped off at each subsequent event. It turns out that onboarding was performing poorly, since many users opted to skip it altogether. With the insight Amplitude provided, we now have to make decisions.
How should we improve onboarding?
It’s understandable that not many people have the patience to give more details and skip onboarding. But the culprit is the first screen, which doesn’t give meaningful visual cues of what to expect, and a button that says “Personalize Anghami”, which is not obvious except for people working at Anghami. A screen that we thought would excite users caused them to skip altogether.
Instead, we could’ve started with the second screen immediately. There’s no need to have a closed door that needs a button to be opened. By having the door open, the user will find it obvious that this is of great benefit: there’s a selection of genres in front of them. Without reading any sentence and without thinking, users now understand what’s going on. They will actually want to choose their favorite genres.
The genres-artists conundrum
Another issue that came to our attention is not everyone who chose genres continued with the process of choosing artists. In fact, 15% of the users dropped at this point. We realize that some users have an affinity towards lesser-known artists who don’t show in a list of popular artists. But that doesn’t justify a 15% drop.
It turns out that while genres like pop, rock, hip-hop and alternative are common terms in the west, middle easterners don’t necessarily know the names of genres they listen to. So much so that we at Anghami created the names of many genres to begin with, like Lebanese Pop and Egyptian Classical. Most users don’t obviously know that Lebanese Pop refers to artists like Nancy Ajram, Elissa, and Carole Samaha. So many users either chose genres randomly, or they thought these genres referred to other artists. That’s why once they saw a list of artists, they didn’t choose any and decided to skip.
Based on this conclusion, we felt we wanted to omit genres altogether. But displaying a random list of artists is nonsensical, and it’s infeasible to display artists from all genres on one screen. But to make the genres more meaningful, each genre will be accompanied by a collage of artists that the user can relate to. Not only that, tapping a genre will immediately open a preloaded grid of relevant artists so that the user can make sure this is the genre they want and they can choose their favorite artists in that genre.
App forward
Anghami’s next update will introduce these changes based on what we learned in the first iteration.
By looking at user behavior in Amplitude, we were able to quantify the value of onboarding, and make conclusions based on the user drop-off rate on each step of onboarding. We expect the design changes based on these conclusions to significantly affect user behavior, and hence increase retention.
