In prediction markets like Probo, retention is related to 2 major aspects:
Product knowledge
Engagement
Predicting outcome of real-world events was a new way of trading for the people of India, especially in a Tier 2 & Tier 3 demography which is not used to the conventional stock trading or mutual funds investment.
Our previous onboarding flows had not kept up with the pace with which the product had grown over the last 2 years. In the last quarter of 2024, as Probo was getting ready for the surge during 2025 IPL season, the marketing this time was planned for TV commercials, that would mean we will have an unprecedented number of users hearing Probo for the very first time, and need a sort of short course on what prediction market is all about.
So we set out on designing an onboarding experience.
Since we had onboarding (of some sort), I started with UX Audit, right at the top of the funnel, starting form Probo website where the user downloads the APK.
How the onboarding (sort of) looked before redesign:
After the audit, I wondered if the users understood:
How is price and probability related on a prediction market?
How does the exchange works?
How are markets resolved?
To that end, I took help of the community team, floating a survey for New Users on mobile app. Here were the results:
There were 2 ways to solve for the education gap:
Show an animated version of the core Trading flow [called Product Demo]
Or, simulate an interactive environment for the new user to experience trading risk free [named as Interactive Video OB]
Based on my interactions with users, I was fairly confident that Interactive Video based onboarding would work the best, because that's how users who retain on platform for longer seem to have in common. They watch some YT videos or have a friend explain it to them face to face.
Both the ideas seem to hold potential and since Probo was close to the high-volume season of the year (IPL League), we decided that Solution 1 would be tested first as that was faster to ship.
We experimented with what converted the user right from the first intro page on the app to the OTP submit screen after installing the app. There was a drop in users who entered their phone number but did not submit OTP.
Our hypothesis was that our website looked very different than the first few screens of the mobile app.
We experimented with 3 versions to see which leads to user converting more:
Current carousel based light UI version with phone input
Graphic showing trust markers for Probo
Social proofing
Turns out, the Variant 2 outperformed in converting more users to sign up. Later, we updated the same graphic for our Probo website as well.
Next, we created a product demo flow.
We chose markets that were already settled so that users understand how probability and prices are correlated. For example, India won the T20 World Cup, so choosing 'Yes' on that market prediction card will lead to wins for the user.
We also created animated GIFs to explain how contracts are matched, markets are settled, and finally, where the winnings are credited.
This resulted in a flow shown in the following image.
This onboarding flow was made live in a phased manner, starting with 5% of New Users, subsequently making it for 100% base over a period of 4 weeks. The following results held true:
This onboarding flow proved that education was indeed missing from our new user journey. With 65% of new users going through the flow and correlating increase in deposits on same day, well, that just showed that those who understood the platform's workings, were depositing money.
Deposits are important, because that directly impacts the retention of users on our platform. The data showed that new users (age on Probo is less than 7 days) who deposit in their first week of sign up are 50% more likely to stay on Probo.
What remained to be experimented next was:
How to push the onboarding completion rate even more?
Trading on opinions is still nuanced, there is a good chance that the animated Lotties in this onboarding were too fast or didn't solve doubts of the users completely. Could we improve on that?
We answered these question with another onboarding experiment, called Interactive Video Based Onboarding.