Consulting · Consumer app · India · Completed 2022
Teaching people worked better than pushing them.
The short version
Khel Group had a large user base and an expensive acquisition problem: getting somebody to install an app is one cost, and getting them to become an active, retained user is another entirely. We mapped the journey from first brand contact through to first game, benchmarked it against apps inside and outside the category, and found something simple. The app asked new users nothing about themselves and then treated all of them identically. We replaced that with five questions and a set of journeys that branched off the answers — and separately, began routing players who were struggling towards instruction instead of towards another game.
The journey was built for the business, not the user
We walked the whole thing: first brand interaction, app download, app launch, first interaction, first game. Then we benchmarked it — not only against other apps in the category, which mostly shared the same faults, but against consumer apps in unrelated sectors that are good at this.
The finding was blunt. The journey was abrupt and impersonal, and the in-app communication was almost entirely transactional. Every prompt existed to move the user one step closer to a commercial action. Nothing existed to find out who they were or to help them get better at the thing they had come to do.
An onboarding flow that only pushes is optimising the first day at the cost of every day after it. It converts the users who were going to convert anyway and loses the ones who needed a reason to stay.
Which is a design failure before it is anything else. A new user who has never played the game, a lapsed player returning after two years and an experienced player switching platforms need three completely different first sessions. The app was giving them one.
Five questions
The replacement was an onboarding survey, deliberately short, asked before anything else happened.
Every answer branched. Somebody who had never played was taken somewhere different from somebody who had played for a decade elsewhere. Somebody who said they played free games was not immediately pushed towards cash ones.
Implementation ran on CleverTap, which is what made the branching maintainable — journeys defined once and triggered by attribute rather than hard-coded into the app.
Five questions is a cost. Every field in an onboarding form loses users. The reason it paid here is that the answers were used — a survey that collects preferences and then serves everybody the same experience is worse than no survey, because you have spent the goodwill and given nothing back.
Sending a losing player to learn
The second piece is the one we would point to.
We built journeys triggered by how a user’s recent games had gone. Somebody losing repeatedly, or making the kind of errors that indicate they have not understood the rules, was routed to how-to videos and to the platform’s Rummy Academy.
The conventional instinct at that moment is the opposite — a struggling user is disengaging, so you push harder. Offers, prompts, another round. That earns something today and it produces a user who leaves shortly afterwards having enjoyed none of it.
Somebody losing because they do not understand the game does not need encouragement. They need to be taught, and the platform already had the material to do it. Sending them to it improved both retention and average gameplay — which is to say the responsible option and the commercial one turned out to be the same option.
That is not always true, and it is worth being honest that we did not have to choose here. When those two things do diverge, a case study is not where you find out what a firm decides.
What it produced
How to read this
Two points, not forty. Conversion moved from 5% to 7%. That is a 40% relative improvement and a two-point absolute one, and both descriptions are accurate. We show both because quoting only the 40% would make a modest absolute movement sound like a transformation.
Retention and gameplay are directional. The client reported both as improved and we were not given figures. We have left them as described rather than inventing precision.
The figures are the client’s. They came from the platform’s own analytics and CleverTap implementation, reported to us during the engagement, and were not audited by us.
We cannot separate the two changes. The onboarding survey and the performance-triggered education journeys shipped in the same programme. Which of them produced the conversion movement, and which the retention movement, we do not know.
Where this category stands now
This work was completed in April 2022, when real-money skill gaming was lawful in India. India’s Promotion and Regulation of Online Gaming Act 2025 came into force on 1 May 2026 and prohibits online money games, including real-money rummy, notwithstanding the long line of judicial decisions that had treated it as a game of skill. Operators were required to cease money-based play or convert to non-monetary formats.
This page is published as a record of past work, not as an offer of services in a category that no longer operates in India. The transferable part of it — journey mapping, onboarding personalisation and triggered content — applies to any consumer application, and that is the reason it is still here.
Constraints that shaped the work
We shipped both changes together, which was right for the client and wrong for the evidence. A staged rollout would have told us which mechanism did what, and on a platform with this much traffic the test would have taken weeks rather than months. We did not propose one, and the result is a page that reports an outcome without being able to attribute it.
The onboarding survey also has a cost we did not measure. Every question in a signup flow loses some proportion of users before they finish it. We know the net effect was positive because conversion rose. We do not know what the survey cost at the top of the funnel, and a business deciding how many questions to ask would want that number rather than our reassurance.
If your onboarding converts badly
Check whether your app asks a new user anything at all before it starts asking them for something. Most do not, and then serve a single journey to people arriving with completely different levels of knowledge and completely different intentions.
And look at what you send someone who is failing at your product. If the answer is another offer, you are optimising the wrong week. The users who need help and get it are worth more than the ones who need help and get a discount.
Read another
Every figure sourced from the client’s own systems, with the basis stated and the exclusions named.Advik Hi-Tech
The CAI Store
Fine Jewellery Manufacturing
Measurement note. Scope, journey design and implementation approach are from our own project records. Conversion, retention and gameplay figures were reported to us by the client from the platform’s own analytics during the engagement and were not independently audited by us. Platform user numbers are the company’s own published figure. Regulatory statements reflect Indian law as at the date of publication.
