Trade X had a problem most growth teams never solve cleanly: banned from the Google Play Store, with a ₹25 CAC ceiling that made pure ad scaling impossible. This is the full story of how we designed a referral engine that eventually started generating its own referrals.
Trade X was building something genuinely new in India: an opinion trading platform where users could participate in events — sports, politics, entertainment — and win or lose based on outcomes. The product was engaging. Users who understood the game came back daily.
The problem was getting them there. Trade X had been banned from the Google Play Store, eliminating the primary install channel for any Indian consumer app. Every user had to be acquired through paid ads, creator content, or direct APK file distribution.
Paid acquisition was working but had a hard ceiling: sub-₹25 per sign-up, sustained. As spend scaled, costs were rising, not falling. The team needed a new acquisition lever that didn’t depend on ad auctions.
Real Money Gaming referral programmes have a specific failure mode: the incentive to abuse them is extremely high. Users on gaming platforms think about expected value. A reward claimable without genuine intent will be exploited — fake accounts, self-referrals, device farms. We had seen this kill competitor programmes.
The additional constraint: Trade X couldn’t afford a cost-centre referral programme. Rewards needed to be economically self-sustaining — funded by the activity the referrals generated, not by a fixed budget.
Any mechanic rewarding users before they generated platform revenue would be abused or become unprofitable. The design had to solve acquisition, activation, and fraud prevention simultaneously — with one mechanic.
Each failure taught us something specific about where the design was breaking down.
| Mechanic | What Happened | Why It Failed | Verdict |
|---|---|---|---|
| Reward on signup only In-app currency on sign-up, no deposit required |
High referral volume. Numbers looked good immediately. | Almost none of the referred users deposited or traded. We were paying for empty accounts. | ✗ High volume, zero quality |
| Two-sided cash reward Withdrawable cash for both referrer and referee |
Fraud appeared within days. Device farms, fake accounts, self-referral rings. | Withdrawable cash on a gaming platform is an open invitation. Bad actors found the path immediately. | ✗ Fraud risk unacceptable |
| Reward too small Low-value reward to minimise cost |
Near-zero sharing behaviour from existing users. | Below the psychological threshold where sharing feels worth the social effort. People don’t refer friends for trivial amounts. | ✗ Insufficient motivation |
The mechanic needed to be meaningful enough to motivate sharing, impossible to abuse without real intent, and self-funded by the activity it generated. No off-the-shelf template solved all three. We had to design it from scratch.
The winning mechanic came from a reframe: instead of "how do we reward referrals?", we asked "what does the referred user need to do to become real — and can the reward make that happen automatically?"
Credited on sign-up but non-withdrawable and only usable in a trade event. To use it, the new user had to add real money and make their first trade. The reward forced the activation step.
The referrer earned 10% of the 2% platform fee on every trade by users they referred — ongoing, not one-time. The more their referrals traded, the more they earned.
The ₹20 non-withdrawable credit wasn’t a reward in the traditional sense — it was an onboarding mechanism disguised as a reward. The referred user didn’t feel pushed to deposit — they felt like they were using free money. Completely different psychological framing from a forced deposit requirement.
The referrer revenue share meant referrers had a financial reason to refer active traders, not just anyone. Quality was self-selected at the source.
To abuse the programme, a bad actor needed to create a fake account, add real money to the wallet, and make at least one trade. At that point, they had become a real user. The “fraud” converted into exactly the outcome we wanted.
No arbitrage existed. The cost of abuse was identical to the cost of genuine participation. Zero engineering resources spent on fraud detection.
Fraud dropped to near-zero without any detection infrastructure. What looked like a fraud problem was an incentive design problem. Solve the design, the fraud disappears.
A referral programme needs critical mass to become self-sustaining. We seeded it across two parallel channels.
Trade X had already built a network of 800+ YouTube and Telegram creators across gaming, finance, and sports. Their compensation was performance-based (Cost per Deposit, Revenue Share, or Cost per Sign-up). We briefed them differently depending on their audience — no single script:
Post-win: Surfaced the referral prompt immediately after a payout — the highest-motivation sharing moment. Post-deposit: Explained the programme to new users who had just added money, framed as "earn while you play." A weekly leaderboard added social status incentive alongside the financial one.
Most referral programmes plateau at a steady state. The Trade X programme crossed into second-order referrals: users who had been referred were themselves becoming referrers — generating further referrals with no prompt from us. The incentive structure was self-replicating.
Motivated by 10% revenue share on all future trades by referred users
Credit only usable in a trade event — activation forced by design
Fraud eliminated here — real money required before any reward realised
Referrer earns ongoing revenue share; referred user understands the platform
Same revenue share incentive kicks in — second-order referrals begin, programme sustains itself
The programme was economically self-funded — referrer rewards paid from revenue generated by referred users. Trade X went on to 4M+ sign-ups and 700K+ paid users, establishing itself as the category leader in opinion trading in India. Growth metrics contributed directly to the $5M fundraise.
The ₹20 credit worked because it forced the exact behaviour that makes a new user real: the first trade. The best mechanics make the activation step part of the reward experience itself — the new user isn’t completing onboarding to get a reward; they feel like they’re using a reward while completing onboarding.
The instinct when fraud appears is to build detection. The better solution is to make fraud structurally uneconomic. If the reward can only be realised by doing what a real user would do anyway, there is no exploit to find.
Most referral programmes are linear — you put fuel in, users come out. A programme with second-order referrals compounds. The difference: whether referred users have the same incentive to refer as the original referrers did. If yes, the loop closes and the programme sustains without constant fuel.
Seven actionable lessons delivered over 7 days. Free, no obligation. Built on the same frameworks used at Trade X and SaveSage.