Rewards Optimization
Product: Consumer mobile travel app
Outcome: Scaling and monetizing a travel rewards program
Context
I owned Hopper’s cash back rewards program, Carrot Cash Back (CCB), a core growth lever designed to differentiate Hopper from other OTAs.
The program was intentionally:
Flexible: Earn on any product, redeem on any product
Fast: Rewards available 48 hours after booking
Simple: $1 earned = $1 redeemed
When I joined, CCB had launched on flights only and was broadly available to all users. The next challenge was to scale the program across verticals (hotels, cars, homes) — and make it economically sustainable.
Primary KPIs:
Unique conversion
Purchase frequency
Contribution profit / ROI
I owned the program end-to-end:
UX & merchandising (wallet, home screen, funnels)
Platform mechanics (earning, redemption, liabilities)
Business model & economics
Fraud monitoring and controls
The Problem
How might we scale a cash back program that meaningfully increases bookings while controlling cost and reaching profitability, across multiple verticals with different margins?
Goals
Establish a reliable measurement framework for incremental impact
Scale CCB across all major travel verticals
Maximize incremental bookings while reducing reward cost
Move the program toward profitability
Approach & Key Decisions
1. Build a rigorous experimentation framework
To understand whether CCB was truly additive, I designed a foundational experiment:
Treatment: Earn cash back across all eligible verticals
Control: No cash back
This allowed us to isolate causal impact on conversion, repeat bookings, revenue, and ROI.
Within treatment, we ran A/B and multivariate tests by vertical. To measure performance, I built a SQL-powered cohort dashboard tracking:
Conversion and repeat rate
Revenue and profit per transaction
ROI over time by install cohort
2. Scale CCB across verticals
Revamped cash back UX in the flights funnel
Built CCB support for hotels
Partnered with cars and homes PMs to integrate cash back into their products
Scaling required influencing other teams to prioritize loyalty work by sharing impact data, tooling, and resources.
3. Optimize reward economics by vertical
More rewards increased conversion — but with diminishing returns.
For each vertical, I:
Conducted competitive analysis of OTA reward rates
Modeled ROI scenarios based on margin forecasts
Designed experiments testing different cash back percentages (e.g., 2% vs 3% vs 5%)
This allowed us to select optimal reward levels by vertical rather than a one-size-fits-all approach.
4. Reduce cost through behavioral targeting
As margins tightened and competing incentives emerged, cost efficiency became critical.
Data showed many users would book on Hopper with or without cash back. We needed to focus spend on users whose behavior was actually influenced by CCB.
I led two key experiments:
CCB Opt-In: Users explicitly enrolled in the program to earn rewards, increasing intentionality and limiting liabilities to motivated users.
CCB Claiming: Users earned rewards but had to actively claim them before redemption, introducing a light gamification layer.
We compared opt-in, claiming, opt-in + claiming, and control cohorts across conversion, bookings, and cost.
Results
Up to 15% incremental bookings per install cohort
Opt-in model reduced reward cost by 65% while maintaining conversion impact
The program approached ROI-positive economics, even amid margin pressure
