Set the strategy and sequence for modernizing checkout across every paid product.
Case study 03 · Commerce modernization
Modernizing checkout across cash, virtual currency, and an entire product catalog.
How I led the evolution of Tarot.com's desktop-oriented commerce system into a responsive, measurable platform—and built phased experimentation into the way checkout improved.
Business outcomes
13%Conversion
lift
lift
50+Digital
products
products
9Checkout
variants
variants
2Payment
currencies
currencies
Executive summary
I redesigned the commerce decision system behind every paid product—without interrupting live revenue.
Owned customer logic, experience direction, experimentation, prioritization, and phased adoption.
Connected cash, Karma Coins, balance, identity, payment, entitlement, and fulfillment states.
Improved the system while protecting transaction accuracy, customer access, and live revenue.
The core product logic
The best path depended on the customer's balance.
Every customer saw a path shaped by their balance, purchase history, VIP status, and the product itself—not the same default sequence for everyone. That's what the balance check was really deciding.
Purchase intentCustomer chooses to
unlock their reading
unlock their reading
Balance checkEnough Karma Coins?
Fulfill automatically
The system deducts the Karma Coins in the background and opens the purchased reading—no checkout required.
Reading revealedNoAdditional payment required Payment confirmed
Choose a payment path
Fulfill purchase
Reading revealedA/B testing
We tested which conversion lever made checkout easier to complete.
The experiment held the selected reading and payment methods constant while changing how customers evaluated the purchase: a simplified default, a visible value comparison, or a time-bound promotional offer.
Control
Reduce visible choice
A collapsed selector presented one Karma Coin package at a time. We tested whether less visible complexity would help customers complete checkout faster.
Variant A
Expose the value ladder
All packages appeared together for easier price and value comparison. We tested whether transparency would increase confidence and package value.
Variant B
Add a timely incentive
A contextual sale introduced visible savings without changing payment mechanics. We tested whether urgency would improve purchase completion.
Marketing activation system
The checkout became a self-serve growth platform.
Marketing moved from engineering tickets to same-day targeted promotions. Every idea used to compete for roadmap capacity now launches the moment it's ready.
Engineering-dependent
- Marketing request
- Engineering ticket
- Backlog prioritization
- Delayed launch
Marketing configured it directly
- Define the offer
ProductCurrencyDiscount
- Select the audience
PurchaserVIP statusLifetime spendProduct affinityAccount age
- Same day launch. No engineering ticket.
Scaling methodology
We turned controlled rollout into a repeatable experimentation capability.
New checkout experiences moved from one product to the broader portfolio only after customer behavior, commercial performance, and operational reliability were validated.
One productFull portfolio
Release narrowly
Start with one product or a controlled cohort with meaningful volume.
Measure behavior
Track conversion, payment path, customer state, price response, and failures.
Compare safely
Evaluate the new flow against the control and operational guardrails.
Scale the learning
Extend validated improvements across the paid-product portfolio.
Guardrails held constant
Revenue continuityPayment successFulfillment accuracyError rate
What I learned
“More ways to pay only create value when the product removes the decisions customers should not have to make.”
The strongest checkout was conditional: balance, product, device, and purchase history determined what the customer needed to see next.
Optimization had to include reporting, recovery, balance accuracy, and fulfillment—not simply the fields inside the payment form.