Test checkout copy in a Mumbai retail app
Before the decision
Refunds rose after a fee display changed, while checkout starts increased from an unrelated campaign. The product team sees more checkout activity, while the finance team sees refund costs rising. The decision concerns the contribution earned from comparable app users.
Read the companion story: Two checkout pages and one misleading winner
The problem
Should the team choose Upfront fees or Current checkout for the next 12 months?
Decision owner: The business unit lead.
Conversion per visitor can be misleading if traffic quality changed.
Option economics
Scroll the table to see all figures.
| Option | Volume (paid baskets) | ₹ per unit | Fixed cost | Net contribution |
|---|---|---|---|---|
| Upfront fees | 97,000 | 54 | ₹38 lakh | ₹14.38 lakh |
| Current checkout | 1,01,000 | 51 | ₹39 lakh | ₹12.51 lakh |
Make the call
- Calculate net contribution for Upfront fees and Current checkout. Which assumption makes the comparison most fragile?
- If the leading option delivers 10% fewer paid baskets, does your recommendation change?
- What evidence would you ask for before the team makes an irreversible commitment?
Reveal the worked solution
Start by validating Upfront fees. It contributes ₹14.38 lakh after fixed costs and leads by ₹1.87 lakh. Treat that gap as the amount of unmodelled cost or lost contribution the option can absorb before the ranking reverses.
- Frame the decision
Specify the eligible population and the outcome before collecting data. Treat the figures as planning estimates. Use comparable groups to measure the effect before rollout.
- Calculate both options
Upfront fees: 97,000 paid baskets × ₹54 less ₹38 lakh in fixed costs gives ₹14.38 lakh. Current checkout: 1,01,000 paid baskets × ₹51 less ₹39 lakh in fixed costs gives ₹12.51 lakh.
- Stress-test the forecast
Upfront fees leads by ₹1.87 lakh in the base case. If Upfront fees delivers 10% fewer paid baskets (87,300) at the same margin and fixed cost, it contributes ₹9.14 lakh. Holding the other option constant, the stressed result falls behind. This isolates volume risk. Check whether both options would respond to the same change in demand.
- Set a decision threshold
Holding the other option’s result fixed, Upfront fees needs at least 93,538 paid baskets to match or beat it under the model. Randomise before checkout and measure net contribution per assigned user.
Conversion per visitor can be misleading if traffic quality changed. Compare the cost of a limited test with the loss the business would face if the full commitment failed. Ask the team to agree the decision criterion before collecting more evidence.
Check whether your answer names the trade-off, cites a number from the exhibit and specifies a reversal condition.
What if the whole market moves?
The worked solution tests a change to one option at a time. Here, both options gain or lose the same percentage of volume. Unit contribution and fixed costs stay constant.
| Volume scenario | Upfront fees | Current checkout | Higher contribution |
|---|---|---|---|
| Both lose 10% of volume | ₹9.14 lakh | ₹7.36 lakh | Upfront fees |
| Base assumptions | ₹14.38 lakh | ₹12.51 lakh | Upfront fees |
| Both gain 10% of volume | ₹19.62 lakh | ₹17.66 lakh | Upfront fees |
Explain which option you would take forward. Check whether capacity, price or customer mix would also change as the market moves.
Practise this case as an interview