What this checkpoint covers
The questions draw on the average order value levers guide, the break-even ROAS glossary entry and the free shipping threshold test design. They test decisions: which number to judge a change by, where a lever belongs and what to do when a test goes off course. Three questions add the arithmetic: working out a break-even ROAS from a margin, checking a reported ROAS against it and reading a threshold test by contribution per visitor.
Task
Write a short answer to each question, with your reason, before reading on; for the questions with figures, show your working. Count an answer as right only when your reason matches the explanation. For each miss, note what in the scenario should have led you to the right answer, then reread the linked section.
Questions
- A gift shop raised its free delivery threshold and its average order value (AOVAverage revenue per order.) went up. The owner wants to call it a success. What else must you know before agreeing?
- A store sets its free delivery threshold a little above its average order value. Few shoppers reach it and margin falls. What did it look at, and what should it have looked at?
- A skincare store's new three-step bundle sells only after the discount is raised well beyond what the team planned. What does that tell you?
- A checkout pop-up offering an upgrade raised the value of completed orders, but fewer orders completed. Where should the offer go instead?
- Two product lines report the same return on ad spend. One carries a much thinner margin than the other. Are both campaigns equally healthy?
- A store worked out its break-even ROAS (Break-even ROASMinimum ROAS to avoid a loss.) from product cost alone. Returns in one category are frequent and postage is free over a threshold. What is wrong with its floor?
- A team running a threshold A/B TestControlled comparison of two variants. sees average order value reach significance after a few days, while contribution margin per visitor has not. The manager wants to stop and adopt the new threshold. What do you advise?
- Halfway through a threshold test, the store launches a site-wide sale. What happens to the test, and what should the team do?
The last three questions use invented figures.
Questions nine to eleven: a candle shop and a threshold test
- A candle shop sells a gift set for £40. Its cost of goods is £14, it pays £4 for delivery and £1 in payment fees, and returns cost it an average of £3 per order. Ignore tax. What is its margin on the order after all of that, and what is its break-even ROAS? What floor would it have set from product cost alone?
- The shop's ad platform reports a ROAS of 2.0 on £1,000 of spend on the gift set. Does the campaign clear break-even, and what did the attributed sales do to profit?
- A threshold test ran to its planned length with 10,000 visitors in each arm. Control: 400 orders, average order value £50, contribution of £15 per order after cost of goods, delivery, fees and returns. Challenger: 340 orders, average order value £58, contribution of £17 per order. The manager points to the 16 percent rise in average order value. What is contribution per visitor in each arm, and what do you advise?
Answers explained
- Whether profit per visitor rose. A higher threshold raises the average among those who buy almost by definition, but it can cost orders and absorbed postage. Check conversion and contribution per visitor alongside the average, and look at the next order too (measuring order value against margin and conversion).
- It used the average instead of the distribution. The average hides where baskets cluster. Set the threshold just above a large cluster so most shoppers can reach it by adding one item (delivery thresholds).
- The fit is weak. A bundle of products genuinely used together sells on saved effort with a modest saving. Needing a deep discount means the price is doing the persuading; fix the fit or drop the bundle (bundles, kits and cross-sells).
- On the product page or after purchase. Interruptions at payment cost more orders than they add in value. A tier comparison on the product page or an offer on the confirmation page carries no such risk (upsells and volume pricing).
- No. Break-even comes from margin alone, so the thin-margin line has a higher floor and may lose money on every sale at the same reported return. Set and read targets per line (how to interpret break-even ROAS).
- The floor is too low. Margin should be counted after delivery, payment fees and expected returns. Leaving them out makes losing campaigns look profitable, and the category with frequent returns needs its own, higher floor (what moves break-even ROAS).
- Keep the planned metric and sample. Average order value answers an easier question than the business is asking. Run whole weekly cycles to the planned sample, ideally past the return window, and judge on contribution margin per visitor (sample size reasoning).
- The sale swamps the threshold effect. Discounts move basket values in both arms and push many baskets over either threshold. Pause the test during the sale or extend it and exclude the sale period, and schedule future tests away from promotions (pitfalls).
Questions nine to eleven worked through
- A 45 percent margin and a break-even ROAS of about 2.2. Contribution on the order is £40 − £14 − £4 − £1 − £3 = £18, which is 45 percent of £40, so break-even ROAS is 1 ÷ 0.45 = 2.22. From product cost alone the margin would look like (£40 − £14) ÷ £40 = 65 percent and the floor 1 ÷ 0.65 = 1.54, so leaving out delivery, fees and returns understates the floor by about 0.7 (how break-even ROAS is calculated).
- No: it loses about £100. A ROAS of 2.0 on £1,000 of spend is £2,000 of attributed revenue. At a 45 percent margin that is £900 of contribution against £1,000 of spend, a loss of £100 on the attributed sales, even though 2.0 would have looked comfortable against the cost-only floor of 1.54. To break even the campaign needs about £2,220 of revenue for every £1,000 spent (1,000 ÷ 0.45) (reading break-even ROAS in context).
- Contribution per visitor fell, so do not adopt on order value alone. Control: 400 × £15 = £6,000, which over 10,000 visitors is £0.60 per visitor. Challenger: 340 × £17 = £5,780, or £0.578 per visitor. Average order value did rise 16 percent (£58 ÷ £50) and contribution per order rose 13 percent (£17 ÷ £15), but conversion fell from 4.0 to 3.4 percent, which is 15 percent fewer orders, and that outweighs the larger baskets: contribution per visitor is about 4 percent lower. A gap this small may be within the test's normal spread, so the advice is to keep the control and not to count the larger baskets as a win, rather than to call the challenger a proven loser (reading the results).
Where to go next
If the lever or break-even questions caught you out, redo raising average order value for a coffee roaster. Then move on to the third module, on retention and lifetime value, which builds towards designing a repeat purchase programme for a spice merchant, or return to the intermediate e-commerce level.
