What this checkpoint covers
Every question is set at Chainglow Bicycle Works, an invented bicycle retailer selling online and from one shop. The scenarios draw on the marginal returns and forecasting guide, the budget increase incrementality test and the marketing profit and loss guide.
Task
- Write an answer and a decision for every question before scrolling further, marked confident or unsure.
- Compare with the explanations and note the signal you misread on each miss.
- Reread the linked section for any question you got wrong while confident.
Questions
- Chainglow's brand search campaign has the lowest cost per acquisition of any channel, and the marketing manager proposes moving the paid social budget into it. What do you need to know before agreeing?
- Using the paid search history in the example below, at which spend level would you stop, and how sure are you?
- Chainglow wants to raise paid social sharply for the spring season. Outline the test you would run before committing the increase nationally.
- After a previous increase, the social platform reported a cost per acquisition comfortably under target, while a regional test showed a marginal cost per incremental sale above it. Which figure decides next quarter's budget?
- Finance reports a smaller marketing cost than Chainglow's own tracker. The agency invoices media and its fee on one bill, salaries sit in a central people budget and cycle-to-work discounts reduce revenue. How do you reconcile the views, and which return figure do you present?
- The owner asks how many extra bikes a budget increase of half again would sell. What do you give her?
- A cash squeeze forces a winter cut. The finance director fears losing sales in proportion to the average cost per sale. Is that fear well founded?
Chainglow Bicycle Works: paid search history
Target cost per sale from contribution margin: £45. Each row is a comparable month from the last two years, with sales counted in Chainglow's own order system rather than the ad platform.
| Monthly spend | Sales | Average cost per sale |
|---|---|---|
| £4,000 | 200 | £20.00 |
| £6,000 | 260 | £23.08 |
| £8,000 | 290 | £27.59 |
| £10,000 (current) | 300 | £33.33 |
Answers explained
- The marginal cost of brand search, not its average. Brand search reaches people already looking for Chainglow, so its first share of spend is cheap and its last share often buys sales that would have arrived through organic results. Check its marginal CPACost of each conversion. and impression share before moving money in; a saturated channel is the worst home for more budget however good its average. See the marginal returns and forecasting guide.
- Stop at the second level, read as a range. The steps cost progressively more per extra sale, and the third step is already well above target (worked below). Treat the stopping point as a band, because the steps were taken in different months.
- A geo budget-lift test. Match regions on sales history, assign test and control within pairs at random, set the affordability threshold first, make the lift large enough to see, count all-channel sales from Chainglow's own systems and run for a full consideration cycle, which for bicycles is long. Guardrails: blended cost per sale and branded search. The budget increase incrementality test sets out the design.
- The marginal cost per incremental sale. The platform credits sales it touched; the test shows what extra spend caused. Hold or reduce the increase, and record the IncrementalityConversions caused by marketing. result where future forecasts can use it.
- Map every cost once, with finance. Ask the agency to invoice media and fees separately so media is not inflated; agree whether salaries are in or out and keep that convention; show discounts as a cost of winning sales. Present return on investment (ROIProfit relative to cost.) on contribution after all marketing costs, beside the channel return on ad spend. The marketing profit and loss guide covers the reconciliation.
- Scenarios with ranges, not one number. Give outcomes for holding spend, a modest rise and the full rise, each built from estimated marginal costs that worsen as spend grows, and recommend a staged test before the large step. Multiplying the new budget by today's average overstates the gain.
- No. The last share of spend is the least productive, so a cut loses fewer sales than the average implies. Cut the steps with the worst marginal cost first and forecast the loss at their marginal rate.
Question two worked through
Step from £4,000 to £6,000: £2,000 for 60 extra sales, about £33 each, under the £45 target. Step from £6,000 to £8,000: £2,000 for 30 extra sales, about £67 each. Step from £8,000 to £10,000: £2,000 for 10 extra sales, £200 each. The stop point sits between £6,000 and roughly £7,000; cutting from £10,000 to £6,000 frees £4,000 and loses about 40 sales, each of which cost far more than it earned. A learner scoring six or seven moves on; four or five rereads the sections behind the misses.
Where to go next
If you missed questions one, two or seven, redo the tea merchant reallocation exercise. Otherwise move to the next module with the sofa brand measurement exercise, and keep the intermediate digital marketing level to hand.
