Advertising Spend and a Misleading Revenue Peak
Brightline, a subscription-box company, reports the following results for the last four quarters.
| Quarter | Revenue ($k) | Ad spend ($k) |
|---|---|---|
| Q1 | 400 | 40 |
| Q2 | 460 | 60 |
| Q3 | 500 | 90 |
| Q4 | 520 | 120 |
The marketing director says: 'Q4 was our best quarter. Revenue was the highest of the year, so every additional ad dollar is paying off. We should double Q4's ad budget next year.' The finance lead notes that the company also changed its packaging in Q3 and ran a holiday promotion in Q4. No holdout or matched-market test was run. The board must decide whether to approve the larger budget.
1. Compute revenue per ad dollar for each quarter and the incremental revenue per incremental ad dollar between consecutive quarters. What pattern do the numbers show?
Read the worked answer
Revenue per ad dollar: Q1 400/40 = 10.0; Q2 460/60 = 7.67; Q3 500/90 = 5.56; Q4 520/120 = 4.33. The ratio falls every quarter. Incremental revenue per incremental ad dollar: Q2 (460−400)/(60−40) = 60/20 = 3.0; Q3 40/30 = 1.33; Q4 20/30 = 0.67. The marginal return is falling, and by Q4 an extra $1 of ad spend is associated with only $0.67 of extra revenue. The director's statement that 'every additional ad dollar is paying off' confuses the level of revenue with the return. Q4 has the highest revenue, but it also has the highest spend and the lowest average and marginal productivity of that spend.
2. Identify the logical flaw in the director's inference and explain why seasonality and reverse causation matter here.
Read the worked answer
The director infers from 'highest revenue' to 'advertising is working best.' That is a level-versus-trend error. Revenue has risen steadily all year, so Q4 being highest is expected even with no change in advertising. The relevant comparison is against a counterfactual: what would Q4 revenue have been at Q3's ad spend? The data cannot answer that. Seasonality is an obvious confound: Q4 may include holiday demand, so part of the revenue rise is not attributable to ads. Ad spend may also be responding to expected demand rather than causing it; managers may spend more when they expect a strong quarter. The declining marginal ratio is evidence against doubling the budget, not for it. Finally, revenue is not profit: the ad spend itself reduces the net return.
3. What analysis should the board request, and what decision rule should follow from it?
Read the worked answer
Request a controlled incrementality test, such as a geo holdout: in matched markets, hold ad spend at the prior level while increasing it in others, then compare incremental revenue net of ad cost. If a holdout is impossible, ask for a regression that includes seasonality, price, promotions, and prior-quarter demand, and report the marginal return with a confidence interval. The decision rule should be: double the budget only if the estimated marginal revenue per ad dollar exceeds $1 after accounting for seasonality and the cost of goods. On the current table, the trend points the other way: the marginal return fell from 3.0 to 0.67, so doubling Q4's spend is likely to lose money. The director's claim is not supported, and the burden is on the proposal to produce counterfactual evidence.