In one line
Incrementality is the sales or profit a lever adds compared with not using it. Attribution is how credit for sales is shared out across levers; last-click attribution gives each sale to the last ad the shopper clicked.
What it does
- The honest test is a counterfactual: the same period with the lever switched off. In the real world that means a holdout or geo experiment; in Countermove the engine can re-run the quarter without it, with the same random draws.
- Last-click reports credit ads close to the purchase, such as search and retail media, with sales that would often have happened anyway. Brand media is rarely clicked, so it gets little credit.
- Some levers pay later: brand media builds memory, listings open stores, content gathers reviews. A one-quarter view misses most of their return.
- Some levers borrow from later: a deal's lift is followed by a dip, so its return shrinks as you count more quarters.
Business and brand
Impact on the businessP&L, cash, share price
- Budgets that follow last-click reports drift towards levers that look efficient and away from those that build the brand.
- Experiments cost money and patience, but they are how you find out which spend is wasted.
- Always ask over what period a return was measured: the ranking of levers can flip.
Impact on the brandMemory, loyalty, pricing power
- Most of brand building's return arrives after the quarter it is spent in.
- Deals that pay this quarter can cost buyers' willingness to pay full price later.
- A brand that judges everything on short-term returns slowly starves its future demand.
Try it
Switch each lever off
The engine re-runs Brightwater's quarter without each lever in its plan, one at a time, and follows both versions for two years. Change the period you count and watch the ranking change; then compare a last-click report on retail media with the re-run.
Gross profit per €1, this quarter: Brightwater with each lever against without it
- Brand media (€1.31m)€0.54 per €1
- Activation (€0.56m)€0.88 per €1
- Retail media (€0.15m)€0.61 per €1
- Content and reviews (€0.08m)€0.00 per €1
- Price deals (€1.20m)€1.25 per €1
- Listing fees (€0.28m)€0.00 per €1
Spend in brackets is the first quarter's. Gross profit counts deals and listing fees back in, so every lever is judged before its own cost; above €1 it pays for itself.
Units Brightwater sold with the first quarter's brand media minus units without it, by quarter. Below zero is sales the lever borrowed from later.
| View | Sales credited | Per €1 |
|---|---|---|
| Spend | €0.15m | · |
| Last click | €0.46m | €3.07 |
| Re-run: the quarter | €0.17m | €1.15 |
| Re-run: two years | €0.26m | €1.75 |
Sales at the prices shoppers paid. Last click credits retail media with every online sale through a sponsored result (11% of Brightwater's search clicks), including shoppers who would have bought anyway. The re-run counts only the sales that would not have happened without it.
Notice: Counted over this quarter, price deals returns most (€1.25 of gross profit per €1) and brand media least (€0.54); content and reviews and listing fees show nothing yet, because they pay in later quarters. Switch to two years and the order changes. A last-click report credits retail media with €3.07 of sales per €1; the re-run finds €1.15.
Runs Countermove's market engine: the reference soft-drinks market, three brands, rivals holding their plans so only your change moves the result.
Sources
- Lewis & Rao (2015), Quarterly Journal of Economics: the unfavorable economics of measuring the returns to advertising
- Blake, Nosko & Tadelis (2015), Econometrica: consumer heterogeneity and paid search effectiveness
- Gordon, Zettelmeyer, Bhargava & Chapsky (2019), Marketing Science: a comparison of approaches to advertising measurement, evidence from big field experiments at Facebook