← Every exhibit
What each measurement instrument can carry
The instrument is not judged by whether it produces a number. It is judged by the question that number can answer.
| Instrument | What it observes | Strongest defensible output | Invalid leap |
|---|---|---|---|
| Touchpoint attribution | Recorded exposures and conversions | Which touches are present in the observed path | Those touches caused the conversion |
| Marketing mix modeling | Aggregate outcomes, spend and controls over time | A scenario relationship under stated model assumptions | The coefficient is experimental lift |
| Holdout or lift test | Treated and untreated units under a designed intervention | Incremental outcome under the tested conditions | The result transfers unchanged to every channel |
| Calibrated portfolio view | Model scenarios plus experiments and operating bounds | A decision with a stated confidence and downside | One dashboard is the truth |
Swipe or scroll horizontally if the table is wider than your screen.
Reference & Evidence
Source: Author's synthesis of Gordon et al. (2019), Blake, Nosko & Tadelis (2015), Lewis & Rao (2015), and Johnson, Lewis & Nubbemeyer (2017).
Related exhibits
-
The calibration log
From the essay Marketing mix modeling is a calibration problem
-
Accumulation versus activity, indexed over twelve quarters
From the essay Why growth compounds, and activity doesn't.
-
Observed credit is not incremental lift
From the essay A marketing attribution model needs a counterfactual