| Counting metrics (e.g. GRPs, Clicks) |
|
| Expert Opinion |
|
| Long-term Cohort Studies, Customer Lifetime Value |
|
| Market Mix Models |
|
| Multi-touch Attribution - Machine Learning ("New") |
|
| Multi-touch Attribution - Rules-based ("Classic") |
|
| Quasi-Experiments with Synthetic Controls |
|
| Experiments |
|
| Randomized A/B Tests ("King of the Hill") |
|
| Geo-Holdout Tests |
|
| ID level non-randomized |
|
| ID level non-randomized, but controlling for demos |
|
| ID level non-randomized, but controlling for demos and brand propensity |
|
| ID level non-randomized, but controlling for demos, brand propensity and media exposure |
|
| ID level non-randomized, but controlling for demos, brand propensity, media exposure, and using counterfactual modeling |
|
| Randomized Controlled Trials - Intent to Treat |
|
| Randomized Controlled Trials - PSAs or Ghost Ads |
|
| Weighted Index |
|