Seasonal mix-shift: separating holiday organic uplift from paid efficiency
Holiday organic uplift gets credited to paid under careless readouts. A synthetic lab holds paid spend fixed and varies organic installs using AppsFlyer definitions.
When organic installs jump in a holiday week, blended CPI and ROAS look better even if paid efficiency is flat.
Define the calculation before using it
Enter paid installs, paid spend, baseline organic installs and holiday organic installs. Compute paid CPI and blended CPI before vs after the organic spike. Synthetic inputs.
Work through the synthetic example
| Input | Synthetic value |
|---|---|
| Paid installs | 8,000 |
| Paid spend | 40,000 |
| Baseline organic | 2,000 |
| Holiday organic | 5,000 |
| Output | Baseline week | Holiday week |
|---|---|---|
| Paid CPI | 5.0000 | 5.0000 |
| Blended CPI (spend / all installs) | 4.0000 | 3.0769 |
| Organic share | 0.2000 | 0.3846 |
Paid CPI unchanged; blended CPI flatters the holiday. AppsFlyer: organic = no media source (article).
Facts versus limits
| Field | Evidence or limit |
|---|---|
| Organic definition | AppsFlyer primary |
| Adjust attribution page | Cross-check only |
| Causal holiday uplift | Not proven by arithmetic |
