In 2019, Patagonia stopped advertising on Facebook and Instagram not for an ethical whim, but because the data showed that reaching its target through those channels cost four times more than niche media like mountaineering magazines or outdoor podcasts. Behind this decision there is the calculation of the Affinity Index, a metric that measures how much a segment of population is over- or under-represented in a certain average than the general average.
Real problem
Many media plans confuse the cost per thousand impressions (CPM) with the cost per significant impression. This leads to investing budget on audiences who will hardly buy the product, wasting resources and damaging the brand perception.
Conceptual model
The Affinity Index is calculated as:
Affinity Index = (percentage of target in the audience of average / percentage of target in total population) × 100
An index of 100 indicates that the segment behaves as the average, above 100 indicates over-representation, below 100 under-representation.
Strict formalisation
| Average | % Target in Public | % Target in Population | Affinity Index | Interpretation |
|---|---|---|---|---|
| Alp magazine (alpinism) | 38% | 6% | 633 | Concentration 6 times higher |
| Trail Talk Podcast | 29% | 6% | 483 | Almost 5 times higher |
| Instagram (generic) | 7% | 6% | 117 | Close to average |
| TikTok (general) | 3% | 6% | 50 | Half the target compared to the average |
These numbers indicate that each impression in the specialized magazine is worth six times more in terms of the probability of reaching the target than in generic Instagram.
Example or case study
For an outdoor premium brand with 6% of the online population target, the nominal CPM is not enough to evaluate efficiency. The CPM in-target is calculated by dividing the nominal CPM by (Affinity Index / 100). Thus a low CPM on an average with Affinity Index 50 can be less efficient than a higher CPM on an average with Affinity Index 300.
| Average | CPM | Affinity Index | In-Target CPM |
|---|---|---|---|
| Generic TikTok | 2€ | 50 | 4.00€ |
| Programmatic display | 2.5€ | 70 | 3.57€ |
| Generic Instagram | 5€ | 117 | 4.27€ |
| Mountaineering magazine | 60€ | 633 | 9.48€ |
Lab / exercise
Basic level
Calculate the Affinity Index by two media of your choice using public or simulated data, comparing the percentage of target in the audience with that in the population.
Intermediate level
Use a customer database to segment the target and calculate the Affinity Index by product categories or interests, as shown in the SQL example.
Research-grade level
It integrates CRM data with customer match on Meta or Google platforms to obtain Affinity Index based on actual customers and not on demographic proxy.
Suggested datasets and materials
Anonymous CRM Dataset, Meta Audience Insights report, panel survey online.
Typical error to avoid
Planning the media mix based only on the nominal CPM without considering the Affinity Index leads to wasting budgets on unimportant audiences.
Quiz or checkpoint
- What does an Affinity Index indicate above 100?, Why is the CPM in-target a more reliable metric than the nominal CPM?, What are the main methods for calculating the Affinity Index?