Paid, owned, earned: assigning a role before assigning a sale.
Key distinctions
Three conditions before calculating.
Distinct control
Paid controls the media buy, owned controls the asset, and earned controls neither the distribution nor the final message. That difference changes risk, speed and reproducibility.
Mechanism before metric
Define whether the lever creates reach, attention, consideration, conversion, loyalty or recommendation. An identical metric does not mean an identical mechanism.
Explicit interaction
Cross effects are estimated as interactions or sequences. Adding up platform attributions overstates the total when several families reach the same people.
Method
Building a measurable paid-owned-earned system
- 01Assign the roles
For each action, write population, mechanism, horizon and expected decision; refuse vague objectives such as visibility with no associated behaviour or measurement.
- 02Map the sequences
Connect paid exposure, owned visit, sign-up, CRM follow-up, external mention and conversion; identify the points where the same action is claimed several times.
- 03Choose the evidence
Use experiments for paid incrementality, journey tests for owned, time series or events for earned, then test their interactions.
- 04Allocate by contribution
Compare marginal cost, capacity, saturation and lead time of each family; earned receives no media cost but requires product, relationship, content and response.
The role — control — horizon — evidence matrix
Each family has a priority function, a limit of control and a compatible method of proof. A sale must not be credited three times.
| Family | Role | Control | Horizon | Evidence | Sales |
|---|---|---|---|---|---|
| Paid | Buy reach | High | Short | Experiment | 900 |
| Owned | Convert and retain | High | Medium | Journey test | 280 |
| Earned | Create external evidence | Low | Variable | Series / event | 120 |
| Interaction | Reinforce effectiveness | Shared | Sequence | Factorial design | 90 |
1,390 sales of which 90 come from the interaction
A set-up produces 900 incremental paid sales, 280 owned, 120 earned and 90 additional when paid and earned are present together. Unit contribution is €85. The costs are €70,000, €12,000 and €18,000.
Gross contribution: (900 + 280 + 120 + 90) × 85 = €118,150. Costs: €100,000. Net contribution: €18,150. Ignoring the interaction would underestimate the result by €7,650; crediting it to both families would overestimate it by the same amount.
The next allocation must preserve the interaction while testing whether the 90 sales persist when paid is reduced.
Acceptance conditions
What must be true to act.
- 01
Every action has an explicit role, population and horizon.
- 02
Direct and assisted measurements are not added without deduplication.
- 03
Paid-owned-earned interactions are estimated separately.
- 04
The marginal cost and the capacity of each family guide the allocation.
Limits
What this analysis does not prove.
- Earned is hard to assign and can react to the same events as sales.
- Fragmented identities make the sequences incomplete.
- An isolated experiment measures long-term reputation effects poorly.
References
Works cited.
- Stephen, A. T. & Galak, J. (2012), “The Effects of Traditional and Social Earned Media on Sales”, Journal of Marketing Research.
- Batra, R. & Keller, K. L. (2016), “Integrating Marketing Communications”, Journal of Marketing.
- Gordon, B. R. et al. (2019), “A Comparison of Approaches to Advertising Measurement”, Marketing Science.