What We Know

Where does the money come from when the user is not the payer?

Short answer

Where the relationship layer cannot be billed directly, value has tended to accrue to whichever input is scarce rather than to the party holding the relationship. Serving one audience while billing another creates a standing tension over whose interests the product optimizes, and indirect economics have to be defensible once the customer can see them.

Based on 3 core studies and 2 related studies.

What history suggests

  • When the relationship layer cannot be billed, value tends to accrue to whichever input is scarce rather than to the party holding the relationship.
  • Serving one audience while billing another creates a standing tension over whose interests the product optimizes.
  • Trust-based distribution can carry a product a long way, but the trusted party's willingness to recommend depends on how the economics look to the end customer.
  • Once the customer can see the commission, indirect economics have to be defensible on their own terms.
  • Widely used reference points and benchmarks are not automatically the source of value. The monetizable asset is often adjacent to them.

What changes the answer

Conditions and contexts where the evidence differs.

  • Which input is scarce differs by market and can change over time, moving capture to a different layer than the one that holds it today.
  • Trust-based distribution works until the economics become visible to the end customer, at which point the recommender's own position is at stake.
  • Wide usage and billability come apart. The most used reference point in a market is often not the thing that can be charged for.

What to look at in your business

Practical application, not historical finding.

  • Which input in the chain is genuinely scarce today, and who holds it?
  • When the payer and the user disagree, whose interest does the product follow?
  • Would the current economics survive being shown to the end customer?
  • What is used widely but not billed, and what is billed but narrowly used?
  • How dependent is distribution on one trusted party's continued recommendation?
  • What would have to change for the scarce input to shift to another layer?

Core evidence

Related research