Data ReserveAI data licensing research
Section 07 / Video and podcast roundups

4. How should we evaluate monetisation experiments without mistaking them for licensing income?

Data Reserve research deskSection 07Page 6 of 14Last updated 4 November 2026

Episode: What TIME learned from a month of selling ads to bots — The Media Copilot, publication dated 10 September 2026. Host: Pete Pachal. Guest: Mark Howard, introduced as TIME's chief operating officer. Watch: publisher-linked YouTube video. Evidence reviewed: original show notes and available transcript; direct video retrieval was throttled.

Howard discusses an experiment placing sponsored material in machine-readable pages, while Pachal asks about measurement, disclosure and the buyer's actual purchase. It is adjacent to content licensing, not a disclosed archive-licensing price benchmark. The guest's explanation of why AI systems may retain information should not be treated as independently verified model behaviour.

Exercise: distinguish the payer, product, claimed outcome and measurement method for any proposed experiment. Require a control or comparison that can separate an observed result from the seller's preferred explanation. State what result would cause you to stop.

Ask afterwards: “Are we selling permission to use editorial content, an advertisement, a technical service or several products together?” The answer determines which internal stakeholders must approve the proposal and what finance should report.

Sources cited on this page

  1. What TIME learned from a month of selling ads to bots ↗mediacopilot.ai
  2. publication dated 10 September 2026 ↗mediacopilot.substack.com
  3. publisher-linked YouTube video ↗youtube.com