Thirty Per Cent Prominence on a Nine Per Cent Catalogue
A regulator is consulting on a prominence obligation: on-demand services must give domestic titles at least 30% of 'prominent placements'. A mid-size subscription streamer holds 4,100 titles, of which 9% are domestic. Its homepage has 14 carousels, and the top three drive 61% of click-throughs. Compliance would require either raising the domestic catalogue share to 30% (which it says it cannot do at current acquisition budgets) or filling roughly nine of the 14 carousels with domestic titles, producing visible repetition. The streamer proposes raising domestic catalogue share to 15% and giving domestic titles 30% of homepage slots for their first 90 days; it also wants the metric defined as recommendation impressions rather than slot count. An internal A/B test found domestic titles placed in the top carousel received 3.1 times the impressions but completed at only 1.4 times the rate of the drama carousel; the recommender ranks by predicted completion.
1. Evaluate the streamer's proposal against the likely objectives of the prominence rule.
Read the worked answer
The rule's likely objectives are discoverability, plurality and correcting a market failure: global catalogue scale means domestic titles are cheap to acquire but expensive to surface, so they lose a competition the algorithm is not designed to win. Judged against that, the proposal is a partial answer with two weak points.
Strong points: raising catalogue share from 9% to 15% is a real supply commitment, and a 30% slot guarantee addresses surfacing directly rather than hoping the recommender corrects itself.
Weak points: the 90-day window treats prominence as a launch burst. After day 90 the titles fall back into a system that ranks on predicted completion, which is exactly the mechanism that disadvantaged them. A schedule that can be satisfied by rotating titles in and out of a window is also easy to game and produces churn rather than a durable library. And filling nine carousels from a 9% catalogue will produce repetition that damages the user experience, which in turn reduces future demand for domestic titles - the policy would undermine its own cultural objective.
A stronger package: a permanent floor on prominent placement (for example 20% rising to 30%), a cap on repeat placements so distinct titles are surfaced, disclosure of how the recommender weights domestic titles, and investment in metadata, dubbing and artwork, which is often the binding constraint on discoverability.
2. Which measurement definition should the streamer argue for, and what are the gaming risks of each option?
Read the worked answer
Three candidates, each with a different failure mode.
Slot share is transparent and cheap to audit, but rewards placement over exposure: a service can satisfy it by filling the fourteenth carousel, which almost nobody scrolls to. It also cannot distinguish one title repeated nine times from nine distinct titles.
Recommendation impressions measures exposure, but as the A/B test shows, impressions can be bought with placement while completion stays poor. Here the domestic titles drew 3.1 times the impressions but completed at only 1.4 times the rate, so an impressions metric would be satisfied by exactly the behaviour that annoys viewers. It also invites throwing impressions at users to hit a number.
Viewing share measures the outcome the policy actually cares about, but penalises new and small titles, rewards incumbents, and can become self-fulfilling if the algorithm optimises for it.
The streamer should argue for a composite: a primary outcome measure (share of viewing of domestic titles against a published trajectory), a secondary exposure measure (share of top-three-carousel impressions), and a discoverability indicator (share of the domestic catalogue receiving a minimum impression threshold per month). Composite metrics are harder to game jointly and let the regulator see whether exposure is converting. The streamer should also concede independent audit, since arguing for a self-reported metric will not survive consultation.
3. What unintended consequences might the rule create for small domestic producers, and how should they be mitigated?
Read the worked answer
The obvious risk is tokenism: a small number of domestic titles absorb the guaranteed slots because they are the safest performers, while the long tail gains nothing. A second is dependency: producers begin to model their output on whatever the platform's promotion algorithm rewards, which narrows the range of domestic content the policy was meant to diversify. A third is that slots become a scarce rent, and larger domestic suppliers with existing platform relationships capture them.
There is also a subtler harm. If promotion pushes under-performing titles in front of audiences, the immediate reception may be poor, and the platform learns that domestic content underperforms - entrenching the bias the rule targets. That argues for pairing prominence with investment in the things that actually drive completion: localisation, artwork, trailers and metadata quality.
Mitigations: count distinct titles rather than placements so repetition cannot satisfy the obligation; tier obligations by provider size so mid-size services are not required to do what only large catalogues can; ring-fence a share of prominent placements for independent producers; fund marketing and metadata support; and build in a statutory review at 18 months with published data. The review matters most, because the evidence base for prominence policy is thin and the rule should be capable of being revised.