User-generated content case studies: only your own numbers beat the rate card

No figures from other people's campaigns appear on this page. A percentage you cannot reproduce from your own records is decoration, and in this category the published ones are almost always cost per asset, which is the number least connected to what an agency actually spends. What follows is the small set of things to record so that your own campaigns become readable as cases, and three divergences that turn up whenever somebody does the recording for the first time.

What to take away

What to record, per campaign

Six fields, and none of them requires a product to capture.

The last row is the one nobody keeps and the one that pays. A list of rejection reasons across a year is the most accurate specification review an agency can own, and it costs a sentence per asset.

Divergence one: the unit that was paid for is not the unit that ran

Rate cards quote a price per asset. Campaigns run variants, and variants are where the actual work sits.

An order for a set of assets with several openings each is a different purchase from an order for the same number of finished films, and it is frequently priced as though it were the same. The divergence appears when somebody divides total spend by assets that ran and finds a figure well above the quote. Nothing improper has happened. The unit was never agreed.

Fix it in the order rather than in the analysis. State how many distinct openings, how many edits, and what a variant means for pricing. The specification items that most often go unstated are set out in the pillar guide to commissioning short form content.

Divergence two: review time scales with brief quality, not with volume

This is the one that surprises agencies most. Doubling the order does not double the review burden. Halving the clarity of the brief does.

The mechanism is simple. A precise specification produces deliveries that either match or clearly do not, and both outcomes are quick to judge. A vague specification produces deliveries that are arguably fine, and arguable deliveries consume meetings. A supplier that looks cheap and needs three rounds of notes per asset is not cheap, and the cost lands on the people least able to log it.

Time one campaign honestly, including the reading and the internal discussion. The number is usually larger than the fee that was negotiated over, which is what total cost of ownership means when it is applied to a service rather than to software. That is the same accounting problem described in the guide to campaign workflow software, where the visible price sits on an invoice and the real price sits in somebody's week.

Divergence three: the license expired before the asset did

An asset that is still performing when its term ends is a cost that arrives with no warning. Either the campaign stops, or the rights are renegotiated from the weakest possible position, which is after the other side has learned that the material works.

The prevention is a field, not a process: the expiry date recorded on the file, in the library, where a person will see it. Where a person appears on camera, likeness permission may run on its own clock and needs its own field.

What that permission covers, and why payment does not transfer ownership, is a question of copyright rather than of contract preference. The United States Copyright Office publishes a plain explanation of what copyright protects, worth reading before a negotiation rather than during one.

How to read a supplier's own case study

Skip the outcome and read the setup. The mechanics are usually accurate because they are not the part being sold: how many creators, what the brief specified, how delivery was structured, how long it took. That material transfers. Reading several of them beside each other is a comparison exercise, and how to structure one is set out in the guide to structured supplier comparison.

Then ask the question that reveals whether the story is evidence. What would have had to happen for this to be written up as a failure? A case that cannot answer it is a testimonial. A supplier who answers it honestly has told you more about how they work than the case study did.

Two further absences are worth noticing. No published case in this category states review hours, and none states the assets that were commissioned and never ran. Both are knowable and neither is flattering, which is why they are missing rather than because they are hard.

Turning your own records into something useful

After two campaigns you have enough to compare suppliers on cost per asset that ran. After four you can see whether revision rounds are a supplier property or a brief property, which is the single most valuable thing this exercise produces. After a year the rejection reasons will have rewritten your specification for you.

Keep the record with the work rather than in a personal folder, and keep it in a form that outlives whichever tool you are using this year. Where the same creators are booked for their audience rather than for footage, the economics change and the sourcing question belongs with the guide to creator marketplaces.

Common questions

How many campaigns before the numbers mean anything?

Two for cost comparison, four before you can separate supplier effects from brief effects. Anything earlier is a single sample wearing a percentage.

Should we share these numbers with clients?

Cost per asset that ran, yes, because it survives scrutiny and answers the question that follows every creative invoice. Review hours are an internal figure until you are ready to price them explicitly.

Our supplier disputes our rejection reasons. What now?

Good, that is the conversation working. Rejection reasons written at the time are arguable evidence; rejection reasons reconstructed later are opinion, which is why the field has to be filled the same day.

Can we use a competitor's published case as a benchmark?

Only for the mechanics. The figures are computed on definitions you cannot see, and in this category the definition of an asset is exactly what is in dispute.