Most affiliate briefs specify hashtags, mention requirements, and disclosure language, then leave everything that determines revenue to the creator.
We audited 5,871 affiliate videos across nine FMCG product lines in Indonesia and ran frame-by-frame teardowns on the extremes. The gap between a video that earned Rp 119 million and one that earned nothing traced back to decisions made before anyone opened a camera.
Nine of them. Each one is a field you can put in a brief this afternoon.
1. The Opening Frame
Specify: product visible in frame one, named within three seconds.
In the mosquito repellent teardown, every zero-revenue video opened on the problem. A burning mosquito coil, smoke, ash. Every top performer opened on the device or its pack, with the benefit stated immediately.
Problem-first openings are standard copywriting advice and they cost you category clarity at the exact moment a viewer decides whether this video concerns anything they might buy. In long-form the setup earns attention. In a three-second scroll it is a delay.
This is the first line we would write, and almost nobody writes it down.
2. One Differentiator, Not a List
Specify: the single benefit this video argues. Name it. Exclude the rest.
This one is our recommendation rather than a measured result: the top mosquito repellent video still covered several benefits. But it led with one, and our beat sheet gives seconds two through six to exactly one claim before moving to proof.
Creators given a full feature list will read the full feature list. If you want one differentiator, the brief has to contain one differentiator.
3. Proof Type
Specify: what the camera must show, and in what setting.
Winning videos showed hands doing things. Inserting the refill, pressing the mode switch, plugging into a real wall socket in a real room. In the pest adhesive teardown, failing videos leaned on spectacle proof, such as a bottle test, with no real use setting.
The distinction that mattered was use-context versus spectacle. A demonstration in the environment where the product gets used outperformed a more impressive demonstration with no context attached.
Worth noting what the teardown did not find: across six videos for an insect repellent line, none showed the product working. No dead mosquito, no overnight result. The strongest available proof was never requested, so it was never filmed.
4. Approved Claims, Written for Speech
Specify: the exact sentences a creator may say, and a short list of what they may not.
Affiliate videos carry the claims a brand answers for: efficacy comparisons, duration, running cost, safety around children. Whoever wrote them, the brief should fix the wording rather than leave it to the creator's memory.
Hand creators language they can actually speak on camera. A regulatory dossier is not usable at 24 seconds. If the approved wording is unusable, creators improvise, and improvisation is where claim exposure comes from.
Pair it with a negative list using real phrases rather than categories. "Do not say it works better than every other brand" lands better than "avoid comparative claims."
5. The Objection to Remove
Specify: one or two, named.
Our beat sheet gives seconds fourteen through twenty to answering a specific hesitation. Installation difficulty. Running cost. How long a refill lasts. Whether it is safe around children.
Which objection matters is category-dependent and you know it better than the creator does. A pest control buyer worries about whether it works. A baby care buyer worries about what is in it. Naming the objection in the brief is the difference between a video that handles it and a video that repeats the benefit again.
6. The Exact Offer
Specify: price, bundle contents, promo condition, refill availability.
The failing videos in the mosquito repellent teardown still pointed viewers to the cart. What they lacked was a reason to act now. The instruction was present. The motivation was not.
Top performers attached something concrete: a specific kit price, how long a refill lasts, a stated promo window. Creators cannot invent a bundle price. If the brief does not hand them one, the close defaults to "check my cart," which is what most of the zero-revenue videos said.
7. CTA Placement
Specify: cart location wording, positioned at 70 to 85% of runtime.
In every top performer we tore down, the call to action landed in that window and never earlier. A 14-second video put it at 11 seconds. A 44-second video put it at 35.
Earlier than 70% and the case has not been made. Later and there is no room to state it properly. Give the creator the wording too. A named spot like "cek keranjang kuning kiri bawah" is clearer than "check my cart," but one zero-revenue video named the same spot, so treat it as necessary, not sufficient.
8. Continuous Voiceover
Specify: spoken narration throughout, unless music-only is a deliberate choice.
In the pest adhesive and mosquito repellent teardowns, every top-performing video carried continuous spoken narration. It is not a universal rule: one of the top-earning shampoo videos was a short music-only product montage.
For a product whose case has to be explained, the voiceover carries the selling logic. Without it, the on-screen text and the product shots have to do all the work, so make music-only a decision rather than a default.
9. Duration Ceiling
Specify: a maximum, not a minimum.
In the pest adhesive teardown, top performers averaged 31 seconds and zero-revenue videos averaged 37. The longer videos were not adding proof, they were repeating claims.
The second-highest earner in the sample ran 14 seconds and had room for four beats: price surprise, bundle reveal, mode explanation, cart location.
Briefs that set a minimum duration produce padding. Set a ceiling instead.
The Assembled Structure
Those nine fields resolve to the beat sheet we distilled from the teardowns. It is a recommended structure, not an average measured across every category.
| Time | Beat | Brief field |
|---|---|---|
| 0:00–0:02 | Product visible, one specific claim | Fields 1, 2 |
| 0:02–0:06 | The single differentiator | Fields 2, 4 |
| 0:06–0:14 | Proof in use context | Field 3 |
| 0:14–0:20 | Named objection removed | Field 5 |
| 0:20–0:28 | Exact offer, then cart location | Fields 6, 7 |
Fields 8 and 9 apply across the whole runtime rather than to any single beat.
What This Does Not Fix
A brief cannot rescue a badly matched product. Baby care ran an 86.5% zero-revenue rate against 16.6% on a plug-in device line, using the same seller and the same creator pool. Trust-gated categories need credibility signals that no amount of beat-sheet discipline supplies.
It also will not make a discovery-category video convert like a problem-category one. A reed diffuser video converting at 0.12 sales per thousand views is not executing badly. It is being watched by people who were not shopping.
The brief controls the gap between a competent video and a converting one. It does not control whether the category was winnable.
For why category choice sets the ceiling, see intent categories versus discovery categories. For the evidence behind these nine fields, see the analysis of 5,871 FMCG affiliate videos.
Why It Is Worth the Hour
Of 92 creators in this sample, 23 produced both top-earning videos and videos that sold nothing. The same people, capable of the winning version, delivering it inconsistently.
What separated their good work from their bad work was mostly what they were handed. That is the cheapest variable in the entire program to change.