How to Brief a UGC Video (With a Copy-Paste Template)

How to brief a UGC video: why the brief decides the outcome

When a UGC video flops, the post-mortem usually blames the creator, the edit, or the algorithm. A common pattern across DTC teams tells a different story: the brief was two lines long — "make a video about our serum, fun energy" — and the video performed exactly as well as that sentence deserved.

The brief is the only part of production you control completely. Creators interpret. Editors guess. Algorithms do whatever they do. But the inputs — what the product actually does, who it is for, what single job this video has — are yours, and ambiguity in any of them compounds downstream into revisions, delays, and generic footage that looks like everyone else's ads.

Think of it this way: a vague brief does not produce a worse version of the right video. It produces a confident version of the wrong video. The creator fills your gaps with the most generic interpretation available, because that is the safest guess when no one told them what mattered. You then pay for footage, watch it underperform, and conclude the format does not work — when the real failure happened in a doc before anyone hit record.

This guide covers the 10 fields a working brief needs, the reasoning behind the hardest ones, worked examples you can adapt today, the mistakes that quietly sink batches, and a copy-paste template at the end you can fill in within about 20 minutes. Whether you produce with in-house creators, freelancers, an agency, or an AI-native pipeline, the same discipline applies: the brief is where the video is actually made.

The 10 fields that matter (and what each one prevents)

A complete brief fits on one page. Ten fields, each one preventing a specific, predictable failure. The point of the list is not bureaucracy — it is that every blank field becomes a guess, and every guess becomes a revision.

Notice the pattern: the fields are not about creativity. They are about removing ambiguity. A brief is not a place to be inspiring; it is a place to be unmistakable. The creative range lives inside the hook directions and references — everything else is a constraint that keeps the creativity pointed at something that converts.

A useful sanity check before you send any brief: read each field and ask, "Could a smart stranger who has never seen my product execute this without emailing me a question?" If the answer is no anywhere, that blank is a future delay. The next sections unpack the fields teams get wrong most often.

UGC VIDEO BRIEF — [Brand] / [Product] / [Date]

1. PRODUCT TRUTH (one sentence, no adjectives)
What it actually does: ___
What it does NOT do: ___

2. AUDIENCE (one person, not a demographic)
Who: ___
What they've already tried: ___
What they're skeptical about: ___

3. ONE JOB FOR THIS VIDEO (no "and" allowed)
This video exists to: [stop the scroll / explain the mechanism / overcome objection X / drive add-to-cart]
Success metric: [hook rate / hold rate / CTR / CPA]

4. CLAIMS WE CAN MAKE (with a proof source for each)
- ___ (source: ___)
- ___ (source: ___)

5. CLAIMS WE CANNOT MAKE (banned words and lines)
- ___
- ___
Compliant swaps (regulated categories):
- Instead of "___" say "___"

6. PROOF ON HAND (assets we have rights to use)
Reviews / screenshots: ___
Demonstrable property the camera can show: ___
Numbers we can defend (with source): ___

7. HOOK DIRECTIONS (3–5 one-line angles, not scripts)
- Objection-first: ___
- Mechanism: ___
- Specific moment: ___
- Contrast / before: ___
- Pattern interrupt: ___

8. CTA (one action only)
Exact line: ___
Destination: [TikTok Shop / product page / link in bio]
Offer shown on landing page (must match the video): ___

9. SPECS & PLACEMENT
Aspect: 9:16 · Length: ___ seconds · Captions: on/off
Running as: [organic / Spark Ads / paid only]

10. REFERENCES
2 videos we like (and WHY): ___
1 video we hate (and WHY): ___

Product truth: the field that protects your comment section

Product truth is the foundation, and it is the one most brands fudge — because writing down what your product does not do feels like leaving money on the table. The opposite is true. Overselling does not just risk ad rejection; it loses the sale in the comments, where a single "didn't work for me" reply can deflate an otherwise strong video.

Write two sentences, no adjectives allowed. Adjectives are where overselling hides. "Revolutionary, fast-acting, dermatologist-loved" tells a creator nothing they can film. "It is a 2% salicylic acid leave-on liquid for oily, blemish-prone skin" tells them exactly what the product is and what frame to film it in.

Worked example (illustrative): say you sell a $34 vitamin C serum. Compare the two versions of product truth a creator might receive:

• Weak: "Our serum is an amazing glow-boosting treatment that transforms your skin."
• Strong: "What it does: a 15% vitamin C serum applied in the morning to support a brighter-looking, more even-toned complexion over weeks of use. What it does NOT do: erase dark spots overnight, replace sunscreen, or fix acne."

The strong version is filmable, it is honest, and it is compliant — and the "does NOT do" line is the part that protects you. It tells the creator not to imply an overnight transformation, which keeps the video out of the territory where commenters call it a scam and the platform flags it. For supplements, pet products, and skincare especially, product truth is also your first compliance gate: if the true description already avoids cure-and-treat language, the rest of the brief inherits that safety.

Audience: write one person, not a demographic

"Women 18–45 interested in wellness" is not an audience. It is a media-buying parameter, and it produces footage that is technically aimed at everyone and emotionally aimed at no one. The fix is to write one specific person with a specific problem, in their own words.

The most useful version of this field has three parts: who they are, what they have already tried, and what they are skeptical about. The "already tried" line is the secret weapon — it tells the creator your buyer is not a blank slate, which changes everything about the hook. Someone shopping for their fourth retinol does not need to be told retinol exists; they need to be told why the last three irritated their skin and this one might not.

Worked example for the illustrative $34 vitamin C serum:

• Who: a 29-year-old with combination skin who reads ingredient labels and has been burned by products that pilled or oxidized.
• Already tried: two cheaper vitamin C serums that turned orange in the bottle within a month.
• Skeptical about: whether this one is stable, or just better packaging around the same problem.

Now the creator has something to work with: a video that opens "if your vitamin C has ever turned orange, here's why" speaks directly to a real objection, instead of a generic "glow up your skincare routine." One sharp person beats a broad demographic every time, because specific resonates and generic gets scrolled. You are not narrowing your market — you are giving the algorithm and the viewer a clear signal of who should stop scrolling.

One job per video: the rule most briefs break

The most violated rule in UGC briefing: one video, one job. A video that tries to hook cold viewers, explain the mechanism, handle the price objection, and close the sale does all four jobs at a quarter strength each. It feels efficient to cram everything in. It is the fastest way to make a video that is busy and forgettable.

Write the job as a sentence a stranger could verify: "This video exists to overcome the 'I've tried vitamin C and it irritated my skin' objection for buyers who already know the category." Then attach the one metric it lives or dies by — hook rate, hold rate, click-through rate, or cost per acquisition — so the kill decision later is arithmetic, not opinion.

Here is a simple decision tree for assigning the job:

1. Is the audience cold (does not know the category)? The job is usually to stop the scroll and create curiosity — optimize for hook rate and hold rate.
2. Do they know the category but not you? The job is usually to differentiate or overcome a specific objection — optimize for hold rate and CTR.
3. Do they know you and are close to buying? The job is usually to push the click with proof or an offer — optimize for CTR and CPA.

This is also why volume works. When you are running 30 to 180+ creatives a month, no single video has to do everything. Each one gets a narrow job, the data tells you which jobs and angles deserve more creatives, and the portfolio — not any individual asset — carries the campaign. A common pattern many DTC teams find is that a small minority of angles do most of the work; you cannot know which in advance, which is exactly why each video needs one clean job and one clean number.

Claims you can and cannot make: write both lists

Most briefs list what to say. Almost none list what cannot be said — and that second list is where ad rejections, platform takedowns, and painful revision rounds come from. The "cannot" list is not legal paranoia; it is the cheapest insurance in the entire process. Ten minutes writing it can save a full revision cycle and a paused ad account.

Write both lists side by side. Every claim in the "can" column needs a source you could produce if challenged: a study you have the rights to cite, a verified review, a demonstrable product property. Everything else goes in the "cannot" column by default — the burden of proof sits with the claim, not the skeptic.

For regulated categories, the safest move is to model the compliant phrasing directly in the brief so the creator never has to improvise it. A practical translation table for skincare and supplements:

• Words like "cures", "heals", "treats", and "clinically proven" are off-limits in UGC unless qualified legal counsel has signed off on the exact wording and you hold the evidence — and platform policy is usually stricter than the law itself.
• Describe the experience, not the medical outcome. "For blemish-prone skin" survives review; "clears acne" often does not.
• Income, weight-loss, and before/after claims carry their own platform rules — check them per channel before briefing, because what passes on one platform can be auto-rejected on another.

None of this is legal advice — your category and region may impose their own rules, and when in doubt the conservative phrasing is the one that keeps the video live.

Proof on hand: list the assets before you write the script

A claim without proof is a liability; a claim with proof is a creative asset. The proof field exists to surface, before production starts, exactly which assets you have the rights to put on screen — because a brief that asks a creator to "show social proof" without specifying what is available just generates another clarification email.

Inventory four buckets and note what you actually hold the rights to use:

• Reviews and testimonials: screenshots of verified reviews, with permission to display them. A real five-star review reading "finally one that doesn't pill" is worth more than any line you could write.
• Demonstrable properties: things the camera can prove on its own — a texture, a pump count, a before-and-after of a stain lifting, water beading on a coated surface. These are the most persuasive proof because the viewer verifies them in real time.
• Numbers you can defend: a survey result, an ingredient percentage, a usage stat — only if you can cite the source and it is not dressed up as a medical claim.
• Earned credibility: press mentions, awards, or a founder's relevant background, where you have the rights to reference them.

Worked example: for the illustrative $34 serum, the strongest available proof might simply be a verified review screenshot plus a filmable demo of the serum staying clear and colorless after the cap is opened — directly answering the "does it oxidize?" skepticism from the audience field. That is a more convincing 15 seconds than any adjective. List the proof first, and the hooks and CTA almost write themselves around it.

Hook directions and CTA: aim the video, don't script it

A brief that contains a finished script is a brief that bets the entire video on one guess. Give directions instead: three to five hook angles, each one line, and let testing decide which earns a follow-up batch. The hook carries a large share of the outcome — most viewers decide whether to keep watching in the first two to three seconds — so you want range here, not a single locked idea.

Five reliable hook archetypes to draw from:

• Objection-first: open on the thing skeptics already believe — "I assumed every vitamin C serum just turns orange."
• Mechanism: show the how before the what — "this is why it stays stable."
• Specific moment: the exact use-case, not the lifestyle montage — "the 8am step I stopped skipping."
• Contrast / before: what the routine looked like before this product existed.
• Pattern interrupt: an unexpected visual or statement in frame one that buys you the next two seconds.

The CTA needs three things pinned down in the brief: the exact action ("tap the cart", not "check us out"), the exact destination (TikTok Shop, product page, or link in bio), and offer parity. Offer parity is the silent conversion killer — if the video promises one thing and the landing page shows another, you pay for the click and lose the sale at the door. One CTA per video. Two CTAs is zero CTAs, because a viewer given two choices usually makes neither. Keep the CTA soft enough to match the platform: on TikTok especially, a hard-sell close can suppress reach, so "tap the orange cart to see the price" often outperforms "BUY NOW."

Specs, placement, and references: the cheap fields that prevent re-shoots

The last three fields feel administrative, which is exactly why they get skipped — and skipping them is how a perfectly good video comes back cropped wrong, the wrong length, or "not what I imagined." These are the cheapest fields to fill and the most expensive to leave blank.

Specs and placement. State the aspect ratio (9:16 vertical is the default for TikTok, Reels, and Shorts), the target length, whether captions are baked in, and — critically — where the video will run. A video built for organic feed, one built for Spark Ads, and one built for paid-only placement are not interchangeable; the safe-zone for on-screen text and the CTA treatment differ. Naming the placement up front prevents a re-export and a crop disaster.

References. The fastest way to align on taste without writing a paragraph of adjectives is to point at examples. The format that works:

• Two videos you like — and one sentence each on why ("the pacing," "the unpolished phone-camera feel," "the way they handle the price objection").
• One video you dislike — and why ("too polished, reads like a TV ad," "the hook is a slow logo intro").

The "why" is the whole point. A reference without a reason invites the creator to copy the wrong attribute — you wanted the editing pace, they copied the lighting. Naming what you like turns a subjective vibe into an actionable instruction, and the "dislike" example sets a boundary that saves a revision round. Together with specs, these two fields convert "I'll know it when I see it" into something a stranger can actually hit on the first try.

What most brands get wrong: the five quiet brief-killers

The failures that sink batches are rarely dramatic. They are small, repeatable, and easy to fix once you can name them. Watch for these five:

1. Multiple jobs in one video. Already covered, and still the number one issue. If your job sentence contains the word "and," split it into two videos.
2. A "can say" list with no "cannot say" list. Listing the claims you want without the claims you must avoid is how ad accounts get flagged. The absence of a banned-words list is not neutral — it is an open invitation for the creator to improvise into trouble.
3. Demographics instead of a person. "Women 25–40" produces generic footage. One specific buyer with one specific objection produces a hook.
4. A finished script instead of directions. Locking the exact words before any data exists bets the batch on a single guess and removes the creator's ability to do what they are good at.
5. References with no reasons. "Make it like this video" copies the visible surface, not the mechanism. Always attach the why.

A sixth, subtler one deserves a mention: treating the brief as one-and-done. The first batch is a hypothesis, not a verdict. The teams that compound results are the ones who feed performance data back into the next brief — promoting the angles that worked, killing the ones that did not, and updating the "already tried / skeptical about" lines as they learn what actually moves their buyer. A brief is a living document, and the second one is always sharper than the first because the data wrote part of it.

Why a tight brief makes a fast first production run possible

On the Growth engine, the first 20 creatives are produced within 72 hours of fit approval. That speed is not a production trick — it is what becomes possible when the brief answers every question before work starts. Every blank field in a brief turns into a clarification email, and every clarification email adds a day. The brief is not paperwork that slows you down; it is the thing that lets you move fast without producing fast garbage.

An AI-native production engine removes the physical bottlenecks of traditional UGC — shipping product to creators, scheduling shoots, waiting on availability and revision queues. For context, a single human-creator video typically runs $200–$600+ all-in (~$150 base before product, shipping, and revisions), which is what makes high-volume testing slow and expensive by hand. What no engine can remove, AI-native or otherwise, is ambiguity. A vague brief produces fast, plentiful, mediocre creatives; a tight brief produces fast, plentiful creatives that are actually aimed at something.

The math is what makes the discipline pay off. Producing 30 to 180+ creatives a month only earns its keep if each one carries a clear job and a clear metric. The discipline matters because the engine's job ends where yours begins: we produce the creative at volume, then you run the ads and pick the winners, and the engine produces more like them. A brief with one job and one number per video is what makes that read clean a week later — you know which angles to ask for more of, and which to drop.

If you want to try the discipline against real volume, the Starter Engine is built for exactly this entry point: 30 creatives a month for $3,000, billed monthly with no long-term contract, so you can find which angles work before scaling up. You apply for free, and a fit-review confirms it makes sense before any invoice. But the engine only pays off if the brief is tight. Fill in the template below, and you have everything the production run needs on day one.

FAQ