Last month I counted the briefs sitting in our inbox. Forty one of them. The average length was about four hundred words. The average amount of useful information in those four hundred words was roughly two sentences. Everything else was plot summary.
That is the problem. Creators think a brief is a story recap. It is not. A brief is a set of constraints. And after three years of designing thumbnails for channels in finance, fitness, gaming, tech and a few niches I still cannot pronounce, I have landed on a hard opinion: if your brief cannot fit in five lines, you have not decided what the video is about yet.
Five lines. That is the whole system. I am going to walk through what each line does, why the long version fails, and what happens on our side when a brief is good versus when it is mush.
Why long briefs produce worse thumbnails
Here is the mechanic nobody tells you. A designer reading a long brief does not absorb more. He absorbs the last thing he read and the thing that was repeated most. So a nine hundred word brief does not give me nine hundred words of direction. It gives me whatever survived the skim, which is usually the part where you described the intro.
Worse, long briefs hide indecision. When somebody writes three paragraphs about a video, they are usually writing their way toward a point rather than stating one. I have to guess which sentence was the point. I guess wrong maybe a third of the time, and then we burn a revision round on something that was never a design problem in the first place.
Short briefs force the decision upstream. If you have to compress the video into one line of promise, you find out fast whether the video actually has a promise.
A brief is not a description of the video. It is a description of the reason someone should stop scrolling. Those are different documents and most people send me the wrong one.
We include title consultation in every pack for exactly this reason. Half the time the brief improves the moment somebody says the title out loud and hears that it does not promise anything.
Lines one and two: the promise and the register
Not the topic. The promise. "Budgeting tips" is a topic. "I lived on 800 dollars for 30 days" is a promise. One of those tells me what to put on screen and one of them does not.
The eight word cap is not arbitrary. Eight words is roughly the length of a thumbnail text overlay that still reads at 320 pixels wide. If your promise cannot survive that compression, the thumbnail is going to be a paragraph in a box, and paragraphs in boxes lose.
Write this line as if it were the title. Often it becomes the title. That overlap is the point.
The register
One word or two. Angry. Smug. Curious. Panicked. Warm. Deadpan.
This is the line people skip, and it is the line that changes the most about the final file. Same promise, different register, completely different design. "I lost 40 thousand dollars trading" in panic mode is a wide-eyed face, hot reds, a chart falling off frame. The same line in deadpan mode is a flat stare, a muted background, tiny type, and it works because the restraint reads as confidence.
We test both registers on our own channels before we push a new style to anyone. On one finance channel around 80k subs, the calmer register outperformed the panic version by a meaningful margin over a run of about a dozen videos. I want to be careful here: that is one channel, one audience, one period. It is not a law. It told us the audience read loud panic as clickbait and quiet certainty as authority. On a gaming channel in a similar subscriber band the result went the other way. Register is audience-specific, which is exactly why I need you to tell me yours instead of letting me infer it.
Lines three and four: the anchor and the boundary
Name the single object, face or moment the thumbnail is allowed to be about. One. Not three.
"Me, the broken laptop, and the invoice" is three thumbnails fighting for the same rectangle. Pick the laptop. Or pick your face. The third element is almost always the one that turns a clean composition into a collage, and collages die at small sizes.
If you genuinely cannot pick, that is useful information too. It usually means the video has two videos in it.
This line is also where you tell me what footage exists. If the anchor is a moment from the video, say the timestamp. Nothing slows a job down like a brief that specifies a shot which was never filmed. When the only available still is soft or low resolution, we can often work with it, and I have written before about what actually happens when we go about rescuing a soft screenshot rather than faking a replacement. But it is cheaper for everyone if you flag it in the brief.
The boundary
What must not appear. This is the most underrated line in the whole system.
Boundaries are faster than preferences. "No red arrows, no shocked face, no numbers over 3 digits" removes an entire region of the possibility space in one sentence. Compare that to "I want it to feel premium but still energetic and approachable", which removes nothing and means nothing.
Common useful boundaries:
- No face on this one, the story is the object
- Do not use the word "secret" anywhere
- Keep it consistent with the last four uploads
- Break consistency deliberately, this is a departure video
That last pair is the real fork in the road. Whether a thumbnail should match the channel or interrupt it is a strategy question, not a taste question, and it is the whole argument in how much variety a channel can survive. If your brief does not answer it, I will answer it for you, and I will answer it conservatively.
Line five: the shelf life
Is this video meant to work this week or for the next two years?
A trending video can lean on a face everyone recognises right now, a reference that will be stale by spring, text that assumes context the viewer already has. An evergreen video cannot do any of that. It has to explain itself to a stranger who finds it in eighteen months from a sidebar recommendation.
Designers make different choices for those two jobs. Sharper topical hooks and higher contrast for the first, cleaner and more self-explanatory framing for the second. I have gone deep on that split in the case for building an evergreen library, and it changes typography, crop, and how much cultural context we are willing to assume.
One word answers this line. Evergreen. Trending. Done.
What five good lines look like in practice
Here is a real shape of one, anonymised the way we always do:
- Promise: I cancelled every subscription for 90 days
- Register: dry, slightly smug
- Anchor: my face, half lit, single cancelled card in frame
- Boundary: no red X marks, no dollar figures in the text
- Shelf life: evergreen
That is thirty odd words. It is more directive than the four hundred word version would have been, and I can start designing off it immediately. The dedicated designer on that account does not have to interrogate the client to begin.
Compare that to the long version, which usually reads like: "So in this video I talk about how I went through all my subscriptions, there were about fourteen, and I break down which ones I kept, and there is a funny bit in the middle where my wife reacts, and I want something eye-catching that stands out."
Eye-catching that stands out. Every brief says that. It is the design equivalent of ordering "food".
The part where the brief stops mattering
I want to be honest about the limits of this. A perfect brief does not guarantee a winning thumbnail, because a thumbnail does not perform in isolation. It performs next to eleven other thumbnails in a sidebar, under a title, at a size you did not design for.
That is why we mock every concept into the actual surfaces before we call it done, and why we look at concepts in the feed rather than admiring them at full size on a clean canvas. Plenty of designs that look brilliant at 1280 wide collapse into grey soup on a phone. The brief cannot save you from that. Only checking can.
Unlimited revisions exist in our pack for the same reason. A brief is a hypothesis. Sometimes the hypothesis is wrong and the second round is where the video actually gets found. That is normal, and I would rather run three rounds on a real idea than one round on a vague one.
When you should talk instead of writing
There is one case where five lines is not enough, and that is when you do not yet know what the channel is doing. If you are mid-rebrand, or you are trying to move from one audience to another, or your last ten videos have no visual through-line at all, a brief is the wrong tool. You need a conversation about the system before you can brief individual videos into it.
That is the entire purpose of a thumbnail consultation: we look at the last twenty uploads, find what is actually holding the channel together visually, and then briefs become easy because the constraints already exist.
Live content is its own animal too. The brief compresses even further because you are often writing it hours before you go on, and the design has to work as both a pre-stream poster and a post-stream VOD. We handle those differently under live stream thumbnails, and the shelf life line becomes the most important of the five.
If you want to see how these constraints resolve into finished files across very different niches, our work is the fastest way to judge whether this way of thinking matches yours.
What this costs you, and what it saves
Our pack is 250 dollars for 8 custom thumbnails, one per video, which works out to about 31 dollars each. Every one is designed for that specific video. We do not run templates, and we test new styles on our own channels before they touch a client account. Turnaround runs 48 to 72 hours per thumbnail. If you want a second variant to test against the first, that is a 15 dollar add-on, and a title pack is 25 dollars if you want the words worked on properly alongside the image.
None of that pricing changes based on how good your brief is. What changes is how many rounds we spend. A five line brief usually lands in one or two. A four hundred word brief usually lands in three, because the first two rounds are me discovering what you meant.
The same principle applies well outside thumbnails. When we started taking on podcast work at our sister studio, the video podcast side of what we do, the first thing we built was a five line intake form, for exactly the same reason. Constraints beat descriptions in every creative handoff I have ever been part of.
Here is the last thing, and it is the one I actually believe most. Write the five lines before you record the video, not after.
If you cannot write the promise line before shooting, the video does not have a promise yet and you are about to spend six hours editing something nobody asked for. If you cannot name the visual anchor, you are going to finish the edit and discover you never filmed a usable frame.
The brief is not paperwork you do for the designer. It is the cheapest possible test of whether the video is worth making. Five lines, ten minutes, before the camera turns on.
If you want a second pair of eyes on the first few, send them over. I read every one that comes in.