The problem
Most creators think income works like this: make a good video, get sales, earn commission. That is only part of the story. Between the video and the commission sits a system that decides which videos get shown, how often, and to whom. That system is GMV Max, and it runs on two settings the brand controls. If you don’t understand those two settings, you will blame your video for problems your video didn’t cause.
Budget is the money on the dashboard each day. Target ROAS, which TikTok calls Target ROI inside GMV Max, is the return the brand is asking for. A target of five means the brand wants roughly five dollars in sales for every dollar of ads.
The budget is not the spend
This is the part most people get wrong. A brand sets a ten thousand dollar daily budget and a target of five. That does not mean TikTok will spend ten thousand dollars. It means you may spend up to ten thousand, but only where you believe you can hit about five times back. If TikTok can’t find enough places where it expects to clear that bar, it spends less. Sometimes much less. A hard target can leave most of the budget sitting unused.
So there are really two numbers, and only the second one matters to you:
- Stated budget. The figure on the dashboard.
- Deployable spend. Whatever TikTok is actually willing to spend while staying near the target.
The target decides how wide the door is
Picture a fifty dollar product. At a target of five, TikTok can afford about ten dollars of ads for every sale. At a target of two and a half, it can afford about twenty. That is double the room to buy impressions, reach new viewers, and find out which videos convert. So when a brand lowers its target, the door opens wider. When it raises the target, the door narrows. TikTok says this directly: more competitive targets let the system spend more and reach more people, and stricter targets restrict spend.
GMV Max allocates. It does not run a fair test.
Imagine two hundred creator videos attached to one product. It is tempting to think each video gets an equal share of the budget. It doesn’t. GMV Max is not running an experiment where every video gets a fair shot. It is constantly asking one question: where is my next dollar most likely to produce enough sales? Which means the room a video lands in matters as much as the video itself.
| The room | What happens to your video |
|---|---|
| Big budget, reachable target | Lots of money in play. Plenty of room to test new videos before spend concentrates. |
| Big budget, hard target | Looks well funded, but little actually gets spent. Few videos get a real test. |
| Small budget, reachable target | TikTok would buy more, but there isn't enough money to explore. Proven videos eat most of it. |
| Small budget, hard target | The hardest room to break into. Little money and a high bar. |
The first row is where new videos get discovered. The last row is where good videos quietly die.
The example
Two creators post videos of the same quality, on the same day, for different brands.
| Creator A | Creator B | |
|---|---|---|
| Daily budget | Large | Small |
| Target | Loose | Strict |
| Videos competing | A modest pool | A very crowded pool |
| What happened | Enough impressions for TikTok to learn whether it converts | A few hundred views, never enough evidence, never entered paid distribution |
Creator B did not make a bad video. Creator B’s video never got tested. Those are very different problems and they need very different responses. A new version won’t help if the room is the problem.
The whale effect
Once GMV Max finds a few videos that reliably beat the target, the system has an easy decision. Why spend money finding out whether the hundred and forty-seventh video works when it already has a proven place to put the next dollar? So early on, spend is spread across many uncertain videos. Later, it piles up on a handful of winners. That is why a brand can be spending heavily on a product while hundreds of creator videos get almost nothing. The money is there. It just isn’t being shared.
Why launch windows exist
Brands often loosen the target and raise the budget when a product launches. Both gates open at once. More money gets deployed, more impressions get bought, more videos get a real test, and more creators get a chance to become one of the funded winners. Then the launch settles, the brand tightens the target, exploration shrinks, and spend concentrates on the videos that already proved themselves.
That is the window. You are not just looking for products that sell. You are looking for the period when TikTok has been given enough permission to go find new sellers.
Three questions before you post
The usual question is “what’s the commission?” That is the last thing to look at, not the first. Ask instead:
- How much is actually being spent on this product? Not the budget. The real spend. You can’t see the number, but you can see the effect: are recently posted videos getting distribution, or do only a few old videos have views?
- How hard is the target relative to what the product can do? A product that converts easily can clear a strict target. A product that converts poorly can’t clear a loose one. Are several different creators converting on it right now?
- How many videos are competing for that spend? A fat pool split across a few hundred videos is a different opportunity than a thin pool split across a few thousand.
Put together: a high commission on a product with almost no real spend, a strict target and a crowded pool can be a worse deal than half that commission on a product with heavy spend, a reachable target and a thinner pool. The second product gives your video far more chances to get funded.
Your commission is not video quality times commission rate. It is closer to the chance of a real test, times the chance of clearing the bar, times how far it scales if it wins, times what you earn per order. That first term is the one creators forget.
TikTok can factor affiliate commission into how it allocates spend for some advertisers. If two videos perform about the same and one carries a much higher commission, the cheaper one may get more traffic. A high-commission video can still win, but its sales have to be clearly stronger. So a commission bump cuts both ways: you earn more per order, and the bar to get funded rises. Lesson 45 is about reading that change when it lands on a brand you already work with.
Read the room before you read the video
If you diagnose creator videos, this changes how you read a bad result. A video with a strong Hook, Payoff and Click that fails only at Convert is often a room problem, not a script problem. The brand tightened the target, the product page is weak, or demand cooled. That video is worth keeping and pointing at a product with a healthier room.
A video that never got past a few hundred views hasn’t failed. It hasn’t been tested. Don’t file it as a losing script, and don’t brief the creator a new version of it either. Wait for a window, or move it. And when the whole roster’s Convert drops in the same week, look at the room before you look at anyone’s creative. Fourteen creators don’t all get worse at once.
The simplest version of all of it:
- Budget sets the size of the pie.
- The target sets how much of the pie TikTok is actually willing to serve.
- How many videos are posted sets how many are competing for it.
- Performance decides who gets more after the first taste.
The video is the last variable in that chain, not the first.
Do this in Verdict
You can’t see budget, target or the size of the pool directly, so Verdict never guesses those numbers. It reads their effects.
- 1Open the Pre-Slot Screener before you write anything. It checks whether recent videos are breaking out, whether more than one creator is converting, whether winners are turning over, and which way the payout is moving. That is the room, read from the outside.
- 2A video still showing HOLD has not had a real test yet. Leave it alone. It is not a losing script until the room has spoken. Lesson 9 covers when a video has earned a verdict.
- 3When a test does fail, Learnings freezes the room it failed in alongside the call, so a starved video is never filed as a bad script and re-tried later as if it were one.