TikTok Shop fees and what selling there really costs

The referral fee on TikTok Shop is 6% in most categories, which reads as cheap next to any marketplace. It’s also the smallest of the four costs the channel imposes on you. Add affiliate commission, fulfilment and the ad spend the Shop Sales objective now requires, and a unit clearing 35% contribution on your own store can land under 16% there — or under zero.

What does TikTok Shop charge a seller?

The published seller referral fee is tiered, and the tiers are the part most summaries leave out. It’s worth reading the schedule as a schedule rather than as a single number, because two of the four rows change what a launch looks like.

The promotional row is the one to plan around. A new seller pays 3% for 30 days from their first sale, then the rate doubles. If you validate a product inside that window and read the margin you saw as the steady state, you’ve measured a discount and called it a business. Push the first sale to the day your inventory and your creator pipeline are both ready, so the cheap month covers your real volume rather than your test volume.

What does the affiliate programme actually cost?

More than the referral fee, and more than the rate card on any single product suggests. Across US TikTok Shop in 2025 the average commission affiliates actually earned was 13.07%, up from 12.27% in 2024, a figure circulated from Momentum Works and Tabcut's February 2026 report. Treat it as unverified: we could not find either number in the public version of that report, and neither firm publishes it openly. Use it as a starting estimate and replace it with your own realised rate as soon as you have one.

That’s a realised figure, not a posted one, and the distinction matters in both directions. It blends the products nobody picked up, which pay no commission at all, with the products creators competed over, which often carry 20% or more. So 13.07% is the market’s answer, not yours. If your product is the kind that needs a rate above the average to get any creator attention, your line is worse than 13.07% and the market average is quietly flattering you.

What does one unit look like once everything is counted?

Here is the same product on your own store and on TikTok Shop. Selling price $24, cost of goods 30%, pick-and-pack plus shipping $3.60, advertising at 20% of price. Those four are yours and you’ll want to replace them with your own. The two new rows are the platform’s.

Nothing exotic happened there. Two rows appeared and contribution more than halved. The reason it more than halved rather than falling by the 19% those two rows represent is that contribution is a residual, so a fixed slice off the top of revenue takes a much larger slice off what is left.

Now add returns. The NRF and Happy Returns study published on 15 October 2025 puts the online return rate at 19.3% of sales. If a return costs you $6 net of what you recover, that is $1.16 spread across every unit sold, and the two columns become $7.24 and $2.66.

Move one more line and the channel column disappears. Run advertising at 30% of price rather than 20% — a 3.3x reported ROAS, which isn’t unusual here — and TikTok Shop contributes $0.26 a unit while your own store still contributes $4.84. Offer affiliates 20% instead of the 13.07% average because otherwise nobody picks the product up, and the column is negative at -$1.40.

Why is the advertising line the hardest one to fill in?

Because the ad product reports on itself. GMV Max is now the default and the only ad type you can create for the TikTok Shop Sales objective. TikTok says sellers who adopted GMV Max saw an average 30% increase in GMV, which is a before-and-after for the sellers who switched rather than a comparison against manual video shopping ads — and it comes from TikTok's own UK launch announcement, so it is a vendor reporting on its own product and should be read as such.

The structural issue is separate from whether that claim holds. GMV Max credits Shop purchases that occur inside the campaign to the campaign, including purchases from customers who never saw an ad. Organic video, affiliate content and search inside the app all produce Shop orders, and while a GMV Max campaign is live a share of those land in its reported total. The ROAS it shows you is therefore inflated by GMV you would have received anyway, and the inflation grows with how strong your organic and affiliate presence is. The better your creators do, the better your ads look.

Is the 15% to 35% of GMV rule of thumb any use?

You’ll see total platform-tied cost quoted as 15% to 35% of GMV, alongside a $15-30 product sweet spot. Both bands are widely repeated and neither traces to a primary source, so treat them as unverified. They are also, on the worked example above, wrong in an instructive way.

Add the four lines the bands are meant to cover on that unit — referral fee $1.44, commission $3.14, fulfilment $3.60, advertising $4.80 — and you get $12.98 against $24 of revenue. That’s 54%, not 35%. Either the bands are describing sellers with much cheaper fulfilment and much lower ad dependence than the example, or they were never measured. Both are possible. Neither is a reason to use them in place of your own six rows.

Is the channel growing enough to justify the take rate?

On volume, yes, and by a lot. Momentum Works and Tabcut put US TikTok Shop GMV at $15.1B in 2025, up 68% from $9B the year before, with beauty and personal care the largest category. They do not publish a GMV figure per category.

The seller figure is the one worth sitting with. More than 2,000 stores cleared $1M in GMV on US TikTok Shop in 2025, against 803,500 stores in total. That is the supply side of your affiliate line professionalising, and professional affiliates negotiate. A market where that many operators have found the channel works is not a market where the realised commission rate is heading down.

The honest counter-argument is that none of this says anything about your product. A 68% growth rate is a statement about the channel, and the four rows in your P&L are a statement about you. Plenty of sellers are inside that $15.1B at a loss.

How do you decide before you list?

1. 01

   Build the six-row unit table first

   Price, cost of goods, fulfilment, referral fee, affiliate commission, advertising. If you can’t fill in the last one, use 25% of price as a placeholder and mark it as a guess. A guess you have labelled is more useful than a blank.

2. 02

   Set the affiliate rate from the market, not from your spreadsheet

   Look at what the top-selling products in your category are offering right now. If they’re all at 20% and you built the plan on 10%, you’ve built it on a product nobody promotes.

3. 03

   Derive a separate maximum CAC for this channel

   Contribution here is lower than on your own store, so the ceiling on what an order can cost is lower too. Carrying one CAC target across both is how a channel goes negative while every dashboard shows green. The break-even arithmetic is the same; only the contribution you feed it changes.

4. 04

   Decide what the promotional 30 days are for

   At 3% you’re seeing a rate you won’t see again. Use the window to learn what sells, and re-run the unit table at 6% before you commit inventory to anything.

5. 05

   Read blended spend over blended GMV from week one

   Set the habit before the numbers get big enough to argue about. GMV Max’s ROAS will always look better than the blended figure, and the gap between them is your organic and affiliate GMV being counted twice.

The mistake underneath all of this

Sellers compare channels on the fee, because the fee is the number the channel publishes. Everything else in the stack — the commission, the fulfilment profile, the fact that the ad product is mandatory and grades its own homework — has to be assembled by hand, so it usually isn’t.

Assemble it. One product, six rows, twenty minutes. If the answer is that TikTok Shop clears less than your own store but still clears, that is a real reason to be there: it’s incremental demand you weren’t reaching. If the answer is negative, no amount of volume fixes it, and the channel is a subsidy you are paying to strangers. There are a few more of these arithmetic traps in the ecommerce numbers that do not survive checking.