If you sell a TikTok Shop seller in Thailand who runs creator-exclusive vouchers and creator-specific prices as a standing part of the affiliate programme, creator content is probably your largest source of orders and, quite possibly, your largest source of refunds at the same time. On 21 March 2026 the basis of the Thai transaction fee changed. It had been charged on what the buyer actually paid; it is now charged on what the buyer paid plus whatever the platform put in. For a seller running creator-exclusive vouchers, this lands on the exact mechanism the programme depends on. A creator voucher is, structurally, a platform subsidy wearing a creator's name: the buyer pays less, the platform or the seller funds the difference, and the difference used to sit outside the fee base. It no longer does. Every baht of voucher you add now increases the number the fee is calculated against, which means the voucher has three costs instead of two — its own value, the fee charged on its own value, and the commission booked on the order it produced.
Take one ฿399 order with a ฿50 creator-exclusive voucher. Three lines come out of it, and most sellers only book the first. The voucher itself: ฿50, which is 12.5% of the basket, and the line everyone remembers. The fee uplift: the fee is now charged on ฿399 rather than on the ฿349 the buyer actually paid, so the base grew by the full ฿50 subsidy and you pay the prevailing rate on that ฿50 as well — take your own rate from the back office and multiply, it is the line nobody books. And the commission: depending on how your campaign is set, commission may be calculated on the pre-voucher basket rather than on what landed in your account, which quietly moves the creator's take up by the same proportion. Add fulfilment, add a returns provision, and compare against contribution. The voucher that looked like a marketing cost is now a margin decision, and it has to clear the same bar as any other cost line.
What a creator voucher order actually costs after 21 March 2026
The reason this caught so many Thai sellers out is that the voucher was never modelled as a cost with a fee attached. It was modelled as a discount, and a discount is a reduction in revenue you can see in the settlement line. What changed is that part of the voucher is now also a base that a percentage is applied to, which is invisible unless you go looking for it in the fee detail rather than in the order summary. It compounds in a specific way: the bigger the voucher, the bigger the subsidy, the bigger the fee base, the more fee you pay on money you never received. A seller running a ฿100 voucher on a ฿399 basket is paying the fee on roughly a quarter of the basket that no buyer ever paid for. And the mechanism is the one creators ask for most, because a voucher is the easiest thing in the world to put in a live script — say the code, watch the orders come.
There is a second-order effect that matters more than the fee itself once the voucher is large. A deep voucher trains the audience to wait. If the same creator runs a ฿100 code every session, the buyers who would have paid ฿399 learn to buy during a session instead, and the voucher stops producing incremental orders and starts discounting ones you already had. That is why the right question is not 'how much voucher can I afford' but 'how many orders does this voucher produce that would not have happened at the lower voucher'. The fee change makes the arithmetic worse; the cannibalisation is what makes it unrecoverable.
Step 1 — Put the subsidy into the fee base and recount one order
Open one creator voucher order and reconstruct it line by line from the fee detail, not from the order summary. Write down: basket before voucher, voucher value, what the buyer paid, the platform subsidy, the fee base the platform actually used, the fee charged, the commission base and the commission charged, fulfilment, and a returns provision. Then recompute the same order under the old basis — fee on what the buyer paid — and put the difference in its own column. That column is the cost of the rule change, per order, in your own numbers. Multiply it by monthly voucher orders and you have the number worth acting on. Most sellers find it is not large enough to abandon vouchers and not small enough to ignore, which is exactly the situation where a per-SKU decision beats a policy decision.
Do this on three orders, not one: a small voucher, a mid voucher, and the deepest voucher you run. The relationship between voucher depth and total cost is not linear once the fee base moves, and three points are enough to see the curve. If the deepest voucher costs more than double the mid voucher for less than double the orders, you have your answer without building a model.
Reconstructing three orders by hand is where most teams give up, and the fee detail is not the friendliest screen to work from. If you want the columns already laid out, register with DAMI and ask the team for the free subsidy-base fee cost sheet: subsidy, fee base, fee charged, commission base and contribution in one row, with the old-basis comparison built in.
Step 2 — Find where the voucher stops paying for itself
The voucher stops paying for itself at the point where the next baht of voucher buys less contribution than it costs. You do not need a model to find it — you need two voucher depths run on the same SKU with the same creator, for the same number of days, with orders and refunds both counted. Compare contribution per order, not order count, because a deeper voucher almost always produces more orders and that is exactly the number that will mislead you. In practice the break tends to come earlier than sellers expect, because the three costs move together: voucher value up, fee base up, commission base up, while the incremental buyer a deeper voucher reaches is by definition less willing to pay. Set the depth one notch below the break and leave it there.
Run it on the SKU, not on the shop. A voucher depth that is profitable on a ฿899 item can be outright loss-making on a ฿199 one, because the same ฿50 is 6% of one basket and 25% of the other. This is also why a single voucher policy across the catalogue is the wrong instrument: it forces every SKU to live with a depth set for the average one, and the average SKU is the one nobody is actually optimising.
Step 3 — Decide who funds the voucher and put it in the campaign
Decide who funds the voucher before you argue about its size, because the answer changes the arithmetic. A seller-funded voucher is a discount with a fee attached. A platform-funded subsidy is the one that creates the fee-base uplift you are now paying for, which means the cheapest subsidy is not always the cheapest option — a platform subsidy you did not ask for can raise your fee base without raising your orders. Write the funding source into the campaign rather than leaving it to whatever the default is at setup, and check the setting on every campaign created before the rule change, because defaults set then were chosen under the old basis.
Re-examine campaigns built before 21 March 2026 specifically. Whatever was true about their economics when they were created, the basis underneath them moved, and a campaign that was marginal then can be negative now without a single setting having been touched by anyone. This is the cheapest audit available to a Thai seller this quarter: it costs one afternoon and it finds money that is leaving every day.
Step 4 — Move vouchers onto SKUs you can defend, not onto the whole catalogue
Give vouchers only to SKUs whose contribution survives the full three-line cost, and take them off everything else. In practice that means the top of the catalogue by margin, the SKUs whose refund rate is below your category line, and the ones where a voucher demonstrably changes the order rather than discounting it. For the rest, use a creator-specific price instead of a voucher where the structure allows it, because a lower list price keeps the fee base and the buyer payment the same number, while a voucher splits them and puts the difference in the base. Where neither works, the honest answer is to pay the creator more and discount the product less — a higher commission on a clean price is often cheaper than a voucher that inflates the base it is charged on.
Write the rule as a per-SKU ceiling rather than a campaign-level number: no SKU carries a voucher deeper than a named share of its basket, and the share is set by margin band. A ceiling travels — it survives staff changes, new creators asking for deeper codes, and the next fee change, because it is attached to something that does not move when policy does.
How DAMI holds voucher, commission and fee base on one SKU record
DAMI brings an 8-million-plus creator database, competitor and competitor-store creator search, automated outreach tasks, AI-assisted outreach, sample management, campaign management, store analytics, and multi-store team dashboards into one place. A voucher decision is three numbers that have to be visible at the same time, and they live in three different screens today. Three steps put them on one record.
Step 1 — Attach the voucher to the SKU, not to the creator
Set voucher value and funding source on the SKU's campaign record, alongside its margin band. Output: every SKU that carries a voucher shows the depth, the share of basket it represents, and who funds it, in one place.
Step 2 — Keep commission and fee base on the same record
Hold the commission rate and the fee basis that applies to the SKU next to the voucher, so the three cost lines are read together. Output: no voucher decision is made against a single number, and the uplift from the subsidy is visible before the campaign goes live.
Step 3 — Review depth against contribution after each cycle
After each cycle, compare voucher depth to contribution per order on the same rows. Output: the SKUs where the voucher has crossed its break point surface in one sort, before the next campaign is built on the same assumption.
Frequently asked questions
Does the subsidy really increase the fee I pay, even if the platform funded it?
Yes, on the current basis. The fee is charged on buyer payment plus platform subsidy, so a subsidy you did not fund still enlarges the base you are charged on. This is the part most sellers miss: the cheapest-looking funding source is not automatically the cheapest outcome, and a campaign with a generous default subsidy can cost more than a smaller seller-funded voucher.
Should I stop using creator vouchers entirely?
No. Vouchers still convert, and they are the one mechanic a creator can deliver inside a live script. What changed is that a voucher now has to clear a margin bar rather than a marketing budget. Keep them on the SKUs that clear the bar at a depth below the break point, and take them off the ones that do not.
Is a creator-specific price better than a voucher?
Often, where the structure allows it. A lower list price keeps buyer payment and fee base as the same number, so there is no subsidy sitting in the base. A voucher splits them and the difference lands in the base. Compare both on the same SKU for one cycle before deciding, because some buyers respond to a code and not to a lower price.
Can I work this out in a spreadsheet?
For the first recount, yes — three orders, three columns. It stops being enough when the question becomes per-SKU and recurring: which SKUs are above their ceiling this week, and did the depth change last cycle move contribution. That needs voucher, commission and fee basis refreshed on the same rows every cycle, which is where a sheet turns into a task nobody repeats.
Conclusion
Every cost line in this article moves without anyone deciding to move it, and every one of them is cheaper to fix before the next campaign than after it. DAMI brings an 8-million-plus creator database, competitor and competitor-store creator search, automated outreach tasks, AI-assisted outreach, sample management, campaign management, store analytics, and multi-store team dashboards into one place.
Fifteen minutes today: open the campaign you created before 21 March 2026 with the largest voucher, find its fee detail for last month, and write down three numbers — total voucher value, total fee charged, and orders. Divide the first two by the third. If the fee per order is higher than it was on a comparable non-voucher order, the campaign is paying for the subsidy and you now know by how much.
Start with DAMI and run the next cycle from numbers you can defend.
