If you sell a cross-border TikTok Shop business in Indonesia without a local entity, now being pushed out of the shelf and into content, the numbers that decide your margin have been moving for a year, and most of them moved in the same direction. Most Indonesian sellers did not choose content commerce. They were moved into it. Organic recommendation for cross-border storefronts thinned through the first half of 2026 while policy moved the other way: search and recommendation now favour local small and medium businesses, and the service-fee relief announced for certified local enterprises does not extend to sellers without a local entity. At the same time the commission schedule stopped being a single number. Ordinary orders sit at the top of the band, live and ad-driven orders near the bottom, and the gap between those two ends is wide enough to decide whether a SKU makes money. That turns creator collaboration from a marketing question into a routing question: same product, same creator, same audience, different channel, different margin. This guide works through the four decisions that keep a creator programme priced on facts instead of on last year: what the two-track commission schedule actually charges you, how to split the catalogue by contribution margin before choosing a channel, how to write the commission basis into the agreement so it is not relitigated later, and what to change in the creator roster once shelf traffic keeps thinning.
Everything below is written to be used directly — in the brief you send a creator, or in the sheet you hand your finance team.
What the two-track commission actually charges you
The two-track schedule charges by how the order was created, not by what was sold. An order that arrives through a product card on a creator's video — the shopper watches, taps the cart, buys — is priced at the upper end of the band. An order that arrives inside a live room, or through paid amplification of a creator asset, is priced near the bottom. The published ceiling for ordinary orders and the published floor for live and ad orders are roughly twenty times apart, and while almost no seller sits at either extreme, the direction of the difference is what matters: the platform is pricing shelf-style demand higher than content-style demand. For a seller whose traffic used to come from search and recommendation, that is a direct transfer of margin to whoever can create demand inside content. It also explains why so many Indonesian catalogues that looked healthy on paper stopped looking healthy once recommendation volume fell — the orders that disappeared were the profitable ones to fulfil, and under the new schedule the expensive ones to book.
There is a second layer that rarely appears in the fee table: who is allowed to be charged less. The service-fee relief introduced for certified local micro, small and medium enterprises is not available to a cross-border seller without a local entity, so two stores selling the same imported product can face different total costs on the same day. That does not make the cross-border route unviable, but it does mean the creator plan has to carry more of the margin back. Content that a local competitor can afford to run at a thin commission is content a cross-border seller has to run at a commission it can defend with conversion rather than with reach.
Step 1 — Split the catalogue by contribution margin first
Before choosing a channel, sort the catalogue by contribution margin per order, not by revenue. The arithmetic is: selling price, minus landed product cost, minus platform commission at the rate that actually applies to that route, minus fulfilment, minus a returns provision that reflects the category, minus creator commission. What is left is the only number a channel decision can be made against, and it has to be calculated per route, because the commission line changes with the route while every other line does not. Three bands are usually enough. Band one: contribution margin above roughly a third of selling price, which can absorb a full commission and still fund paid amplification. Band two: margin between a fifth and a third, which can carry a creator commission but not a commission plus amplification. Band three: margin below a fifth, where any route that adds cost without adding conversion destroys the order. The common mistake is to sort by gross margin instead. Gross margin ignores fulfilment and returns, and in a market where return handling has become a seller cost, that omission is large enough to move a SKU from band two to band three without anyone noticing.
Do this once, in one sheet, and keep it current, because the inputs move. When the commission schedule changes, only one column changes; when a carrier raises rates, only one column changes; the bands stay comparable. Sellers who rebuild the table from scratch each time end up comparing numbers that were never measured the same way, which is how a loss-making SKU keeps getting funded for a quarter. Keep it per SKU and per route, and date-stamp it — a margin table without a date is a rumour.
Step 2 — Match each margin band to the cheaper creator route
Route one is the creator video with a product card. The shopper discovers the product in the feed, taps, and buys without a live host involved. It has the widest reach per unit of creator effort, it keeps working after the creator stops posting, and it is the expensive route under the two-track schedule. Route two is the live room, whether the creator's own or a session the seller runs with a creator present. The order is created inside a time-bound event, conversion is compressed into minutes, and under the schedule this is the cheaper route. Route three is paid amplification of a creator asset: the seller takes a video that already converted organically and puts spend behind it, so the order is attributed to the amplified asset and priced at the lower end. Matching is mechanical once the bands exist. Band one SKUs can run route one, because the margin can pay for reach. Band two SKUs should run route two, because they need the conversion density of a live event to make the commission work. Band three SKUs should not run creator routes at all until either price or cost changes; putting them in a live room only produces volume at a loss.
The mistake is treating these as a ladder — video first, then live once you are bigger. They are not stages, they are instruments with different cost structures. A seller who only runs lives gets no evergreen asset; a seller who only runs videos pays the top of the commission band on every order and has no way to compress conversion when a SKU needs a push. Most healthy catalogues run all three at once, on different SKUs, for different reasons.
Step 3 — Write the commission basis into the agreement
The commission basis has to be written down before the first video, because the two things that cause disputes are always the same: what the percentage is calculated on, and what happens when the order comes back. Four clauses do most of the work. First, the base: commission on the item price after seller discount but before shipping, stated explicitly, because 'commission on the order value' means different things to different people. Second, the treatment of returns: whether commission is reversed on a refunded order, and if so, how the creator is told. Third, the settlement rhythm: how often the creator is paid and what evidence they receive, because a creator who cannot see which order paid and which reversed will assume the worst. Fourth, the term: how long the rate holds, since a commission agreed in a month when the schedule was different is not the same commission in the next month. Write these four into the brief, in the same place the deliverables are described, and have the creator acknowledge them before the sample ships.
It is worth saying the awkward part out loud: under current schedules a refunded order can reverse the creator's commission while the seller still absorbs outbound shipping, return handling, and in some categories part of the platform commission as well. Both sides lose, and neither side caused it. Sellers who explain that before the first order get fewer arguments later than sellers who discover it in a dispute. The clause is not a legal formality; it is the thing that keeps a mid-tier creator from quietly moving their links to a competitor's platform.
Step 4 — Rebuild the roster for content, not for the shelf
The roster that worked when traffic came from search is the wrong roster now. Search traffic converts intent that already exists, so the useful creator was the one with the largest reach per rupiah. Content traffic has to create the intent, which puts a premium on creators whose audience believes them — and in Indonesia through 2026 that has meant a visible shift in what converts. Beauty and personal care still carry the largest single share of the platform's GMV, but that share fell through the first half of the year while other categories grew, and the sellers who kept growing were the ones whose content looked like evidence rather than advertising: unboxing with the product in hand, a demonstration that includes the failure case, a review that names what the product is not good at. So rebuild the roster on three filters. Does the creator's recent work contain at least one long-form demonstration, not only fifteen-second hooks? Do the comments contain questions the creator actually answers? Has the creator posted about a product they clearly kept using? A creator who passes all three is worth a lower follower count and a higher commission than one who passes none.
Then rebuild it geographically. If local-entity stores are favoured in recommendation, the creator whose audience is concentrated in Indonesia and who films in Indonesian settings is worth more than a creator with three times the followers whose audience is regional. Audience location is a filter, not a metric to optimise after the fact — check it before the sample goes out, not in the performance report afterwards.
How DAMI holds route, commission and SKU in one 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. For a routing decision the three relevant pieces are campaign management, so the commission and the channel are attached to the SKU rather than to a chat thread, store analytics, so the contribution figure each route is judged on comes from the same source, and the multi-store dashboard, so a seller running a cross-border store and a local-entity store reads one cost sheet instead of two.
Cost decisions are decided by records, not by recollection: which SKU, which channel, which rate applied on which date, and what came back. Start with DAMI and keep the whole calculation in one place.
Frequently asked questions
How far apart are the two commission tracks in practice?
Wide enough to change a routing decision. The published ceiling for ordinary orders and the published floor for live and ad-driven orders sit at opposite ends of the band, and the distance between the two ends is close to twenty times. Almost no seller sits at either extreme, because the rate depends on category and programme, but the ordering is consistent: orders created inside a live room or through paid amplification are priced well below orders created through a product card in the feed. Check the rate card in the seller centre for your own category, and build the margin table on that number rather than on the headline figure.
My store has no local entity. Does that change which route I should use?
It changes the margin you have to protect, not the routing logic. A cross-border seller in Indonesia does not qualify for the service-fee relief available to certified local enterprises, so the same catalogue carries a higher total cost. That generally pushes SKUs down one band, which means fewer SKUs can afford the reach-heavy route and more should sit in live or amplified routes, or wait until price or cost changes.
Do creators prefer one route over the other?
They prefer being paid predictably. Under current rules a creator's commission can be reversed weeks later when an order is refunded, and the dashboard does not itemise which order caused it. That is why so many mid-tier creators now spread links across platforms. A seller who states the settlement rhythm, shows the order-level record, and pays on a fixed date is offering something a rate alone cannot: certainty. In practice that retains creators better than a higher headline percentage does.
How do I know a route is actually cheaper after everything is counted?
Compare contribution margin per order on the same SKU across routes over the same period, and include creator commission, platform commission at the applicable rate, fulfilment in both directions, and the returns provision. The route with the lower commission line is not automatically cheaper — a live order that converts at twice the rate can beat a video order on total contribution even at a similar commission. The comparison only means something when every line is on the same basis and the period is the same.
Conclusion
The sellers who come out of a fee change intact are not the ones who cut the hardest, they are the ones who knew their numbers before it happened. Decide which SKUs can carry a creator, which channel each one belongs on, and what you will do when a line moves — then write it down so the next change is an update rather than a surprise. 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.
Price from contribution, state the basis in the brief, and keep the record. Get started with DAMI and run the next quarter from facts.
