If you are a seller running both a cross-border store and an Indonesian local entity store on TikTok Shop, holding one creator pool and deciding which products and which creators sit on which line, the decision in front of you is no longer whether to add another platform. It is how the same creators and the same assets get split across the places you now sell. The regulation that landed on 17 June 2026 is narrow enough that most cross-border sellers read past it, and specific enough that it changes what your second store is for. Permen UMKM No. 3/2026 requires Shopee, TikTok Shop and Lazada to cut platform service fees — commission and transaction fees together — by at least 50% for micro and small sellers that are verified through the government's Sapa UMKM system and sell only goods made in Indonesia. The cost sits entirely with the platforms; there is no government subsidy behind it. Qualification needs an NIB, a Sapa UMKM application verified by both the government and the platform, and a catalogue with nothing imported in it, and the red line is real: list an imported product and the platform can terminate the discount. Two categories are excluded outright — ready-to-eat processed food, and electronics made by large domestic industrial firms. What this does to a dual-store seller is not "local is cheaper, therefore do local." It splits the two stores into different jobs. One line can carry the fee discount and the local-entity traffic weighting; the other is where you test new products cheaply and keep long-tail imported assortment that would disqualify you if it sat in the same store. The moment both lines carry the same SKU at different prices and the same creators promote both, the discount stops being an advantage and starts being a price war you are running against yourself.
Put the gap in the unit you actually book. Take a dual-store seller doing 12,000 orders a month at Rp89,000 average order value, so Rp1.068bn of GMV. On the fee line, work the published example: a Rp100,000 product at a 15% platform rate costs a cross-border seller Rp15,000 and a verified local seller Rp7,500, so the gap is Rp7,500 per order before you touch anything else. Applied across 12,000 orders that is Rp90m a month, or about Rp1.08bn a year, sitting on a qualification most cross-border sellers cannot meet in full — which is exactly why the practical move is to split which SKUs sit on which line rather than to pretend the whole catalogue can be localised. Then price the self-inflicted part. If one shared pool of 40 creators promotes the same SKU on both lines, and creator-driven orders are 25% of volume, that is 3,000 orders a month exposed to two prices for the same product. Put 30% of that on overlapping creators and you have 900 orders where the buyer can see a cheaper version of what they just bought; at an 8% rate of returns, refund claims or price complaints, that is 72 orders and roughly Rp9.3m of GMV a month, plus the review damage that follows a buyer realising they overpaid. Add the time cost nobody budgets: re-briefing and re-approving the same creator twice because nobody recorded which line they belong to.
What the 50% fee cut actually changes for a dual-store seller
Three mechanics sit behind the headline, and each one is a reason a shared pool starts costing you money. The first is the fee asymmetry itself: platform service fees in Indonesia run 10% to 18% generally, with TikTok Shop's blended take commonly above 25%, and the regulation gives the verified local micro seller at least half of that back while leaving the cross-border seller on the standard rate. The second is that the local store is not only cheaper, it is weighted differently: the same short video tends to draw 20%–35% more views and 20%–40% higher order conversion on a local store than on a cross-border one, and local fulfilment arrives in one to three days instead of seven to fifteen, with roughly 30% lower fulfilment cost and a materially lower return rate. The third is the one that turns an advantage into a fight: the qualifying catalogue has to contain nothing imported. That means the imported long tail you still want to sell has to live somewhere, and if it lives in the same store it costs you the discount, while if it lives in a second store promoted by the same creators it costs you the price integrity. On top of that, the 2026-08 Indonesian rules require sample-based review videos to carry the official commercial partnership label from 31 August, with unlabelled content taken down automatically and repeated violations restricting affiliate access for seven to fifteen days — so content attribution now has to be right per line, not just per brand.
Two qualifications matter more than the headline number. Size: micro means net assets at or under Rp50m or annual revenue at or under Rp300m, and small means net assets at or under Rp500m or annual revenue at or under Rp2.5bn — a seller above those thresholds gets no discount regardless of how local the catalogue is. And purity: the discount is calculated per domestic-goods transaction, so a single imported SKU in the mix does not merely dilute it, it puts the whole qualification at risk. The same regulation also tidies the rest of the relationship — fee changes need three months' written notice, fees collapse into registration, service and promotion categories, and contracts run a minimum of one year with no unilateral rate increases inside the term — which is useful, because it gives you a planning horizon instead of a monthly surprise.
Step 1 — Split products by whether they can be localised, not by margin
Sort the catalogue on one question rather than on margin: can this SKU be sourced, stocked and fulfilled inside Indonesia, and can you evidence that it is domestically made? Everything that answers yes belongs on the local line, because that is the line that can carry the fee discount and the fulfilment advantage. Everything that answers no belongs on the cross-border line: new products you are still testing, imported assortment you do not want to localise yet, and anything whose volume does not justify setting up local stock. The reason to sort on this axis rather than on margin is that margin tells you what a SKU earns today, while localisability tells you which line it is allowed to earn on at all. Write the result down as two lists with an owner each, and give every SKU a line, including the ones you are unsure about — an undecided SKU is the one that ends up on both.
Keep one exception explicit: a SKU that sells well on the cross-border line is a candidate for the local line, not a reason to keep it on both. The migration rule is simple — once a SKU has cleared a defined volume and return threshold on the cross-border line for two consecutive months, move it, and take it off the cross-border line at the same time. Running the same SKU on both lines during migration is how the two prices surface to the same buyer, and it is also how you end up paying creator commission twice on one order.
Two lists with an owner each is the whole artefact, and it only works if it is the same document your creator operator opens. register with DAMI and ask the team for the Indonesia dual-line SKU split sheet so each SKU carries a line, a migration threshold and a named owner, instead of a decision that lives in someone's head and gets re-litigated every month.
Step 2 — Split the creator pool by line instead of sharing one
Once products are split, split the pool with them, because a creator who promotes both lines is a creator who shows one buyer two prices. Assign each creator a primary line and tag them, then keep the overlap deliberately small: local line carries the head and mid-tier creators who can carry brand proof and conversion, cross-border line carries mid-tier and long-tail creators whose job is testing rather than scaling. Where you do have to reuse a creator across lines, separate them by SKU rather than by week — same creator, different product, so there is nothing for the buyer to compare. The rule that makes this hold is that sample and brief flow from the line the creator is tagged to, so a cross-border test never lands on a creator who is already promoting the local version of the same thing.
Expect the commission conversation to differ by line, and budget for it rather than discovering it. Cross-border stores carry lower traffic weighting and weaker creator willingness, so the working practice is to set cross-border commission two to three points above the local line — 15% locally and 17%–18% cross-border is a common shape — and to stop there rather than escalating. What you get for those points is acceptance, not loyalty: if acceptance stays flat when you move from 15% to 18%, the constraint is the store type rather than the rate, and the fix is better targeting, not more commission. That is also why the two lines are worth running as a test: the same category at 18% on one line and 15% on the other tells you which variable the creator is actually responding to.
Step 3 — Set commission and content rules per line, and keep them apart
Content rules differ by line even when the product is the same, and two of them are now enforced rather than advised. From 31 August 2026 every review video produced in exchange for a free sample has to carry the official commercial partnership label whether or not cash changed hands; unlabelled content is taken down automatically, the linked seller account loses compliance points, and repeated violations restrict affiliate access for seven to fifteen days. Separately, listing images and video have to match the actual product, and a mismatch between a listing photo and a video ad's claims counts as a violation on its own, feeding the Account Health Rating that starts at 200 points and deactivates at zero. The practical consequence for a dual-line seller is that the content package has to be built per line: the local version can lean on domestic-manufacture claims, and the cross-border version cannot claim what the local one does.
Assign the label responsibility to the store, not to the brand. A sample that ships from the cross-border warehouse and a video that posts to the local store is the combination that produces both a labelling failure and a localisation claim you cannot support, and it happens most often when one person runs both stores and picks whichever warehouse has stock. Tie the sample, the brief and the label to one line, and make the line a field rather than a habit.
Step 4 — Record which line each creator belongs to before volume peaks
Record the line on the creator record before volume peaks, because the failure mode is not a bad decision, it is a forgotten one. One field does the work: which line this creator is tagged to, which store their samples ship from, and which commission tier applies. Audit it weekly against three things — creators tagged to both lines, creators whose recent content shows a product from the other line, and creators whose commission was changed without the tag changing. Then check the buyer-facing symptom directly: search your brand and your top five SKUs on the platform and see whether two prices for the same product surface in one scroll. That symptom is the one that costs you the return, and it is only visible from outside your own dashboard.
Two Indonesia-side costs land on both lines regardless of how well you split them, and they belong in the plan rather than in the surprise column. From 1 August 2026 the four named platforms withhold PPh 22 income tax at 0.5% of gross turnover under MoF Regulation No. 37/2025, so the settlement line moves before you change anything operationally. And the local-store commission schedule is not the cross-border one: the February 2026 adjustment moved fashion from 2% to 3% and beauty and personal care from 3% to 4% on local stores while leaving electronics at 2%, with cross-border rates set separately and a further 1.8% mall service fee where it applies. Model both lines with their own rates, or the cheaper-looking line stops looking cheaper the first month it runs.
How DAMI keeps two lines in one pool without mixing them
DAMI brings an 8-million-plus creator database, competitor and competitor-store creator search, automated outreach tasks, AI-assisted outreach, AI video creation, sample management, campaign management, social media management, store analytics, and multi-store team dashboards into one place. Three actions keep two lines in one pool without mixing them, and each ends in a record rather than in a rule someone has to remember:
Bind both stores to one creator library and tag by line
Manage the cross-border store and the local store in DAMI's multi-store collaboration under one account, with the creator library shared and each creator tagged to a line. The output is one pool with no duplicate outreach, and a line field that decides which store a sample and a brief go out from.
Run the two lines as separate outreach tasks with their own commission
Create outreach tasks per store in DAMI's automated tasks, each with its own commission tier and its own product set, so a cross-border test never lands on a creator tagged to the local line. The output is two pipelines that share a library without sharing a brief.
Compare acceptance and GMV per creator across the two lines monthly
Use DAMI's targeted plan management to review acceptance rate, content shipped and attributed GMV per creator by line each month. The output is a table that tells you whether the commission gap between lines is buying acceptance or just costing margin, and which SKUs are ready to migrate.
Frequently asked questions
Can a cross-border seller qualify for the fee discount at all?
Not as a cross-border seller. The regulation applies to verified micro and small enterprises registered in Indonesia selling domestically made goods, applied for through Sapa UMKM, and cross-border status disqualifies a seller regardless of order volume. The realistic route is to localise part of the catalogue under a local entity and keep the rest on the cross-border line.
Should I shut the cross-border store once the local one is running?
No, and treating them as sequential is the expensive mistake. The cross-border line is where you test new products cheaply and keep imported assortment that would cost you the discount if it sat in the local store. Keep both, and give each SKU one line rather than closing one.
How much higher should cross-border commission be?
Two to three points above the local line is the working shape — 15% locally and 17%–18% cross-border. Stop there rather than escalating: if acceptance does not move when you go from 15% to 18%, the constraint is the store type rather than the rate, and better targeting beats a bigger number.
Do I need separate content for each line?
Yes, and for two reasons. Claims differ: the local version can lean on domestic manufacture and the cross-border version cannot. Labelling differs: from 31 August 2026 every sample-based review video must carry the official commercial partnership label, and unlabelled content is taken down automatically.
Can I track which line each creator belongs to in a shared sheet?
For the first month, yes. After that the sheet stops holding: it does not stop a sample shipping from the wrong store, it cannot tell you that a creator's recent content shows the other line's product, and it records the tag once rather than per campaign. The failure is not that the sheet is wrong — it is that during a volume peak the tagging is the step that gets skipped, and the sheet still shows last month's assignments.
Conclusion
Layout is an allocation problem: the right SKU and the right creator in the right field, and everything above is about making that split an explicit decision rather than an accident. DAMI brings an 8-million-plus creator database, competitor and competitor-store creator search, automated outreach tasks, AI-assisted outreach, AI video creation, sample management, campaign management, social media management, store analytics, and multi-store team dashboards into one place.
Search your brand and your top five SKUs on the platform and scroll once. If two prices for the same product surface in that scroll, write down which creators appear on both, tag them to one line today, and stop the sample flow from the other. Fifteen minutes, and it finds the overlap before your next return spike does.
Start with DAMI and put the split where the brief is written rather than where the platforms are compared.
