On 25 October 2026, Indonesia’s Ministry of MSMEs announced that micro and small businesses selling domestically produced goods will get at least a 50% permanent reduction in platform fees. To qualify, a seller has to register in the government’s Sapa UMKM system and pass review. Ready-to-eat food and locally produced electronics are excluded. The announcement was reported by Chuwang (chwang.com/news/211454157094, 26 October 2026); fee mechanics on Shopee Indonesia vary by category and seller entity, so treat your seller centre statement as the authority rather than this post.
Read the headline quickly and it sounds like your local competitor just got 50% cheaper. It is not a 50% event. Platform fees on Shopee Indonesia typically run 2–5% of price, so halving them returns 1–2.5 points of margin. If a subsidised seller passes the whole saving into price, their break-even floor drops about 2.3%. The rest of the gap between you and them — the part that actually decides who wins the buy box — has nothing to do with this policy and was there last month. Separating those two numbers is the entire decision: it tells you whether to reprice this week or do nothing at all.
Turn it into arithmetic
Take one unit of the same physical product, sold by two sellers into the same Indonesian category. The figures below are illustrative — pull your own landed cost and your own fee rate from the seller centre rather than trusting mine.
| Per unit (IDR) | Cross-border seller | Registered local MSME |
|---|---|---|
| Selling price | 100,000 | 100,000 |
| Landed cost | 62,000 | 52,000 |
| Platform fee | 4,000 (4%) | 2,000 (2%) |
| Gross profit | 34,000 | 46,000 |
| Gross margin | 34.0% | 46.0% |
The fee line differs by 2,000 rupiah. The cost line differs by 10,000. One of those numbers is five times the other, and only the smaller one was announced on 25 October.
Now convert both into the only figure that sets price behaviour: the floor you can sell at and still keep a target margin.
Floor price = Landed cost ÷ (1 − fee rate − target margin)
At a 10% target margin:
| Seller | Fee | Landed cost | Floor price | Gap vs cross-border |
|---|---|---|---|---|
| Cross-border | 4% | 62,000 | 72,093 | — |
| Registered local MSME | 2% | 52,000 | 59,091 | −18.0% |
An 18-point floor gap is enough to lose every price comparison in the category, and it is tempting to blame the announcement. Decompose it instead. Hold landed cost equal at 62,000 and change only the fee from 4% to 2%: the floor moves from 72,093 to 70,455, which is 2.3%. So of the 18.0% gap, 2.3 points — about 13% — comes from the policy. The other 15.7 points, roughly 87%, is landed cost, and it predates the policy by however long you have been importing.
There is a second formula worth memorising, because it settles the “should I match?” question in one line. A seller whose fee falls by Δf, holding target margin m constant, can fund a price cut of:
Cut% = Δf ÷ (1 − f_new − m)
The denominator is always below one, so the price cut a fee reduction can fund is always larger than the fee reduction itself. Every time, for any margin above zero. That asymmetry is the whole story: they save Δf, you must spend more than Δf to neutralise it.
| Target margin | Cut fundable by a 2-point fee saving | Their saving per 100,000 order | Your cost to match |
|---|---|---|---|
| 5% | 2.15% | 2,000 | 2,150 |
| 10% | 2.27% | 2,000 | 2,270 |
| 15% | 2.41% | 2,000 | 2,410 |
| 20% | 2.56% | 2,000 | 2,560 |
| 25% | 2.74% | 2,000 | 2,740 |
You spend 8% to 37% more than the competitor gained, in absolute rupiah, at every margin band. Matching is a negative-sum trade by construction, not by estimate.
And the full match is not merely expensive — it is impossible. At 59,091 the cross-border seller in this example is taking 62,000 of landed cost against 59,091 of revenue. That is a 2,909 rupiah loss per unit before any fee. There is no price at which you both match the subsidised seller’s floor and stay solvent, which means “match the price” was never an available strategy. It only looks available because the two gaps were being read as one.
The counter-intuitive turn
The reflex response to this announcement is to protect rank by shaving price. Three separate things are wrong with it.
First, the number is too small to be worth the move. A 2.3% cut on a 100,000 rupiah SKU is 2,300 rupiah. Most sellers reprice in steps of 5,000 or 10,000, so the mathematically correct adjustment rounds to zero or overshoots by a multiple. You cannot execute the right number, and the nearest executable numbers are both wrong.
Second, the policy does not change today’s price — it changes who can survive a price war tomorrow. A registered local seller sitting at 46% margin can cut to 30% and still be comfortable. You at 34% cannot cut to 30% without surrendering most of your profit. So the announcement moved the distribution of pain tolerance, not the price. Pain tolerance only becomes price when someone decides to start a war, and that decision is observable: it shows up as a listing that actually moved after 25 October. Most categories will not see one, because most registered sellers would rather bank 2 points of margin than start a fight they do not need.
Third, the eligible set is narrower than the headline. Registration in Sapa UMKM plus a review is a gate, not a formality, and ready-to-eat food and locally produced electronics are excluded outright. If you sell in those two categories, this month changed nothing for you and repricing would be a pure loss. If you sell elsewhere, you still do not know which specific competitors cleared the gate — and assuming the whole category got 50% off is how sellers end up cutting price against a subsidy that was never granted.
There is a diagnostic habit underneath all of this. When margin falls and no competitor’s price moved, the cause is on your cost side: a fee change, a duty change, an ad-cost change. Check the invoice before the price list. A cost event that has not been transmitted into price is, for pricing purposes, not an event — and treating it as one is the expensive mistake. Use smart follow: follow, hold or raise only after the price series shows an actual move; the three inputs it reads are price gap, sales increment and rating, and the price gap is the one that has to move first.
Where we sit in this
PriceMinder monitors Shopee only — ID, VN, TH, PH, MY, SG and BR — on a fixed schedule. Free gets 2 checks a day and 5 products; Pro gets 4 checks a day and 50 products, with 30-day price history against 7 on Free. Smart Follow reads three inputs, price gap, sales-increment and rating, and returns Follow, Hold or Raise with a 14-day review. Alerts reach desktop and Feishu on Free, plus Telegram and Slack on Pro, and there is an MCP with 22 tools.
The limitation matters more than usual here, so state it plainly: we do not read your fee invoice, we do not know which competitor holds Sapa UMKM registration, we do not have a live feed, and we never touch your prices. Two of the three numbers in this article — your landed cost and your fee rate — you have to supply yourself. What we supply is the third: the price history that tells you whether a subsidised competitor actually transmitted the saving into their listing, or quietly banked it. See pricing and plans for the history window, and how the fixed check schedule works for what it catches and what it misses.
Numbered action checklist
- Today: list which of your Indonesia SKUs sit under 50,000 rupiah and which fall in the excluded categories. Excluded means this policy is not your problem this month, and repricing those SKUs destroys margin for nothing.
- Today: compute your landed cost per SKU in rupiah, not percent. Include freight, duty and the FX you actually settled at. This is the number that carries 87% of the gap and the one most sellers have never written down.
- Tomorrow: for each contested SKU, calculate the floor a registered local seller would hold at a 2% fee and a 10% margin. If that floor sits below your landed cost, the SKU is structurally lost and no pricing decision fixes it — decide differentiate, resource, or exit.
- This week: pull the price history of your three closest local competitors for the two weeks around 25 October. You are looking for movement, not for a level. A flat line means the saving was banked and your correct action is none.
- This week: if a competitor did move 2–3%, still do not follow. You would spend more than they gained. Hold, and compete on the non-price attributes the policy does not touch.
- This month: attack landed cost, because that is where the gap lives. Consolidate freight, change the sourcing route, or shift mix toward categories the policy excludes. A one-point landed cost improvement is worth more than the entire fee halving.
- This month: re-check eligibility quarterly. Registration is open-ended, so more competitors can qualify over time and the subsidised set can widen without any new announcement. Use multi-market coverage and the repricing math post to keep the fee-stack assumptions current.
FAQ
Did Shopee Indonesia really cut seller fees by 50%? Not for everyone, and not by 50% of your price. Indonesia’s MSME ministry announced on 25 October 2026 that micro and small businesses selling domestically produced goods get at least a 50% permanent reduction in platform fees, after registering in the Sapa UMKM system and passing review. Ready-to-eat food and local electronics are excluded. The 50% applies to the fee, which is typically 2–5% of price, so it returns 1–2.5 points of margin, not half your costs.
Should I cut my prices to stay competitive? No, and the arithmetic is the reason. A competitor whose fee halves from 4% to 2% can fund a 2.27% price cut at a 10% target margin. To match that 2.27% you give up 2,270 rupiah per 100,000 rupiah order to neutralise 2,000 rupiah of their saving. You always spend more than they gained, because the cut is divided by a number below one.
How do I know whether a competitor actually passed the saving into price? You cannot tell from a single snapshot. A seller who pockets the 2 points and a seller who cuts 2.3% look identical on the day of the announcement and diverge over the following two weeks. You need their price history across that window, which is why the useful question is not what a competitor charges today but whether their listing moved after 25 October.
Does PriceMinder tell me which competitor is subsidised? No. We track Shopee listing prices and your own price history across seven markets, 7 days on Free and 30 days on Pro. Whether a rival holds Sapa UMKM registration is a government record we do not read, and we do not see your fee invoice either. What we give you is the price series to check the hypothesis against, not the registration status behind it.