Thailand’s Trade Competition Commission approved a draft regulation on 21 October that will force large platforms to explain their fee structure and justify each increase before it happens. Buried in the announcement was a number worth more attention than the regulation itself: Shopee has raised its service fees 10 times since 2022 — a cumulative increase of about 433%. Lazada did it 7 times, about 367%.
Sellers felt that. Nobody itemised it.
That is the real problem with a fee increase: it arrives as a percentage, and percentages are easy to nod at and impossible to feel. So here is the arithmetic, in one order, before we talk about what to do with your price.
What the fee stack actually costs
Take a 100 MYR order on Shopee Malaysia, cross-border direct shipping. Rates vary by category and market — this is one worked example, not a quote for your store:
| Cost line | Rate on a 100 MYR order | MYR |
|---|---|---|
| Commission | 18.36% | −18.36 |
| Transaction / payment fee | 3.78% | −3.78 |
| Tech support fee (added Feb 2026) | 5.00% | −5.00 |
| Platform take | 27.14% | −27.14 |
| SLS shipping, 500g parcel | — | −11.00 |
| Product cost | — | −30.00 |
| Cash contribution | 31.86 |
Two things stand out. First, the platform take is 27% before logistics, and 38% after. Second, the 5% tech support fee added in February 2026 is not commission — it applies on top, and it is the line most sellers still forget when they recalculate margin.
Now change one variable. Suppose commission moves up 1.5 percentage points, which is within the range being discussed across MY / ID / TH. To keep the same 31.86 MYR you must sell at:
31.86 + 11.00 + 30.00 = 72.86 fixed-and-target → 72.86 ÷ (1 − 0.2864) = 102.10 MYR
A 1.5-point fee increase demands a 2.1% price increase just to stand still. Miss it and you are not “slightly less profitable” — you are giving away 1.5 MYR per order, which on 2,000 orders a month is a colleague’s salary.
The trap: your competitor’s fee bill is not yours
Here is where the arithmetic stops being arithmetic.
Your competitor may be on a different category rate, a different fulfilment model (official overseas warehouse sits at 15.12% commission but drops shipping time from 7–15 days to 2–4), or a different market. Their break-even is not your break-even. So when fees rise:
- If they raise and you don’t — you look cheaper, volume moves to you, and you absorb a margin loss on every unit of it. Growth that loses money is not growth.
- If you raise and they don’t — you lose the click, the conversion, and the ranking that came with both. Your 2.1% is now a 2.1% haircut on demand.
- If both raise — the market re-prices and nothing happens to share.
You cannot answer this from your own price sheet. You need the gap, and the gap changes daily. This is exactly why “watch my own price and reprice when fees change” fails: it treats a two-player problem as a one-player decision.
Three inputs, not one
The instinct is to reprice on price alone: competitor drops 3%, I drop 3%. In a fee-increase cycle that instinct is actively harmful, because a competitor’s lower price may simply mean they are absorbing the new fee while sitting on stock they need to clear.
The decision needs three inputs read together:
| Input | What it tells you | What it changes |
|---|---|---|
| Price gap | How far you sit from the cheapest credible rival | Whether you can raise at all |
| Sales increment | Whether the category is accelerating or stalling | Whether holding share is worth a margin hit |
| Rating | Whether buyers still prefer you at a higher price | How much premium the gap tolerates |
Read together they produce three calls, and only three:
- Follow — competitor cut, category is growing, your rating is at parity. Match, and keep the volume.
- Hold — competitor cut but their velocity is not picking up, or your rating is materially higher. Do nothing; their discount is buying them nothing.
- Raise — you are the cheapest by a wide margin with no conversion advantage to show for it, or a fee increase landed and your gap gives you room. Take the 2%, keep the volume.
The failure mode of every repricing tool is collapsing those three inputs into one. Price alone generates Follow calls on a Hold situation, which is how sellers end up in a margin spiral during a fee cycle.
Where we fit
We built PriceMinder for exactly this decision, and it is fair to say plainly what it does and does not do.
It watches the gap for you. Add a competitor product from any Shopee page with the browser extension; it reads the site and product info automatically across ID / VN / TH / PH / MY / SG / BR. Free covers 5 products with 2 checks a day (morning and afternoon); Pro covers 50 with 4 checks a day across morning, afternoon, evening and night. We publish that schedule instead of claiming “real-time” — under Shopee’s risk controls nobody can honestly promise more.
It tells you when something moved. A competitor price change past your threshold (3% by default) pushes a notification to Feishu, Telegram, Slack or the desktop, depending on plan.
It makes the three-input call. Mark which of your monitored products you actually sell, map the competitors to them, and Smart Follow reads price gap, sales increment and rating together, returning Follow / Hold / Raise with the reason attached. Fourteen days later it shows you whether the call was right — which is the part most repricing tools skip, and the only part that makes you better at this.
It answers questions in plain language. The MCP server exposes the same data to Claude, Cursor and other AI clients, so “which competitors dropped more than 4% this week, and by how much does that undercut me after the new 5% fee” is a sentence you can type instead of a spreadsheet you have to build.
What it does not do: it does not edit your listings, and it does not touch Lazada or TikTok Shop. Price changes stay in your hands — on a fee-increase cycle that is a feature, not a limitation.
A seven-day sequence
If a fee change lands on you this quarter, this is the order that avoids panic:
- Recalculate your take-home. Pull the actual rates from your Seller Centre for each category you sell, not the headline rate. Include the 5% tech fee and the ~2% payment fee.
- Compute the stand-still price. Extra fee ÷ (1 − your total variable rate). That is the number a fee increase costs you, expressed as a price.
- Map your real competitors. Not “the category” — the five listings a buyer actually compares you against, per SKU.
- Read the gap, not the price. Note your position against the cheapest credible rival, before and after their last move.
- Check velocity. If the category is decelerating, absorbing the fee to hold share is usually wrong; if it is accelerating, it is usually right.
- Make one call per SKU — Follow, Hold or Raise — and write down why. The reason is what you review later.
- Review in 14 days. Not to feel good about the call, but to find out which input you over-weighted.
Fee increases are now a recurring event, not an incident. Thailand’s regulation, if it takes effect as drafted by 18 December, will at least make the next one explainable in advance — which turns a surprise into a scheduling problem. That is a much better thing to have.
- See the plans — what 2 checks a day versus 4 actually covers
- How Smart Follow decides — the three-input model in detail
- MCP setup in ten minutes — ask your own price data questions in Claude or Cursor