On 4 November 2026, Shopee Vietnam moved Special SPayLater from one store-wide setting to a per-product setting. Each SKU can now be switched on or off individually, and each SKU can carry its own installment tenure. The 5 November 2026 Chuhaiwang report, citing the Seller Centre guide, notes there is no transition period. Sellers who were covering an entire catalogue with a single installment plan woke up to a catalogue where every listing has to be recosted.
The installment cost previously averaged across the store is now visible per SKU, and averaging is exactly the mechanism that was hiding your losers. If a 0% installment plan costs 4-5% of order value, a VND 10,000,000 order carries VND 400,000-500,000 of subsidy cost. Against a 30% gross margin, that is 13% to 17% of the profit on that product line — previously diluted into a store-level average nobody could act on.
Turn it into arithmetic
Assume, for illustration only, a Vietnam order at VND 10,000,000 with a landed cost of VND 7,000,000, so gross profit is VND 3,000,000, or 30.0%. This is our assumption to make the arithmetic visible. It is not a rate from your store and not a Shopee statistic. Substitute your own landed cost.
The formula is one line:
Share of profit consumed = (installment cost rate × order value) ÷ gross profit
With the 4-5% assumption plugged in:
| Line item | Low case 4% | High case 5% |
|---|---|---|
| Order value | VND 10,000,000 | VND 10,000,000 |
| Installment cost | VND 400,000 | VND 500,000 |
| Gross profit (30%) | VND 3,000,000 | VND 3,000,000 |
| Share of profit consumed | 13.3% | 16.7% |
| Resulting gross margin | 26.0% | 25.0% |
Read the last row. A 30% margin becomes 25-26% on the SKU carrying the plan, while the SKU beside it in the same store, plan switched off, stays at 30%. Under the old setting those two numbers were blended into one, and the high-ticket product quietly subsidised the catalogue. That subsidy is now legible, which makes it a decision you have to take explicitly instead of one the platform takes for you.
One more layer most sellers skip. Say the competing listing on the same model sits at VND 9,500,000, 5% below you, and you still carry the 5% installment cost. Your combined disadvantage is 10% of order value, VND 1,000,000, against VND 3,000,000 of gross profit — a third of the line’s profit. The buyer never sees the gap as a percentage. They see the monthly payment.
The decision rule falls out of the table rather than out of instinct:
- Absorb only when the post-installment margin stays above your floor. With a 20% floor, a 30% margin survives a 5% cost (it lands at 25.0%). A 22% margin does not (it lands at 17.0%).
- Raise only after checking whether the competitor has already moved. A lone 5% increase hands the cart to the seller who did nothing.
- Opt out on the SKU where the installment cost exceeds the margin you can defend. Per-product switching exists so you can say no to one listing instead of the whole store.
That 4-5% comes from the 5 November 2026 Chuhaiwang report on interest subsidy, funding cost and bad-debt risk. It is an assumption we set to make the arithmetic runnable, not a Shopee-published rate. Your real figure is in the setup guide in Seller Centre.
Why a store-wide markup now fails in both directions
The failing playbook is the flat store-wide markup: add a few points to every listing to cover payment cost, then stop thinking. That worked while SPayLater was one setting for the store, because the cost really was roughly uniform. It is now wrong in two directions at once.
Take a two-SKU store. SKU A is high-ticket at VND 10,000,000 with a 30% margin. SKU B is fast-moving at VND 200,000 with a 12% margin. Under the old rule you add 4% across the board: VND 400,000 on A, VND 8,000 on B. Against a blended store-wide cost, that was fine.
Now the setting is split. A goes on a six-period plan costing 5%, or VND 500,000. B gets no plan. The flat 4% now undercharges A by VND 100,000 per order and overcharges B by VND 8,000. If A ships 100 orders a month, that is VND 10,000,000 of gross profit never collected. The overcharge on B is not free money either: it is a price disadvantage against the competitor who correctly left B out of the programme, and B is exactly the kind of low-ticket SKU where a 4% price difference changes conversion.
So the flat markup undercharges where the cost is and overcharges where it is not. Per-product configuration did not create this problem; it removed the averaging that was papering over it. Keep the flat markup and you will spend next quarter confused about why the high-ticket lines got thinner and the low-ticket lines went soft.
The second failing playbook is pricing first and checking competitors later. Whether you can pass the cost through depends on whether your competitors have finished the same recalculation. Raise 5% alone and you have given 5% to whoever stood still. Check the gap first. That is also why Smart Follow treats the 14-day review as a fixed rhythm instead of a one-off decision: follow, hold or raise is a call you re-make, not a switch you flip.
What the per-product setting actually gives you
Before the change the choice was binary at store level: enrol or not, one tenure for everything, and the platform reported a single blended number, so neither distortion showed up on a fee statement.
Now the granularity is per SKU, which gives the decision a shape: which listings gain more in conversion than they lose in subsidy. That is a question about the buyer. A VND 10,000,000 appliance converts on monthly payment; a VND 200,000 consumable converts on unit price, and the plan mostly buys it a signal that the item is expensive.
Two paths exist in Seller Centre, and they differ in maintenance cost, not strategy. Under SPayLater 0% Xtra in the marketing centre, non-enrolled items sit in a Not Enrolled group and activate with a chosen tenure, while enrolled items sit in an Enrolled group where you change tenure or opt out. Or you work listing by listing from the product page, setting Special SPayLater to a tenure or to No. The first is faster for a bulk review; the second is what you use when a price move and a plan change must land the same day. Either way, what triggers a revisit is a competitor move — the same 30-day curve logic as the rebate price-lock case.
Where we sit in this
PriceMinder monitors Shopee only, across seven markets: ID, VN, TH, PH, MY, SG and BR. Collection runs on a fixed schedule, not continuously: 2 checks per day on the free plan (morning and afternoon), 4 per day on Pro (morning, afternoon, evening, night). Free accounts get 5 products and a 7-day trend; Pro gets 50 products and a 30-day trend. Smart Follow reads three dimensions, price gap, sales increment and rating, and returns follow, hold or raise with a 14-day decision review. Alerts go to desktop and Feishu on the free plan, plus Telegram or Slack on Pro.
The limitation, stated plainly: no continuous collection, we do not change your prices in any form, other platforms out of scope. That is deliberate. A per-SKU installment decision needs one price move made after you have seen the gap, not a machine repricing against a competitor whose plan tenure it cannot see. Four checks a day catches a listing that repriced once in a 14-day window, and that is the cadence the decision runs on. Quotas are on the pricing page, and yesterday’s moves are summarised in the daily price brief.
A five-step checklist
- Today: cost out every SKU carrying a plan. Pull your top 20 SKUs by revenue and write the installment cost next to each, using Seller Centre numbers. Mark red anything where installment cost exceeds 10% of gross profit, which on the example above means above 3% of order value at a 30% margin.
- Today: write the floor down. Set the minimum post-installment margin you accept per SKU. SKUs that break it go on the opt-out list, not the raise list — conflating the two is how a price increase kills a product that simply should not have offered installments.
- This week: pull 30-day curves on your top 10 competitors. Two signals matter: a price move in the last 7 days, and whether it coincided with a tenure change. A competitor dropping hard may be shortening tenure to clear stock, which is a different threat from a cut on a stable plan.
- This week: set one price line. Floor is cost plus installment cost plus minimum margin; ceiling is the competitor median. SKUs outside the band do not get touched this cycle, which is most of them.
- This month: review at day 14. Record how many points of conversion the opted-out SKUs lost and how many of volume the raised ones lost, so the next cycle runs on data instead of on how last week felt.
FAQ
What is the first thing I should do now that SPayLater is per-product?
Cost out every SKU that carries an installment plan. On the worked example, a VND 10,000,000 order at 30% gross margin with a 4-5% installment cost loses 13.3% to 16.7% of its profit. Flag the SKUs where installment cost exceeds 10% of gross profit, then decide absorb, raise or opt out for each one individually.
Competitors have not moved. Can I add the installment cost to my price first?
Not on your own. The two-SKU example shows a flat 4% markup undercharges the high-ticket SKU by VND 100,000 per order and overcharges the fast-moving SKU by VND 8,000. Check whether the competitor’s 30-day price curve and plan tenure have already changed before you move, or you are handing them the buy box.
Can you watch the exact moment a competitor changes their installment plan?
No. We cover Shopee only, on a fixed schedule: 2 checks per day on the free plan and 4 per day on Pro. There is no continuous checking and we do not change your prices. Plan tenure changes usually arrive with a price change, and four checks a day is enough to catch a listing that repriced once in a 14-day window.
Is the 4-5% installment cost a Shopee-published rate?
No. It is an assumption we use to make the arithmetic visible, drawn from the 5 November 2026 Chuhaiwang report discussing interest subsidy, funding cost and bad-debt risk. It is not your store’s rate. What you actually pay is in the Special SPayLater setup guide in Seller Centre, and that version is the one that counts.