# TikTok Shop Thailand: What COD Refusals Cost and How to Cut Them
If you run a beauty brand on TikTok Shop Thailand with a team of ten to fifty people, you have likely already met the line item that quietly decides whether the market works for you. It is not commission, it is not shipping, and it is not creator fees. It is cash on delivery refusals.
Thailand is a COD-heavy market. A large share of buyers will not complete a purchase unless they can pay the courier at the door, which means COD is not a checkout preference you can decline to support. It is the entry ticket. And the cost of that ticket is the refusal rate — the parcels that travel all the way to a buyer’s address and then travel all the way back.
Most sellers entering Thailand budget for shipping and commission and treat refusals as an afterthought. This article is about putting a number on a refusal, understanding what causes them, and cutting them before the parcel leaves the warehouse.
Why COD is not optional here
In markets where cards dominate, COD is a minor option that a small minority of buyers use. In Thailand the proportions are closer to reversed for a meaningful share of the buying population.
Three reasons, and all of them are structural rather than temporary.
Card penetration. A large portion of Thai buyers do not hold or use credit cards for online purchases. Debit infrastructure is more widespread, but online card usage is still far below Western European or North American levels.
Trust in the transaction. Paying on delivery removes the risk of paying for something that never arrives or does not match the description. In a market where marketplace fraud has been widely reported, that protection is worth more than convenience.
The unbanked and underbanked population. A meaningful share of Thai adults are unbanked or underbanked. For them, COD is not a preference — it is the only way to buy online.
The consequence for a beauty brand: if you disable COD to avoid refusals, you are not reducing risk, you are removing most of your addressable market. The correct response is to keep COD and reduce refusals.
What one refused parcel actually costs
The intuitive estimate is “we lost the shipping fee.” The real number is larger, because a refusal triggers costs in both directions plus a write-off on the goods.
Here is the structure for a mid-priced beauty item.
| Cost line | One-way (THB) | On refusal (THB) |
|---|---|---|
| Outbound delivery | 35 | 35 |
| Return delivery | — | 35 |
| Payment collection fee | 12 | 0 (not collected) |
| Product cost | 180 | 180 (usually unsellable) |
| Repack and handling | 8 | 14 |
| Total per refusal | 235 | 264 |
The product cost is the part teams underestimate. A cosmetic item that has travelled out and back in Thai heat is generally not resellable — the seal is compromised, the packaging is scuffed, and the contents have been through temperatures the formulation was not tested against. Treat it as a write-off in the model, not a recoverable return.
Now scale it. At a 12 percent refusal rate on 1,000 COD orders a month, that is 120 refused parcels, roughly 31,000 THB a month in direct cost, before counting the customer service time and the delivery-performance damage.
The three things that make a buyer refuse
Refusals are not random. They cluster around three causes, and each has a different fix.
Delivery took too long. The buyer ordered something they wanted in a few days, it took two weeks, and by the time it arrived the need had passed. This is the most common cause and the most fixable.
The reality did not match the listing. The shade was different, the size was smaller, the texture was not what the video implied. For beauty this is the single most expensive category, because colour and texture are exactly what video distorts.
The order was never firm. COD removes the payment moment, which means some buyers place orders with no real commitment. This segment cannot be eliminated, but it can be reduced by who you sell through.
Notice that only the third cause is genuinely a COD problem. The first two are inventory and content problems that COD merely exposes faster.
The 7-day delivery cliff
Cross-border sellers hit a specific threshold: once delivery exceeds roughly seven days, refusal rates climb sharply, commonly moving past 20 percent.
The relationship is not linear. Somewhere around the one-week mark, buyers stop waiting for the parcel and start treating it as something that may or may not arrive. After that point they have mentally moved on, and the parcel arriving is an inconvenience rather than a delivery.
What this means operationally:
Under seven days, refusals are mostly the mismatch and non-commitment categories. Content and creator selection are the levers.
Over seven days, refusals are dominated by the delivery itself. No amount of listing accuracy fixes it, because the buyer is refusing something that simply arrived too late.
If you are shipping cross-border into Thailand and cannot get under seven days consistently, the honest conclusion is that refusal rate is a logistics decision, not a marketing one. Either move inventory into the country or price the refusal rate into the unit economics and check whether the market still works.
The three-refusal rule, and who it actually hurts
The platform tracks buyer-side refusal behaviour: a buyer who refuses three deliveries within 60 days loses COD access for 60 days.
Sellers sometimes read this as protection. It is, but the protection is slow relative to your losses.
Consider the arithmetic. A buyer has to refuse three parcels before the penalty applies. If three different sellers each absorb one refusal from that buyer, the buyer’s third refusal is somebody else’s first. The penalty lands after three sellers have already paid the return shipping and written off the goods.
So treat the rule as a backstop that eventually removes the worst offenders from the COD pool, not as a mechanism that protects your margin on any individual order. Your protection has to come from the pre-shipment steps.
Cutting refusals before the parcel leaves
Five levers, roughly in order of impact.
Get delivery under seven days. Everything else is secondary. If your lanes cannot do it, fix that first. TikTok Shop Shipping Time: How Delivery Speed Decides Whether Creator Sales Repeat goes at this in detail.
Make shade and texture honest. For beauty specifically: show the product on the skin tone it is intended for, state the finish in words, and do not rely on a filtered demo to communicate colour. Most mismatch refusals trace back to a video that flattered the product past the point of accuracy.
Confirm the order before dispatch. A short confirmation message or call on higher-value COD orders converts a soft commitment into a firm one, and it is far cheaper than a return leg.
Choose creators by selling record, not follower count. This is the lever that most directly reduces the non-commitment segment — see the next section.
Set an internal refusal threshold by SKU. If a product’s refusal rate runs materially above your category average, stop shipping it COD and either fix the listing or drop it.
Using creator data to find where refusals come from
Refusals are not evenly distributed. Some creators generate orders that arrive and get paid for; others generate orders that arrive and get refused. Telling them apart is a data question, and the relevant data already exists.
Filter creators by what they actually sell. When building a shortlist in DAMI, filter by content preference — shopping videos, shopping livestreams, or no stated preference — alongside one of six follower bands: under 10k, 10k-20k, 20k-25k, 25k-50k, 50k-100k, or 100k+. A creator categorised for shopping video is producing commerce content, which is a different audience from an entertainment account that happens to have an affiliate link.
Check GPM, not views. DAMI’s video detail view reports gross merchandise value per thousand views. Two creators can deliver the same view count with very different revenue per view, and the low-GPM one is disproportionately likely to be generating soft orders.
Verify the real number. DAMI’s product performance view reports revenue after refunds, taxes, and shipping costs. A SKU’s headline GMV can look healthy while its contribution is negative once return and refusal costs are included. Check the deducted figure before deciding a product is working.
Tag the creators you learn something about. DAMI’s creator resource management supports multi-dimensional tagging, which is what turns a one-month observation into a permanent filter. Tag creators by whether their orders get paid for. After two or three months, that tag becomes the fastest shortlist you have — faster than any follower-count filter.
The pattern to look for is a creator whose views are fine but whose GPM sits below your account average. That is usually an entertainment-led audience arriving at a product they did not intend to buy — which is precisely the order that gets refused at the door.
One more Thai-specific warning: the biggest follower counts are not the safest bet here. TikTok Shop Thailand: Why Big Creators Waste Your Sample Budget makes the case for mid-tier creators in this market specifically.
Heat, humidity, and festival peaks
Thailand’s climate adds a category-specific cost that sellers from temperate markets routinely miss.
Heat during transit. Parcels sit in non-climate-controlled vehicles and sorting facilities. Formulations that separate, melt, or degrade at elevated temperatures arrive compromised even when the buyer accepts delivery.
Humidity and packaging. Cardboard softening and label degradation in transit increase both damage rates and the perception of a mishandled product — which feeds refusals even when the contents are intact.
Leakage risk. Liquids, creams, and serums under pressure and heat leak. A leaking parcel is refused on sight, and it also damages the courier relationship.
Practical responses: specify packaging rated for the temperature range, add secondary containment for liquids, and test ship your own products in peak season before scaling. This is one of the few areas where spending more per unit reduces total cost.
Two calendar peaks dominate the Thai year, and both change refusal dynamics.
Songkran — the water festival in April — drives demand for waterproof and cooling products, and it also disrupts logistics for roughly a week. Refusals rise because delivery times slip during the disruption.
Loy Krathong — in November — drives gifting and beauty demand generally.
The operational rule: tighten your promise before a festival, not during it. Set conservative delivery estimates in the two weeks leading into the peak, because the cost of over-promising during a logistics disruption is a refusal spike you cannot recover within the quarter.
If you want the automation side of running this at a small team size, the product page covers creator discovery, scheduled outreach, and the reporting views referenced above.
FAQ
Should I turn off COD to avoid refusals? Almost certainly not. Thailand’s card penetration is low enough that disabling COD removes a large share of your addressable buyers. Keep COD and work on the causes instead.
What is a normal COD refusal rate in Thailand? It varies significantly by category, price point, and delivery speed. The more useful approach is to measure your own rate by SKU and watch the spread, rather than compare against a benchmark that may not match your mix.
Does a refused parcel cost more than the shipping fee? Yes. A refusal costs the outbound leg, the return leg, the handling on both ends, and usually the full product cost, since beauty items that have travelled in transit conditions are generally not resellable.
How does the buyer-side refusal penalty work? A buyer who refuses three deliveries within a 60-day window loses COD access for 60 days. It helps over time, but it does not protect you on any individual order.
What is the single biggest lever on refusal rate? Delivery speed. Cross-border shipments that exceed roughly seven days see refusal rates climb sharply, commonly past 20 percent, and no amount of listing accuracy compensates for a parcel that arrives too late.
