You approved a sample. The colour was deep green, the aroma was sweet and marine, the shot pulled clean on the espresso-style dosing spoon your barista trainers use. Eight weeks later the same lot number in your Bangkok distribution centre is yellowing at the edges, the powder cakes in the bag, and your QA lead is writing an email to the supplier in Japan asking what changed.
Usually nothing changed in Uji. What changed is everything between the Japanese warehouse door and the back-of-house shelf in your outlet in Thonglor.
This article is for sourcing and QA managers at Thai café chains, bakery and beverage manufacturers, and Bangkok importers who already buy Japanese matcha and are dealing with a concrete symptom rather than a definition. It does two things: it gives you a way to locate where in your chain the damage happens, and it gives you the specific lines to put into your next purchase order so the damage stops happening.
The short version of the claim
Matcha degrades through three mechanisms that all accelerate with heat and moisture: oxidation of the catechins and lipids, loss of volatile aroma compounds, and chlorophyll breakdown that shifts the colour from green toward yellow-brown. None of these are switches. They are cumulative. A powder does not "go bad" on a date; it spends a certain number of day-degrees in a certain humidity and arrives at your counter with whatever quality is left.
That means the useful question is not "is this supplier good" but "how much exposure did this lot accumulate, and at which handover." In Thailand there are four handovers that compound in a way importers in cooler, drier markets never experience simultaneously:
- Storage and staging at the Japanese warehouse before the container closes.
- Ocean or air transit, and in particular non-refrigerated LCL consolidation where your pallets sit in a shared box with whatever else is going to Laem Chabang.
- Clearance in Thailand — customs plus Thai FDA processing — during which goods sit in a port or bonded area, not in a chiller.
- Domestic distribution and outlet back-of-house, which in most Thai café and bakery operations is not air-conditioned.
Two levers close most of the gap, and neither of them is "buy a higher grade." The first is pack size matched to how fast a single opened bag is actually consumed. The second is shifting the risk window: choosing shipping mode and order cadence around your real clearance time, and writing shelf life from the landed date rather than the production date.
Why Thailand specifically
Bangkok's climate is the reason a chain that works for a buyer in Rotterdam or Melbourne fails here. Long-term climatological normals for Thailand are published by the Thai Meteorological Department (tmd.go.th), and you should pull the normals for your own province rather than trusting a national average — the Bangkok metropolitan values differ from Chiang Mai or Phuket. What matters for your worksheet is that you record the actual mean temperature and mean relative humidity for the months your shipments land, from that source, and use those numbers rather than an impression.
Two practical consequences follow from a warm, humid ambient, and neither requires a citation because they are physics you can verify with a thermometer in your own DC:
- Opened bags absorb moisture continuously. Matcha is a very fine powder with enormous surface area. Every time a 1 kg bag is opened for dosing in a humid room, it takes on moisture. Caking, clumping in the sifter, and a flat, hay-like flavour are the visible end of that process.
- Condensation happens on the way into the chiller, not just out of it. If you move a sealed bag from a 32°C room into a 4°C chiller and back out again for dosing, you get condensation on and inside the packaging film every cycle. Repeated in-and-out handling of the same bag is often worse than leaving it at stable ambient.
This second point surprises people. Cold storage helps enormously for unopened, sealed, gas-flushed inventory. It hurts when a single bag is cycled daily. That distinction is what drives the pack-size decision below.
Stage-by-stage: the Bangkok landed-quality worksheet
Before you challenge a supplier, quantify your own chain. Build a simple log with one row per stage and fill it in for one real shipment. You need three columns: elapsed days, temperature, and relative humidity. Where you cannot measure, estimate and mark it as an estimate — an honest estimate is more useful than a blank.
Stage 1 — Japan: production to container close
Ask your supplier, in writing, for two dates per lot: the milling date and the packing date, and separately the date the goods left their warehouse. These are not the same thing and the gap between them is invisible on a standard commercial invoice. Also ask what the storage condition was during that gap — ambient, chilled, or frozen — because Japanese tea companies vary and many hold bulk product cold and only bring it to ambient for packing.
Record: days from milling to departure, and the storage condition during that window.
Stage 2 — transit
This is the stage buyers most often assume is short and most often is not. Note the difference between the three modes:
- Air freight is the shortest exposure and the least temperature-controlled per hour, but the hours are few. For first orders, trial lots, and anything under a few hundred kilograms, the freight premium often costs less than the quality loss you are currently absorbing.
- FCL ocean, dry container. A dry container crossing to Southeast Asia is not a neutral environment. Internal temperatures in a steel box in sun can exceed ambient substantially, and the day-night cycle drives moisture movement inside the box. If you ship FCL dry, ask for desiccant and for the cargo to be loaded away from the doors and walls.
- LCL consolidation is the highest-variance option. Your pallets sit in a CFS warehouse in Japan waiting for the box to fill, share the container with unknown neighbours, and sit in another CFS in Thailand waiting for deconsolidation. The consolidation wait on each end is invisible on the bill of lading. Ask your forwarder for the CFS receipt date and the actual sailing date; the difference is exposure you are paying for without seeing.
Reefer LCL on Japan–Thailand lanes exists but availability and premium vary by lane, season and volume, and any figure quoted here would be stale before you read it. Treat this as a question for your forwarder, asked with a date attached: "for a shipment closing in [month], what is the reefer LCL option and premium for two pallets Japan to Laem Chabang, and what temperature setpoint can you hold?" An insulated liner or thermal blanket inside a dry container is a cheaper middle option and worth quoting alongside.
Record: CFS receipt date, sail date, arrival date, mode, and whether temperature control was in place.
Stage 3 — Thai clearance and FDA
This is the stage that is genuinely Thailand-specific and the one most often left out of a supplier conversation entirely.
Food imported into Thailand falls under the Food Act administered by the Food Division of the Thai FDA under the Ministry of Public Health. The authoritative reference is the Thai FDA food portal (fda.moph.go.th), and customs procedures and tariff classification sit with Thai Customs (customs.go.th). What you need to establish for your own operation, and verify directly with those authorities or a Thai licensed customs broker rather than taking it from an article:
- Which licence or registration route applies to your product as you intend to sell it. The requirements differ depending on whether matcha powder enters as a bulk ingredient for further manufacturing or as a retail-ready packaged food carrying Thai-language labelling. Confirm the route with the Thai FDA before your first PO, not after the container arrives.
- Label requirements and who registers them. Thai-language labelling obligations for retail-ready packaged food are set in Thai FDA notifications; if your plan is private-label retail bags, the label content and its approval sit on your critical path, not the supplier's.
- Documentation your supplier must provide. Certificate of analysis, ingredient and manufacturing statements, and any origin documentation for preferential duty. Ask your broker specifically whether any additional certification is currently required for tea of Japanese origin — requirements in this area have changed over time in several markets, and the only reliable answer is a current one from the Thai FDA or your broker, dated.
- Duty treatment. Japan and Thailand are parties to a bilateral EPA and to the ASEAN–Japan agreement, and preferential rates depend on correct HS classification and a valid certificate of origin. Confirm the HS line your broker intends to use for powdered green tea and the documentation the supplier must issue to support the preference. Japanese standards and JAS documentation references are published by MAFF (maff.go.jp), which is useful when your broker asks what a supplier's certification actually attests to.
The operational point for quality: whatever this process takes for you, in days, is a period during which your matcha is sitting in Thailand at ambient conditions. First-time importers and first-time products take materially longer than repeat shipments of an already-registered item, because registration is a one-time gate that the first shipment waits behind. If you have imported before, pull your own records and calculate your actual median and worst-case days from vessel arrival to release over the last several shipments. That number, your number, is the input for the worksheet — not a benchmark from elsewhere.
Record: arrival date, release date, elapsed days, and where the goods physically sat.
Stage 4 — DC and outlet back-of-house
Put a cheap datalogger on the shelf where the matcha actually lives — in the DC, and in two representative outlets, ideally one high-volume and one low-volume. Run it for two weeks. Most operators discover something they did not expect: the back-of-house cupboard above a fridge compressor, or the dry store next to the kitchen line, is warmer than the dining room by a wide margin.
Record: days of stock on hand at each location, and the measured mean and peak temperature and RH.
Reading the completed worksheet
Add up the days. Most Thai buyers who complete this exercise for the first time find that the powder they serve is considerably older, and has spent far more time above 25°C, than they assumed when they approved the sample. Then look at where the biggest block of warm days sits. If it is stage 2, change mode or add temperature control. If it is stage 3, change order cadence and inventory buffer so clearance time is not eating fresh product. If it is stage 4 — and it very often is — no supplier change will help you, and the fix is pack size.
Lever one: pack size matched to weekly draw-down
The single most common mismatch we see in tropical markets is a 1 kg bag sitting open on a shelf in a store that uses 120 g of matcha a week. That bag is open for roughly two months. In an un-airconditioned Bangkok back-of-house, that is the whole problem, and it is not solvable by grade, price or supplier.
Calculate it yourself, per outlet, rather than per chain:
- Measure grams per serve. Weigh the actual dose your baristas use, ten times, and average it. Do not use the recipe card figure; use the measured one.
- Pull weekly serve counts per outlet from POS for the last eight weeks. Use the lowest week, not the average — your worst-case bag life is set by your slowest week.
- Weekly grams = grams per serve × slowest-week serves. Add your production waste and calibration allowance.
- Open-bag days = pack size ÷ weekly grams × 7.
Then apply a decision rule. Set your own acceptable open-bag exposure by running the hold test described at the end of this article — but as a starting frame, in our own handling experience with matcha in warm, humid conditions (this is operational judgement, not a measured study), an opened bag in an un-airconditioned Bangkok back-of-house is on borrowed time well before an opened bag in a temperate, air-conditioned one, and the practical answer is almost always to buy smaller packs more often rather than to buy a more expensive grade.
| Calculated open-bag days at 1 kg | Recommended pack format |
|---|---|
| Under 7 days | 1 kg is fine; the bag is consumed before it matters |
| 7–21 days | 500 g, or 1 kg only if the outlet has air-conditioned dry storage |
| Over 21 days | 100 g or 500 g; stop shipping 1 kg to this outlet regardless of unit price |
The thresholds in that table are a planning frame you should replace with your own numbers once you have run a hold test. The structure of the decision is what matters: pack size is a function of measured draw-down and measured storage conditions, not of purchasing convenience.
Two objections come up and both have answers. Smaller packs cost more per kilogram — true, and you should quote it explicitly, because the comparison is not against the 1 kg price but against the cost of product you currently discard, remake or serve below standard. And smaller packs mean more SKUs and more handling — also true, which is why the sensible pattern is usually a split: bulk 1 kg to high-volume outlets and central production, smaller packs to the long tail.
Lever two: shift the risk window
The second lever is contractual rather than physical. Three changes:
Shelf life stated from landed date, not production date
A best-before date printed at packing in Japan silently transfers all transit and clearance exposure to you. If clearance takes longer than expected, you receive product with a shortened remaining life and no recourse. Instead, specify a minimum remaining shelf life on arrival — a guaranteed number of months remaining at the point the goods are released to you in Thailand. This is standard practice in food distribution and it moves the incentive to the party that controls the schedule.
Pair it with a maximum age at shipment: "goods shipped must have been packed no more than X days prior to the departure date." Together these two clauses close the gap that a single best-before date leaves open.
Order cadence sized to your clearance window
If your own measured clearance time at stage 3 is long, a single large quarterly container means the last bags served are very old and the first weeks after a stockout are a scramble. Splitting the same annual volume into more frequent, smaller shipments costs more in freight per kilogram and less in degraded product. Run the arithmetic with your own numbers: total freight delta versus the value of product you currently write off or serve below spec.
For a first order of a newly registered product, plan the buffer around the registration gate specifically. That gate is one-time; do not size your entire ongoing cadence around it.
Packaging specification written into the PO
Do not assume. Specify.
The PO clause checklist
Take these to your supplier as requests, not assumptions. Some Japanese suppliers can meet all of them; some can meet most; the ones who cannot should tell you before you order, which is itself useful information.
- Pack size. "Supply in [100 g / 500 g / 1 kg] units as specified per line item." State the split by outlet tier if you are running one.
- Barrier material. "Primary pack to be an aluminium foil laminate or equivalent high-barrier structure, opaque to light, with a heat-sealed closure." Matcha degrades under light as well as heat; a clear or semi-transparent pouch is a liability in a bright warehouse.
- Gas flush. "Primary pack to be nitrogen-flushed with residual oxygen at or below [X]% at time of sealing; state the measured or specified residual oxygen level." Ask what level they actually achieve rather than accepting "nitrogen flushed" as a binary. If the supplier uses oxygen absorbers instead, that is a valid alternative — ask which, and ask for the specification.
- Outer packaging. "Primary packs to be packed in sealed cartons; cartons to be palletised and shrink-wrapped with corner protection." Loose cartons in a shared LCL container take handling damage, and a punctured foil pouch is a total loss on that unit.
- Lot traceability. "Each primary pack to be marked with lot number, milling date and packing date." You cannot run the worksheet without dates on the pack.
- Temperature control in transit. "Shipment to move under [reefer at X°C / insulated liner with thermal blanket / dry container with desiccant and non-door-adjacent stow]. Supplier to confirm mode before booking." Ask your forwarder to quote all three so you are comparing costs, not guessing.
- Minimum remaining shelf life on arrival. "Goods to have not less than [X] months of stated shelf life remaining at the date of release from Thai customs." Define release, not arrival, so clearance time sits with the schedule and not silently with you.
- Maximum age at shipment. "Goods to have been packed no more than [X] days prior to vessel departure or flight departure."
- Documentation pack. "Supplier to provide certificate of analysis per lot, ingredient and manufacturing statement, and certificate of origin in the form required for preferential duty claim, delivered by email no later than [X] days prior to arrival." Late documents are the second most common cause of extended clearance after incomplete registration. Confirm the exact document set with your Thai broker and put that list, not a generic one, into the PO.
- Sample retention. "Supplier to retain a sealed retention sample of each shipped lot for [X] months and to supply it on request." This is what lets you settle a dispute about whether degradation happened before or after departure. Without a retention sample, the argument is unresolvable and defaults to blaming whoever is less patient.
That last clause is the quiet one and the most valuable. Once both sides know a sealed reference exists in Japan, the conversation about a yellowing lot changes from opinion to comparison.
Run a hold test before you commit a container
Everything above is a framework. The number you actually need — how many days an opened bag survives in your back-of-house — can only come from your own site. Here is how to get it in thirty days.
- Request samples in the pack format the worksheet points to, not the default 1 kg. If your calculation says 500 g, test 500 g. Testing a 1 kg bag and extrapolating tells you very little, because headspace-to-product ratio changes with pack size.
- Ask for the same lot in two specifications if the supplier can provide it — for example nitrogen-flushed versus standard — so you are measuring the effect of the spec you are considering paying for.
- Place them where the product will actually live. One set in the DC, one in a representative outlet back-of-house. Put a datalogger next to each.
- Open on day zero and handle them as the outlet would — same scoop, same frequency, resealed the same way. A test that keeps the bag pristine measures nothing.
- Assess on days 0, 7, 14, 21 and 30. Photograph the dry powder against a fixed white reference under the same lighting each time. Record aroma on a simple three-point scale by the same two people. Prepare a standard usu-cha and a standard latte and taste both blind against a freshly opened reference bag held cold.
- Record the day the sensory panel first rejects it. That day, minus a safety margin, is your open-bag limit — and pack size follows directly from it.
Thirty days of discipline here replaces a year of arguing with suppliers.
What only the supplier can answer
When you finish the worksheet, you will be left with exactly two questions that no amount of internal measurement can resolve: can this supplier pack the size and gas-flush specification my numbers point to, and what does that do to unit price and minimum order quantity.
Bring those two questions, plus your calculated weekly draw-down per outlet tier and your measured clearance time, to the conversation. A supplier who can see your actual numbers can propose a pack split and a shipment cadence rather than quoting a price list.
The practical next step: request samples in the pack format your draw-down calculation recommends — not the default 1 kg — and run the thirty-day hold test in your own Bangkok back-of-house before you commit to a container. Request samples in your specified pack format, and tell us the outlet count, weekly gram usage and storage conditions you measured, so the samples you receive match the decision you are actually making.
Verify before you rely on this
Thai FDA import licence, food product registration and Thai-language label requirements are administered by the Food Division of the Thai FDA and are subject to change; confirm current requirements directly at fda.moph.go.th or through a licensed Thai customs broker before your next order, and review at least annually. HS classification and preferential duty treatment under the Japan–Thailand and ASEAN–Japan agreements should be confirmed against the current tariff schedule at customs.go.th. Climate normals should be taken from the Thai Meteorological Department at tmd.go.th for your specific province. Japanese standards and certification references are published by MAFF at maff.go.jp. Freight transit times, reefer availability and surcharges change continuously and should be quoted fresh by your forwarder with a date attached.
