Your outlets report the same three things within four to eight weeks of a delivery landing: the powder forms lumps, it sits on the sieve instead of falling through, and the cup tastes flat-sweet with the top aroma gone. The COA at shipment looked normal. The question is not whether something went wrong, but where — in the tea itself, or somewhere between customs release and the back-of-house bench.
This article covers that separation: a cause-isolation sequence a Singapore buyer can run in-house across one or two deliveries, and the specific supplier-side changes that follow from each outcome. If you have not yet connected Singapore's climate and last-mile conditions to matcha quality at all, start with tropical warehousing and last-mile handling for matcha in Singapore. For an end-to-end quality-loss narrative in a different market, see matcha quality loss across the Thailand import chain. For general storage rules that are not Singapore-specific, see matcha storage for businesses. This piece assumes you already have the symptom in front of you.
What caking and flat taste actually tell you
Matcha is a very fine, low-moisture powder with a large surface area. Two separate mechanisms produce the complaints you are seeing, and they leave different fingerprints.
Moisture pickup. Water vapor condenses on or migrates into the powder, particles bridge, and lumps form. Lumps are mechanical: they resist the sieve, they break under pressure, and they appear first at the surface of an opened pack or along the inside of a liner. Moisture also accelerates the reactions that strip the fresh green top note and leave a dull, flat sweetness.
Oxidation and heat exposure. Color drifts yellow-brown at the edges, aroma flattens, and the astringent tail becomes rougher — but the powder may still sieve normally. Flat taste without caking points here.
The useful consequence: caking is a moisture story, and moisture in Singapore is mostly a story about how many times a pack crosses a temperature boundary. Every time a chilled pack is brought into an ambient room and opened before it has equalized, the air that enters is warm and humid, and the surface it meets is cold. That is a condensation event. Our handling experience is that the count of these events after customs release — not the grade of the tea — explains most of the caking complaints we are asked about from tropical markets. That is an observation from how we handle and ship product, not a measured study.
The two candidate explanations, stated precisely
Before you measure anything, write down the two hypotheses so the result can actually decide between them.
Hypothesis A — supplier spec. The matcha left the mill with too much residual moisture at packing, or with a particle distribution that bridges easily, or in a pack format whose barrier and headspace do not hold for the transit and shelf life you need. If this is true, caking will appear in unopened packs held at stable temperature.
Hypothesis B — chain and handling. The spec is fine, but the pack is opened repeatedly in a humid back-of-house room, or it crosses chilled-to-ambient boundaries several times between the distributor's cold room, the delivery van, and your outlet. If this is true, unopened temperature-stable packs will stay clean while opened or cycled packs degrade.
Both can be true at once. The point of the ledger below is to find out which one dominates for your chain, because the two lead to different purchase orders.
The condensation cycle ledger
Run this over two consecutive deliveries. It needs no instruments beyond a kitchen scale, a sieve you already own, and a cheap hygrometer-thermometer for the back-of-house area. Keep one tin per delivery as a sealed control.
What to log, per pack
- Pack identity. Lot number, pack size, whether it is nitrogen-flushed, and the inner-liner type (foil-laminate, metallized film, or plain).
- Transitions before first opening. Count every move from chilled to ambient (or back) that the pack makes after customs release: distributor cold room to van, van to outlet shelf, shelf to your fridge, fridge to bench. Ask the distributor for their leg; do not assume.
- Openings per shift, plus the room temperature and relative humidity at the moment of opening, and whether the pack was allowed to reach room temperature before the seal was broken.
- Scores at week 0, week 4, and week 8 using the rubric below.
Sieve-resistance score (1–5)
Weigh 20 g, pass it through your normal sieve with your normal technique, and score what happens. Use the same person and the same sieve every time; the scale is comparative, not absolute.
- 1 — Falls through freely, no residue on the mesh.
- 2 — Falls through with light tapping; a few soft lumps that collapse instantly.
- 3 — Needs sustained tapping or brushing; visible lumps that break with finger pressure.
- 4 — Needs pressing through the mesh; lumps hold their shape when handled.
- 5 — Hard aggregates that will not break by hand; material left on the mesh after effort.
Taste note
Prepare the same way each time (same water temperature, same dose, same vessel) and record three marks: aroma present or flat, sweetness clean or dull, finish smooth or rough. One line is enough. You are tracking direction, not writing a tasting panel report.
The control pack
This is the part most buyers skip, and it is the part that makes the ledger decisive. From each delivery, set aside one unopened pack in a place with stable temperature — a dry ambient store room that does not swing is better than a fridge that is opened forty times a day. Do not move it. Score it at week 4 and week 8 alongside the working packs.
The decision rule
At week 8, compare the control pack against the working packs.
If the unopened, temperature-stable control has also caked (sieve score 3 or higher) — the problem travels with the product. Something in the spec or the pack is not holding. This points at the supplier side.
If the control is clean (1–2) but working packs score 3–5 — the product arrived sound and degraded in your chain. This points at pack format and handling, not at grade.
If the control is clean and working packs are also clean, but taste has flattened everywhere — you are looking at oxidation and cumulative heat exposure rather than moisture. Shorten the time between arrival and use, and look at how long stock sits at the distributor before it reaches outlets.
Cross-check the transition counts: if the packs with the most chilled-to-ambient crossings before first opening are consistently the worst, you have your answer regardless of how the averages look.
What to change on the supplier side, per outcome
If the control pack caked — change the spec
- Moisture at packing. Ask what the actual value was for your lot, not just whether it passed. Ask whether it is stated on the COA you receive, and if it is not, ask for it to be added as a line item on future lots. A number you cannot see is a number you cannot negotiate.
- Particle distribution. Very fine fractions bridge more readily. If your application is latte and bakery rather than ceremonial preparation, a slightly different distribution may sieve far better at the same sensory quality. Ask what is available.
- Barrier and flush. Confirm the inner-liner material and whether the pack is nitrogen-flushed. A plain liner that is fine for a domestic Japanese shelf is not the same proposition after a tropical transit and several weeks in a Singapore store room.
If only opened or cycled packs caked — change the format
- Smaller packs sized to one shift. The single most effective change is usually to stop reopening a large tin. A pack that is consumed in one shift is opened once. Specify the size against your real per-outlet throughput, not against the unit price per kilogram.
- Nitrogen-flushed units for anything that will be held more than a few weeks before opening.
- A written equalization rule for outlets: chilled packs come out and sit sealed until they reach room temperature before the seal is broken. Decide the waiting time from your own back-of-house measurements — put a thermometer against a pack and time how long it takes to stop being cold to the touch in your actual room. That number belongs to your site, not to a general guideline.
- Fewer cold-chain boundaries overall. Stable ambient storage in a dry, air-conditioned room, with the pack opened once, often outperforms chilled storage that is cycled daily.
If taste flattened without caking — change the cadence
Order smaller quantities more often, and ask your distributor how long stock typically sits with them before it reaches your outlets. If you do not know that interval, you are managing a shelf life whose start date you cannot see.
Why we suggest fixing the chain first
Spec renegotiation takes a production cycle and a new shipment before you learn anything. Pack format and handling changes can be tested on stock you already hold. More importantly, if condensation cycles are the dominant driver — which is what we most often find when we work through this with buyers in tropical markets — then tightening the moisture spec will not fix the complaint, and you will have spent a quarter proving it. Reduce the open-to-ambient cycles first; renegotiate moisture and particle distribution only if caking survives that change in unopened, temperature-stable packs.
Running the ledger on alternatives before you commit
Every branch of the decision rule ends in something you cannot verify from a document: whether a lower moisture-at-packing target, a different particle distribution, or a smaller nitrogen-flushed single-shift pack actually behaves better in your room, at your humidity, with your staff. The only honest test is to run the same ledger on candidate formats side by side with what you buy now, before you move volume.
Tell us the symptom you are seeing, your outlet throughput per shift, and how your stock is stored between arrival and use, and we will send samples in the pack formats worth comparing — including the moisture-at-packing figure for the lot, so your control pack result means something.
Request samples in the pack formats you want to test
Storage and pack-format guidance here should be reviewed against current packing specs and local conditions; if Singapore import or storage requirements change, confirm the current position with the Singapore Food Agency.
