What a best-before date is
A durability date on a dry food is an indication of the period over which the product is expected to retain its expected quality. Two features of that definition do the work. It is about quality rather than anything else, which is why old tea is stale rather than a hazard. And it is set by the business responsible for the product, using its own judgement, rather than by an authority testing the tea. In practice packers apply a standard period — two years and three years are common — across a range with quite different clocks, so the same interval gets printed on a robust black blend and on a green tea whose good period is a fraction of it. The date is therefore an accurate statement of the packer’s policy and a poor statement about the leaf. There is a further reason the interval carries so little information about tea specifically. Durability dating was designed for foods whose deterioration is reasonably predictable from composition and packaging, and dry tea does not deteriorate so much as drift — the change is continuous, the endpoint is a matter of taste, and the rate depends overwhelmingly on storage conditions the packer does not control after despatch. A date is a poor instrument for describing a gradient, and this is a gradient.
Why the interval hides the important number
A best-before date encodes only the end point, and what a buyer needs is the start. Two packets can carry the same durability date while one was harvested this spring and packed immediately and the other was harvested eighteen months ago and sat in bulk before packing. Nothing on either label distinguishes them, and the difference is the entire difference between a fresh green tea and a tired one. This is why the durability date is not merely imprecise but structurally unable to answer the question, and why the useful response is not to look for a longer date but to look for a different field. Where a packing date is given without a harvest date, it narrows the range slightly and still leaves the time between picking and packing unstated. The gap between picking and packing is larger than most buyers assume. Tea is frequently made, sold at auction or by private treaty, shipped, stored in bulk by an importer and packed to order, and each of those steps takes time. For a robust black tea that is unimportant. For a green tea it can consume the entire period during which the tea was worth buying, and the durability date printed at the end of that chain will still be two years out.
What a harvest date states, and in what form
At minimum a year. Better, a season — first flush, second flush, autumnal, spring, or the Japanese vocabulary of ichibancha and nibancha for first and second picking. Better still, a specific window, which for the teas where it matters is normal practice: Chinese green teas are commonly distinguished by whether they were picked before the Qingming festival in early April, before Guyu later that month, or after; Japanese producers date shincha and single-picking teas precisely; Darjeeling lots carry an invoice number that identifies the production run. Each of these is a fact somebody had to know in order to print, and each of them narrows what the tea can be. A seller who states the harvest is also, implicitly, accepting that the tea will look older next year, which is a commercial cost and therefore a signal. Japanese labelling deserves a specific note because it is the most informative in common circulation. Producers routinely state the picking, the prefecture and often the cultivar, and shincha is dated tightly because the entire product is a claim about recency. A buyer who can read those fields, or who buys from an importer who translates them, has more information about a Japanese green tea than is available for almost any other category at the same price.
Where the distinction is decisive and where it barely matters
It is decisive for Japanese green tea, matcha, Chinese spring green tea, first flush Darjeeling, unroasted high mountain oolong and fresh white tea — everything whose selling point is a fragile aromatic fraction. For these, a packet without a harvest year should be read as an unknown rather than as a recent tea. It matters moderately for black tea and roasted oolong, where a year or two changes the cup less dramatically. And it inverts for the ageing categories, where the production year is a fact about the tea’s position in a long trajectory rather than about its freshness, and where a wrapper date is a claim about a date and not about the twenty years of storage that followed it. Knowing which of the three situations applies is the whole skill. Flavoured and scented teas form a further group with their own behaviour. Their added or absorbed aromatics disperse faster than anything in the leaf, so a harvest date on the base tea tells only part of the story and the more useful question is how long ago the tea was scented or blended — a date almost nobody prints. In practice this means flavoured tea is best bought from sellers with fast turnover, and judged on smell at the point of opening rather than on any printed field.
The asymmetry of information, and what to do about it
The seller knows the harvest date and the buyer does not, and printing it costs the seller flexibility — dated stock is harder to sell down slowly. That asymmetry explains the whole pattern of who prints it. Speciality sellers with fast turnover print harvest dates because it advantages them; packers with long supply chains and slow-moving stock do not, because it would not. This means the presence of a harvest date is informative beyond the date itself: it indicates a business whose stock turns over quickly enough to bear the disclosure. The practical move for a buyer is simply to ask. A seller who answers with a year and a season is one kind of business; a seller who cannot find out is another, and neither answer requires interpreting a marketing claim. The same asymmetry explains why speciality sellers often list harvest dates prominently in spring and quietly stop mentioning them by the following winter. That is not deception so much as ordinary commercial behaviour, and it gives a buyer a useful signal: a seller who continues to state the year on stock that is no longer new is accepting a disadvantage in order to be accurate, which is exactly the behaviour worth rewarding with repeat custom.
Reading a date you cannot verify
Dates are claims like any other, and a harvest year on a wrapper is not audited. The reasonable position is neither to trust it fully nor to dismiss it, but to check it against everything else on the packet and in the cup. A tea claiming a recent spring harvest should look and smell like one — bright dry leaf, vivid aroma, and a first infusion with the aromatic top notes present. A green tea claiming this year’s picking that reads flat and olive-coloured is answering the question regardless of what is printed. Where an age claim runs in the other direction, on a pu-erh cake or an aged white, it is unverifiable at any price and colour deepens with warm storage in months as readily as with genuine years. Dates narrow the range of what a tea can be; the cup decides where in the range it sits. There is one more check available on arrival, which is the packaging itself. A tea whose harvest date is recent and whose pouch has been sealed and dated by the packer is internally consistent; a recent harvest date on a pouch that has clearly been sitting open or repacked is not. Consistency between the claims on a packet is not proof, and inconsistency is a reasonable basis for a question — which most sellers will answer straightforwardly if asked.