Why this page gives ranges and descriptions rather than dates
A date implies a threshold, and there is no threshold. What happens instead is a continuous drift whose speed depends on how the tea was stored, so the same leaf can be tired in four months or serviceable in eighteen. Two things are therefore worth stating for each category instead of a number: roughly how long a well-stored example holds its character, and what specifically changes when it stops. The second matters more, because it is what lets a drinker recognise the state of a tea in the cup rather than by arithmetic from a packing date. All ranges below assume the storage that category wants — sealed, opaque, cool and dry for the preservation categories, and stable and ventilated for the ageing ones. They also assume an unflavoured tea; added flavourings run on their own, usually shorter, clock. There is a second reason to describe the change rather than to name a deadline. A drinker who knows what fading looks like in a category can assess a tea they have just been given, or one that has been in a cupboard for an unknown period, without any dates at all — which is the situation most people are actually in. A date answers a question about a packet; a description of the change answers a question about a cup, and the cup is the thing in front of you.
Japanese green tea, and matcha
The shortest clock in the catalogue. Sencha, gyokuro and their relatives are at their best within a few months of processing, hold reasonably for something in the region of a year when sealed and cool, and are noticeably tired beyond that. What changes is the aroma first — the fresh, marine, cut-grass top notes thin out — then the sweetness, leaving a flatter, more generically vegetal cup with the astringency more exposed than it was. Colour shifts from bright green towards olive and then brown, in the dry leaf and in the liquor. Matcha is faster again and is treated separately, because grinding removes every remaining protection. Shincha is a special case in the other direction: it is sold precisely because it is new, and buying it late in the year defeats the point of the product rather than merely reducing it. Two Japanese categories run on their own clocks and are worth separating out. Hojicha, roasted at high temperature, is far more stable than the green teas it is made from and keeps for a year or more without much loss, which is part of its appeal as an everyday tea. Kukicha and other stem teas sit between the two. Genmaicha is limited by its roasted rice as much as by its leaf, since the rice goes stale on its own schedule and contributes a distinctly tired note when it does.
Chinese green tea and yellow tea
Similar in kind, slightly more forgiving in practice, and much more variable between styles. A pan-fired Longjing or Bi Luo Chun bought in spring is at its best that spring and through the following months, and is a different and lesser tea a year later; the pre-Qingming premium exists partly because this is understood. The changes mirror the Japanese pattern with a different vocabulary — the toasted, chestnut, bean-like top notes recede, the liquor loses brightness, and what remains reads as thin rather than harsh. Yellow tea behaves as green tea does, its yellowing step notwithstanding, and is stored the same way. One genuine exception is worth naming: some Chinese greens, notably certain Liu An and Tai Ping styles, have small local traditions of deliberate keeping, but these are specific practices rather than a general licence to let green tea age. There is a further complication specific to Chinese green tea, which is that a good deal of it is roasted or fired more thoroughly than its Japanese equivalent. A well-fired Longjing or a gunpowder green is measurably more stable than a lightly fired bud tea from the same region, so the category spans a wider range of clocks than a single figure suggests. The practical guidance is to judge by how thoroughly the tea was fired — which is usually evident in the toasted character of the cup — rather than by the word green alone.
White tea, which runs on two clocks
The only category whose shelf life depends on what the buyer intends. Drunk as a fresh tea, a silver needle or a bai mudan is at its best in its first year: delicate, hay-like, melon-sweet, and easily lost. Stored deliberately, particularly as a pressed cake, the same material moves over years towards dried fruit, honey, date and a much heavier body, and the trade in aged white tea is built on precisely that. The two paths diverge in storage rather than in the leaf: sealed and cool holds the fresh character, while stable, ventilated and undisturbed permits the aged one. The intermediate case is the trap — a loose white tea left half-sealed in a kitchen for three years is usually neither, having lost its freshness without acquiring depth. More mature grades such as shou mei carry more of the material that ageing acts on, which is why the cheapest grade is often the one worth keeping. For a buyer the practical question is which path a given packet was made for, and the format usually answers it. Loose silver needle in a sealed pouch with a recent harvest date was made to be drunk fresh. A pressed cake, particularly of a more mature grade, was made to be kept. Where a seller offers both, the price difference is often smaller than the difference in intent, and choosing by intent rather than by grade is what stops a buyer ageing the wrong material for five years.
Oolong, where the roast sets the clock
Unroasted, lightly oxidised oolong — most contemporary Tie Guan Yin, most Taiwanese high mountain tea — is close to green tea in fragility and is best within the year. Its failure mode is distinctive: it does not so much flatten as turn sour, and a faintly sour high mountain oolong is almost always one that has taken up moisture. Roasted oolong runs longer and behaves differently again. A fresh roast is often at its least attractive immediately after firing and improves over some months as the roast settles and integrates, then holds for years. Its enemy is reabsorbed atmospheric moisture, which dulls the roast and pushes it towards that same sourness, and the traditional answer is periodic light re-firing rather than tighter sealing. A traditionally roasted oolong maintained this way has no useful upper limit, which is why aged oolong exists as a category at all. There is a middle case worth naming, which is the lightly roasted contemporary oolong that has become common in Taiwanese and Fujianese production. It has enough roast to be more stable than an unroasted tea and not enough to age indefinitely, so it holds well for a year or two and then quietly flattens. These teas are frequently bought in quantity on the assumption that any roasted oolong keeps, and the assumption holds for a traditional charcoal roast and not for a light finishing bake.
Black tea
The most forgiving of the preservation categories, and the one where storage errors cost least. A well-sealed black tea holds its character for a couple of years and declines gently rather than dramatically after that. What goes first is the aromatic top — the muscatel of a second flush Darjeeling, the cocoa and orchid of a Keemun, the sweet-potato note of a Dian Hong — leaving the malt, the body and the tannic structure intact. The result is a serviceable, duller tea rather than a spoiled one, which is exactly why bagged black tea can sit in a supermarket supply chain for a long time without anybody complaining. First flush Darjeeling is the honourable exception and behaves much more like a green tea, since its character is largely in the fragile aromatic fraction its light processing preserves. Two qualifications apply within the category. Chinese black teas with a high proportion of buds — Jin Jun Mei, golden-tipped Yunnan styles — carry more of their character in the fragile aromatic fraction and fade faster than a robust broken-leaf Assam. And smoked teas are among the most stable things in the cupboard, since the smoke character both dominates and persists. Within a category that is generally forgiving, the same principle applies as everywhere else: the more of a tea’s value sits in aroma, the shorter its useful life.
Dark tea and pu-erh
No shelf life in the ordinary sense, because time is the point. What exists instead is a trajectory: young sheng is bitter, brisk and aromatic; the middle years, often described as an awkward period, can be neither one thing nor the other; and long storage brings dried fruit, wood, camphor and a thick sweet body. Ripe pu-erh and most hei cha start further along that road because wet piling did some of the work, and they change more slowly thereafter. None of this happens on schedule — a cake in Guangdong and the same cake in a dry northern flat will be years apart at the same age, which is the storage-versus-age point again. The genuine failure mode is not age but damp: a cake that was stored too wet is musty rather than mature, and that is a permanent condition rather than a stage. Ripe pu-erh has one further characteristic worth knowing, which is that a young shou often carries a distinct fishy or pond-like note from the wet piling stage and loses it over a year or two of airing. That is not a fault to be returned but a stage to be waited out, and a shou bought young and disliked is worth revisiting later. The catalogue keeps a separate troubleshooting record for the note precisely because it is so commonly mistaken for spoilage.
What the printed date is actually telling you
A best-before date on tea is a quality estimate set by the packer, not a statement about anything else, and packers set them very differently — two and three years are common defaults applied regardless of category. That is why a best-before date is nearly useless for the teas where freshness matters most: a sencha with a two-year best-before is being described by a rule of thumb that does not fit it. A harvest date or a production date is the informative one and is much rarer outside Japanese and Chinese speciality supply. Where only a best-before appears, subtracting the packer’s standard period is guesswork. The reliable move is to buy from sellers who state the harvest, and to read the printed date as a packaging requirement rather than as information about the leaf. One thing a durability date does tell you reliably is how the packer classifies their own product. A three-year date on a green tea says the packer regards it as shelf-stable stock rather than as a seasonal product, which is a fair description of the supply chain it came from. Read that way the date is genuinely informative — not about the leaf, but about the business, and by extension about what kind of tea you are buying and how it has probably been handled on the way.