The steep you planned and the steep you got
A multi-infusion schedule assumes that conditions are the same each time and that only the leaf has changed. In practice the temperature of the water reaching the leaf falls across a session, sometimes by a great deal, and that falling temperature is silently doing some of the work the extending times were supposed to do. The result is a session that seems to fade much faster than it should, an eighth steep that gives almost nothing where the record promised a long tail, and a general conclusion that the tea was not what it was sold as. Temperature drift is the most common invisible cause of that outcome, and unlike leaf quality it is entirely within reach and costs nothing to correct. It is worth stressing how invisible it is: nothing about the process looks different, the kettle was boiled, the schedule was followed, and the only evidence is a cup that has gone quiet earlier than expected. That is exactly the kind of fault that gets attributed to the wrong cause for years.
Where the heat goes
Four places, in roughly descending order of size. The kettle, which has been standing since the session began and is no longer anywhere near the temperature it was switched off at — a kettle left ten minutes is a long way down. The vessel, which cools between steeps while the cup is being drunk, and which cools faster if it is thin and faster still if the room is cold. The air, during the pour itself and while the lid is off between infusions. And the leaf, which begins as a dry mass at room temperature and at least stops taking heat once it is warmed through. Of these, the kettle is usually the largest term and is the one nobody watches, because it was boiled at the start and has been mentally filed as done. The vessel gets all the attention because preheating is well known, but preheating addresses the first steep and says nothing about the fifth.
Why it compounds rather than staying constant
The drift is not a fixed penalty applied evenly to each steep, which is what makes it worse than it first appears. A cooler steep extracts less, which leaves more in the leaf, which makes the next steep look weaker still — and the natural response is to extend the time, which pulls the balance towards the later, grippier fraction and produces a cup that is simultaneously thin and drying. Meanwhile the vessel, having just been filled with cooler water, is itself cooler at the start of the following steep than it was at the start of the last one. Two or three rounds of this and the session is somewhere quite different from where the schedule expected it to be. That is why the fault characteristically shows up as a sudden collapse around the fourth or fifth infusion rather than as a gradual decline: several small effects have been quietly reinforcing each other, and the cup crosses a threshold where the drinker notices all at once.
Thermal mass cuts both ways
A thick clay pot takes a great deal of heat out of the first pour and then holds what it has taken, so its early steeps run cooler and its late ones warmer than a thin porcelain gaiwan’s. A gaiwan does the opposite: it takes very little from the water and loses what it takes almost immediately, so it tracks the kettle closely and drifts downward with it. Neither behaviour is better in the abstract, and the choice between them is genuinely a choice rather than an upgrade. What matters is knowing which one you have, because the correction differs completely. A heavy pot needs a thorough preheat — filling it once with boiling water and discarding — after which it largely looks after itself. A thin one needs attention throughout, rewards working quickly, and benefits from a rinse of hot water between infusions. Buying the heavier vessel and then not preheating it is the worst of both, and it is common.
The kettle is the other half of the problem
A kettle switched off at boiling is well below boiling within a few minutes, and a session lasting half an hour will otherwise be brewing its last steeps with water that has cooled substantially — often into a different band entirely from the one the schedule assumed. Reboiling between steeps is the obvious answer and is standard practice in gongfu service for exactly this reason: the traditional arrangement keeps a kettle over a charcoal brazier or a small flame precisely so that the question never arises, which is a piece of equipment design that reads as ceremony and is actually engineering. A variable-temperature kettle with a hold function does the same job with less romance. Anyone doing neither is running an uncontrolled downward ramp across their session and attributing its effects to the leaf, and the correction is a single switch. For teas that want the top of the range this is the single largest available improvement.
Corrections, in order of how much they buy
Reboil or hold the kettle, which addresses the largest term and costs a button press. Rinse the vessel with hot water between steeps whenever there is a gap of more than a minute, which addresses the second. Work faster, which addresses several terms at once, costs nothing, and is the correction most often resisted because a session feels like something that should be unhurried. Use a smaller vessel, which loses proportionally more heat but is refilled sooner from a hotter kettle and so drifts less over a session. And preheat properly at the start, which sets the baseline that everything else drifts from and makes the first steep comparable to the rest. None of these is exotic, none costs money beyond a kettle you probably own, and together they remove most of the drift. Doing only the last of them — which is what most brewing advice recommends — addresses the smallest of the five.
When the drift is the point
Not every style wants a flat temperature, and it would be a mistake to read this as a rule that heat must be held constant. Japanese kyusu practice deliberately raises the temperature across a sequence rather than holding it, because the leaf becomes progressively harder to extract while the times are shortening, and the extra energy is what makes the later infusions work. Grandpa brewing lets the glass cool continuously and depends on it, since a cooling glass is a large part of what makes continuous contact with the leaf tolerable. Boiled preparations hold at one temperature by definition and have no drift at all. The principle is not constancy but intention: the temperature should be where you meant it to be at each stage, rather than wherever the equipment happened to leave it. A session that drifts by design is being brewed; one that drifts by neglect is being allowed to happen.