Second-harvest matcha is not second-rate: a blind cupping

What second harvest means, and what it does not

In Japanese tea the harvest number tells you when the leaf was picked. The first harvest, ichibancha, is the spring flush from late April into May. The second, nibancha, comes in June and July from the growth the bush pushes after the first picking, and autumn tencha follows later in the year. The second-harvest leaf has had more sun and less of the slow, shaded spring growth that builds amino acids, so the general tendency is more astringency and less of the sweetness the spring flush is prized for. What it is not is a lower grade of the same thing: it is a different leaf, milled the same way, at a lower price. We went through the chemistry and the invoice in first harvest vs second harvest. This article is the practical part: how to put the two in front of your team without the labels, and what to do with the sheet afterwards.

Why blind, and why in milk

Price is an anchor. So is the word ceremonial on a bag, and so is the harvest line on a spec sheet. Once a taster knows a bowl cost $290 / kg they find the umami in it. Numbered bowls take that away, and the results change more than most owners expect.

Two rounds matter. Straight, as usucha, the differences between harvests are at their widest, and that is where you learn what each lot is. Then in milk, because that is how most matcha in a café is drunk, and milk and sugar compress the differences. A lot that wins straight does not always win in a latte, and a lot that loses straight is sometimes the one your customers would pick.

Which lots to put in the flight

Four to six lots is enough, and they should span both harvests at more than one price. From our catalog, a flight that works:

One kilo is 250 lattes at 4 g, so the tea cost per cup is the catalog price divided by 250

The sample form sends up to six teas at 20 g each, which is exactly this flight. Twenty grams covers three tasters at 2 g each straight and a 4 g latte for each of them, with a little left over for the iced check. Ask for the spec sheets with the samples, but keep them in a drawer until the reveal.

The method, step by step

1. One person prepares and does not score. They number the samples 1 to 6, write the key down and hide it.
2. Powder first. A small spoon of each on white paper, side by side. Score the colour of the dry powder from 1 to 5. The eye sees the harvest before the tongue does.
3. Straight round: 2 g sifted into the bowl, 70 ml of water at 75 °C, whisked for 15 seconds, the same bowl type for every sample, served within a minute of whisking.
4. Score each bowl 1 to 5 on colour of the tea, aroma, umami (the savoury depth), astringency (the dry grip on the gums), bitterness and finish, and add one line of words. Rinse with water between bowls.
5. Latte round: 4 g, 40 ml of water at 80 °C, 180 ml of milk steamed to 60 to 65 °C. Same milk, same cup, same person steaming. Score colour, sweetness, how much tea comes through the milk, and bitterness.
6. Iced check, optional but useful: one iced latte per sample, left on the counter for 30 minutes, then compared for colour. Why some lots go olive is in our colour article.
7. Before the reveal, every taster writes a guessed price per kilo next to each number. Then open the key.

What usually shows up

Every lot on our product pages carries our own scores for umami, smoothness, aroma and colour hold. Sae sits at 89 umami and 93 colour hold, Hikari at 87 and 92, Megumi at 77 and 86, Tsuyu at 76 and 84. In the straight round a blind flight tends to track those numbers: the first-harvest bowls are greener, sweeter and rounder, the second-harvest bowls have more grip and a plainer, more leafy aroma. Tasters who have never had a first-harvest ceremonial lot straight usually pick it out of the line-up at once.

In milk the gap narrows a lot. A second-harvest Yabukita such as Megumi has the body to be tasted through 180 ml of milk, which is why we list it as a dependable latte base. Some first-harvest lots are so soft that the milk swallows them, and the latte scores come out close to the cheaper bowls. The one that surprises most panels is Tsuyu: Asatsuyu is a naturally low-astringency cultivar, so a second-harvest Tsuyu latte often reads as sweeter and softer than a first-harvest blend costing more. The price guesses are where the room goes quiet, because they rarely line up with the key.

• Colour hold is the clearest harvest marker after 30 minutes on ice. First harvest tends to stay green longer; most second-harvest lots drift toward olive sooner. If your menu is mostly iced, weight this score.
• Astringency in milk is not a fault by default. A little grip is what makes a latte taste of tea rather than of sweet milk. Harshness is a fault; grip is a style.
• Aroma drops off fastest across harvests and matters least in milk. Score it, then discount it for the latte decision.

Reading the sheet: what to buy

If the latte scores are within a point of each other and the straight scores are two or more apart, the answer is usually two lots: a second-harvest café lot for the house latte and a first-harvest lot for usucha and the signature drink, which is the menu structure we recommend in the harvest article. The money is in the first column. Between Sae and Megumi the difference is $105 / kg, which is $0.42 of tea per latte; at 100 lattes a day that is $42 a day and about $1,260 a month, for a difference your panel could not taste through the milk. Spend it instead on the first-harvest lot for the drinks served straight, where everyone could.

If the second-harvest bowls lost both rounds clearly, buy the first harvest and price the menu for it. The point of the cupping is not to prove the cheaper lot wins; it is to stop the harvest line on a spec sheet making the decision for you. How we define ceremonial, café and culinary, and which harvests sit in each, is on the grades page.

Questions buyers ask

Prices and lots named in this article are as published on Nov 2026; the catalog always carries the current list price per kilo.