A matcha latte recipe built to stay the same every time

A good matcha latte recipe is boring on purpose. It is a small set of fixed numbers a team can hit without thinking, because the whole point of a café menu item is that the fifth one tastes like the first. The recipes that circulate online are usually written for a single home cup — a scoop of powder, some hot water, milk, done — and they fall apart the moment a second barista, a busy rail, or a new bag of powder enters the picture. This one is written the other way round: as a build a buyer can hand to a floor and trust.

The base recipe, in grams

Here is the build, written the way a bar should run it. Everything downstream in this article is a reason one of these numbers is what it is.

Scale the dose with the cup, not the other way round: a 16 oz drink takes about 3 g so the matcha is not lost in the milk, and a small 8 oz takes 1.5 g. The ratio holds; the grams move. From here, each variable in turn.

What ratio actually works?

The ratio is the recipe. Two grams of matcha to roughly 300 ml of finished drink is the café baseline — enough for the matcha to read clearly through milk without tipping bitter. A house that wants a bolder, more astringent, more “serious” cup runs 3 to 4 g; a softer, more approachable one stays at 2 g and leans on the milk. Neither is wrong. What matters is that you pick one and write it down in grams.

Spoons are the enemy of consistency here. A level teaspoon of matcha is roughly 2 g, but a heaped one can be closer to 3 — a 40 percent swing that shows up as one latte tasting flat and the next tasting sharp. A cheap 0.1 g scale on the bar removes that variance entirely, and it is the single highest-return change most cafés can make to a matcha programme. If you serve matcha in real volume, the scale pays for itself in complaints you never receive.

Which grade should you build on?

A latte is milk plus matcha, and milk flattens nuance — so the grade you build on is not the one you would whisk straight as usucha. You want a matcha with enough colour and body to carry through dairy or oat rather than one whose delicate top notes vanish under it. In practice that is a mid-range ceremonial or a high culinary grade: vivid green, a clean umami base, and enough backbone to stand up to 250 ml of milk without disappearing.

Two properties matter more than price. The first is colour — a well-shaded, well-milled grade holds a clean green in milk instead of collapsing to khaki, which is most of what a photogenic latte is. The second is how cleanly it suspends, because a coarse powder grits out in milk no matter how good the barista is. Both are decided upstream, in the shading and the mill, not on your bar. We cover the colour question in depth in the guide on matcha colour for lattes, and the dispersion side in matcha solubility.

MATSU runs an eight-grade architecture across Uji, Kagoshima, and Izumo — some grades built for straight whisking, others milled and shaded specifically to hold up in milk at rail volume. You can see the full ladder and where the latte workhorses sit on the grades and pricing overview; pricing runs $390 to $1,050 per kilogram, FOB Japan.

What water temperature, and why not boiling?

Use water at 70 to 80 C (158 to 176 F), never straight off a rolling boil. Boiling water scorches the fine powder: it pushes the flavour bitter and harsh, and it dulls the colour toward the grey-khaki end you spend everything else trying to avoid. Matcha is milled leaf, not a robust black tea — it wants heat that wets it, not heat that cooks it.

You do not need a thermometer to get there. Boil the kettle, then let it stand two to three minutes; or cut roughly one part cold water into three parts just-boiled. In a latte the steamed milk carries most of the warmth to the cup anyway, so the slurry water is only there to wet and disperse the powder — it does not need to be hot enough to keep the finished drink piping. This is also why an iced latte works perfectly with cold water: the temperature was never doing the flavour work in the first place.

How do you stop it clumping?

Matcha does not dissolve — it suspends. Drop a dry dose straight into a jug of milk and the outer particles wet instantly, forming a skin that traps dry powder inside as lumps, and you get the classic gritty latte with green flecks on top. The fix is the slurry step: disperse the powder in a small measure of water first, so every particle wets individually before the bulk of the milk arrives.

• Whisk or froth to a slurry. Sieve the dose if the powder has sat, add 50–60 ml of 70–80 C water, and work it with a bamboo whisk or a handheld milk frother until it is smooth and glossy with no lumps. A frother is faster and more consistent on a bar.
• Then add milk. Steam and pour as you would any latte. The matcha is already suspended, so it folds into the milk cleanly instead of clumping against it.

Technique only takes you so far, though. A coarse or moisture-clumped powder grits out however carefully you slurry, because clean suspension is set in the mill before the tin is ever opened — fineness and de-veined leaf do most of the work. That is why we treat “does it clump” as a grade question first and a barista question second; the full mechanics are in the solubility guide.

Hot versus iced: what changes?

Less than you would think. The dose and the ratio stay the same across hot and iced — what changes is how you disperse the powder, because cold liquid fights you. Cold water is more viscous and wets particles slowly, so a whisk cannot break the powder apart the way it does in warm water. The answer is mechanical: a shaker.

• Iced method. Put the dose, 50–60 ml of cold water, and a scoop of ice into a sealed shaker and shake hard for ten seconds. The violent agitation and the ice separate the particles that a whisk would leave behind. Pour over fresh milk and ice.
• Nudge the dose. Ice dilutes and cold mutes flavour, so many houses run the iced dose 0.5 to 1 g higher than the hot — 2.5 to 3 g where a hot latte takes 2 g — so the drink still reads clearly once it is cold and long.

Iced matcha is where a coarse powder exposes itself fastest, which is why an iced-forward menu leans hardest on a fine, well-milled grade plus a shaker routine. If cold drinks are a big share of your sales, that build deserves its own attention — we walk it through in the guide on matcha for iced lattes, and set the latte alongside the rest of a café menu in matcha drinks.

Which milk should the recipe specify?

The milk is part of the recipe, not a free substitution — it changes both flavour and colour, and a matcha that looks brilliant in one milk can look flat and grey in another. Two options are reliably safe to standardise on:

• Whole dairy. Its fat and protein carry the matcha cleanly, steam to a stable microfoam, and keep the colour honest. The default for a reason.
• Barista-formulated oat. The most dependable plant milk for matcha — it holds a clean green, steams well, and its slight natural sweetness flatters the tea. If you offer one non-dairy default, make it this.

Where care is needed is with thinner or more acidic milks. Some almond and coconut lines can curdle against hot matcha or wash the colour toward grey, and their flavour can fight the tea rather than frame it. None of that means avoid them — it means test any milk before you standardise it, on your actual grade, hot and iced, and see the cup before it reaches a menu. A latte recipe that does not name its milk is only half written.

How do you keep it consistent across a team?

Everything above assumes one honest cup. A café needs the recipe to survive a floor of different hands and a stack of deliveries across a year. The discipline is the same one good coffee programmes already run:

• Write it in grams and millilitres, not scoops and glugs, and post it on the bar.
• Weigh the dose on a small scale, fix one water temperature, and standardise one dispersion method per drink — frother for hot, shaker for iced.
• Name the milk, and treat a milk change as a recipe change to be re-tested, not a swap.

But the recipe can only ever be as consistent as the powder it is built on. If a new delivery is milled coarser, shaded differently, or blended from another lot, your dialled-in numbers quietly stop landing — the colour shifts, the dose reads weaker, the slurry grits. This is where sourcing does the work most cafés never see. Because MATSU is producer-direct — grower-level sourcing without the usual chain of trading houses and distributors — the lot you dial your recipe on is the lot you reorder, milled from the same de-veined leaf by the same people season after season. That stability is the whole reason a fixed recipe stays fixed. If matcha is becoming a real line on your menu, the sourcing side is worth reading in full in wholesale matcha for cafés.

Frequently asked questions