How to Test Matcha Samples for Your Café: A Step-by-Step Protocol

A matcha sample only tells you something if you test it the way you will actually serve it, under the same conditions as whatever you compare it with. Fix everything except the matcha (dose, water, milk, cup, tool, temperature), taste blind, and score against criteria you set before you start. This protocol takes one session of about an hour and fits comfortably within two 10 g samples.

Before you start: decide what you are testing for

Write down the decision this test should answer. Typical examples:

• "Which matcha do we use for our house oat latte?"
• "Is a new matcha at least as good as our current one in milk?"
• "Do we need a separate matcha for a pure serve?"

Then write your acceptance criteria. If you have none yet, setting acceptance criteria covers how. A simple version: "Clearly tastes of matcha through 200 ml of oat milk, no grit, colour stays bright green for at least five minutes."

Step 1: Gather and budget the samples

A 10 g sample is enough for this protocol if you plan it. At 2 g per drink it gives a theoretical 10 ÷ 2 = 5 preparations.

Use	Grams per sample
Visual check (dry powder)	0 (look, do not use)
Latte test, 2 cups at 2 g	4
Straight test, 1 cup at 2 g	2
Repeat or staff tasting	4
Total	10

If you test both the Ceremonial Barista Matcha sample and the Ceremonial Matcha sample, plus your current matcha as a control, keep 10 g of the control aside too.

Step 2: Set up controls

Inconsistency between cups ruins a test faster than anything. Prepare the station before anyone tastes.

• Scale reading to 0.1 g. Spoons and scoops vary too much for a comparison.
• Same water for every cup: same source, same temperature, measured with a thermometer.
• Same milk brand and type, from the same carton, at the same temperature. Use the milk you sell most.
• Same cup and volume as your menu drink.
• Same tool and time: whisk or frother, same duration for every cup.
• Sieve every sample through the same fine sieve.
• Labels: code each sample (A, B, C) on the bottom of the cup. One person prepares; others taste without knowing which is which.

Step 3: Visual and aroma check

Open each sample on a white plate next to the others, in daylight if possible.

• Colour: note how green, bright or dull each powder looks. Compare side by side rather than judging alone.
• Texture: fine and soft, or with visible clumps that do not break under light pressure.
• Aroma: smell the dry powder, then smell it again right after adding a few drops of warm water.

Write notes, but do not score yet. Appearance can bias the tasting.

Step 4: The latte test

This is the main test if most of your matcha is sold with milk.

1. Weigh 2 g of matcha (or your house dose) into the cup through the sieve.
2. Add a small amount of water, about 30–40 ml at around 80 °C, and whisk to a smooth paste without lumps.
3. Steam or heat the milk to your normal service temperature and pour as you would for a guest.
4. Taste within two minutes, then again after five minutes.

What to look for: does matcha flavour come through the milk, or does the drink just taste of milk? Is there bitterness at the back? Any grit in the last sip? Has the colour dulled after five minutes? The side-by-side latte test goes into scoring milk drinks in more detail, including iced versions.

Step 5: The straight test

Even if you only sell lattes, tasting matcha with water alone shows faults that milk hides.

Ippodo's producer guidance for usucha is a useful baseline: 2 g of matcha, 60 ml of water at 80 °C, whisked briskly for about 15 seconds. Use it for every sample.

Taste for sweetness and savoury depth, bitterness and astringency, and how clean or lingering the finish is. A matcha you plan to serve straight should do well here. A matcha for milk may taste stronger here and still be the right choice. The buyer's usucha test covers the vocabulary and what to expect.

Step 6: Score blind

Score each coded cup on a 1–5 scale, per taster, before anyone talks. Use the weights that fit your decision.

Criterion	Weight (latte decision)	Weight (pure serve decision)	A	B	C
Matcha flavour through milk	30	0
Balance in water (bitterness, astringency)	10	35
Texture and smoothness (no grit)	20	20
Colour in the cup after 5 minutes	15	15
Aroma	10	20
Ease of preparation (mixes without lumps)	15	10
Weighted total (score × weight, max 500)	100	100

For a printable version with descriptors per score, use the matcha sample scorecard. Average the weighted totals across tasters. Two to four tasters is enough; include at least one barista who will make the drink daily.

Step 7: Decide

Use a simple rule so the decision does not drift into opinion.

• Clear winner (highest total and meets all acceptance criteria) → order the 1 kg bag of that product.
• Close result (within about 5% of each other) → choose on cost per serving and preparation ease, then run a short in-service trial.
• Nothing meets the criteria → change one variable (dose up by 0.5 g, or a different milk) and retest the top candidate, or ask the supplier which product they would recommend.

If you are comparing the new matcha with the one you already use, keep the current one coded and blind too, so the familiar taste does not win by default.

A note on record keeping

Keep the score sheet, the recipe used and the date. When the first 1 kg bag arrives, run Step 4 once with the bulk product and compare with your notes. It gives you a reference point for every later delivery.

Next step

Order both 10 g samples (€4.95 each excl. VAT at the time of writing; check the product page for the current price) and book an hour before opening. If the latte test decides it, move to Ceremonial Barista Matcha in a 1 kg bag.

Sources

• Ippodo – Basic Usucha (Matcha)