What a stable extraction target looks like on a busy brew bar

A stable extraction target is not a single number pinned to the wall. It is a set of measurements that still make sense when the café gets louder, the grinder warms up, and a second barista steps into the station. Teams often say a brew recipe is stable because one cup tasted right at 8:15 in the morning. That is not stability. Stability means the recipe survives actual service.

On filter bars, the most reliable anchor is beverage weight. Brew time matters. TDS matters. Taste matters most of all. Still, the one measurement that every barista can check quickly, without argument, is the amount of beverage that ends up in the server or cup. When that number drifts, every other conversation becomes less trustworthy.

1. Start with a target the whole team can repeat

Most cafés begin with a recipe written as dose and ratio. That is useful, but incomplete. If a brewer retains more water than expected, or if agitation changes, a ratio alone will not tell the next barista what actually happened in the cup. Add a beverage target in grams and treat it as part of the recipe, not as a side note.

At one London training room I visited last winter, staff were brewing the same Kenyan coffee at 18 grams to 300 grams of water. Some cups landed at 262 grams, others at 274 grams. The team described the coffee as “a little inconsistent.” That phrasing hid the real issue. The recipe was not being finished to the same beverage weight, so the cups could not reasonably taste alike.

• Write dose, total water, and final beverage weight on the same dial-in card.
• Record a preferred drawdown range, not a single heroic brew time.
• Choose one pouring routine for the shift and keep it visible.
• Use the same server or cup set when comparing brews.
• Recheck beverage weight after a paper or brewer change.

2. Watch for signals that show the recipe is holding

The best extraction targets show resilience under pressure. If two baristas can brew within four to six grams of the same beverage weight, and the cup remains within the same sweetness and acidity window, the recipe is likely serviceable. If one barista keeps landing wide of target, do not jump straight to grind adjustment. Review workflow first.

Stable recipes leave a pattern. The bloom looks familiar. The slurry bed settles at a consistent height. The final drawdown does not stall for one brew and race through the next. You do not need a romantic theory about extraction to notice these things. You need calm observation and a sheet that captures the same numbers every time.

One practical standard I recommend is a three-brew check: three consecutive cups by two different baristas, all within a narrow beverage window, with no major sensory drop. If cup two tastes sharp and thin while the numbers look correct, review pouring structure or water chemistry before rewriting the entire recipe.

3. Build service checks that fit a real café

Many brew guides fail because they ask for laboratory discipline in the middle of lunch rush. Strong cafés keep their checks short. They verify dose, beverage weight, and grinder setting at defined moments: open, pre-rush, and mid-afternoon. That rhythm is realistic enough to sustain.

Another useful habit is to separate correction decisions. If beverage weight is off, fix the pour or finish point first. If weight is on target but the cup tastes dull, then look at grind. When both are changed at once, the team learns nothing and the recipe loses its reference point.

The cafés that stay calm during service usually have one shared rule: no one describes a brew as “good” or “bad” without at least one measurement beside the opinion. That rule sounds strict, yet it protects staff. It moves the conversation away from blame and toward the state of the brew itself.

4. Keep the target alive after dial-in

A stable extraction target needs maintenance. Humidity shifts, grinder burrs age, and staff attention rises and falls across the day. The answer is not constant recipe churn. The answer is a short review loop that keeps the original target visible and asks whether reality still matches it.

If your brew bar can state dose, beverage weight, drawdown range, and expected taste profile at any point in the day, you are operating with control. If those answers change depending on who is speaking, your target is not stable yet.