Tool
Extraction yield calculator
Dose, beverage weight and TDS give the proportion of the coffee that actually dissolved — which taste alone frequently cannot distinguish from simple dilution.
Extraction yield
19.8%
The drink is in a fairly ordinary range at 1.35% dissolved solids, and 19.8% of the coffee dissolved. Those are separate things: concentration is set by the ratio, and extraction by grind, time and agitation.
With those allowances the yield lies between 19.3% and 20.4%. Most of that width comes from the refractometer reading. It is a worst-case span, with every error falling the same way, and not a statistical interval; two readings that differ by less than its width have not been shown to differ.
Within the range most filter brewing guidance treats as typical. That is a convention derived from tasting panels, not a law: a brew inside it can still taste bad, and one outside it can be excellent.
How this was worked out
Dissolved coffee = beverage weight × TDS
440 × 1.35% = 5.94 g
Extraction yield = dissolved coffee ÷ dose
5.94 ÷ 30 = 19.8%
The percentage of the dry coffee that ended up dissolved in the cup. It says nothing about whether the coffee is good.
- What to change first
- Nothing, on this number. If you want the cup stronger or weaker, change the ratio: that moves strength a great deal and extraction comparatively little. If the flavour itself is off, change the grind.
- Keep the same
- Dose, water and method. Change the grind alone and measure again in the same way: the same point after stirring, the sample cooled as the instrument's instructions say. A second reading only means something if one thing differs from the first.
- What this cannot tell you
- A refractometer infers dissolved solids from how the sample bends light, and the reading moves with sample temperature and with fine particles left in the sample. A small difference between two yields can be the method and not the brew, so read trends over several brews. The ranges used here are conventions from tasting panels, not limits.
An evidence-based range, not a measurement
What to do with the number
- Before acting on it: what a reading can and cannot tell you: Measuring strength and extraction: what a reading cannot do
- You want to know how far the scale and the instrument can move it: Measuring strength and extraction: where error enters
- It is a filter brew and you want to change the next one: Dialling in filter coffee: what each control does
- It is espresso and you want to change the next shot: Dialling in espresso: strength and extraction
- The number looks fine and the cup does not: Make my coffee better
How it works
Beverage weight multiplied by TDS gives the mass of dissolved coffee in the cup; divided by the dose, that is the extraction yield. Beside the figure the tool shows a span: the yield recomputed with each of the three measurements off by an allowance you set, all in the direction that lowers it and then all in the direction that raises it. With a dose weighed to a tenth of a gram, a drink weighed to a gram and a reading good to 0.03 points, a 19.8% yield becomes 19.3% to 20.4%. The tool also checks the result for plausibility, because the commonest input error — using brew water instead of beverage weight — produces a number that looks reasonable and is a quarter too high.
What the inputs mean
- Beverage weight
- The mass of liquid in the cup, weighed. Not the water poured in: a coffee bed retains roughly twice its own weight, so the two differ substantially.
- TDS
- Total dissolved solids, as a percentage, from a refractometer. It measures concentration — how strong the drink is — and on its own says nothing about extraction.
- Extraction yield
- The proportion of the dry coffee that dissolved. A weak drink can have a normal yield (it is simply dilute) or a low one (it is under-extracted), and those need opposite fixes.
Worked examples
A 30 g pour-over
- Dose 30 g
- Beverage 440 g
- TDS 1.35%
Result: 5.94 g of coffee dissolved, which is a 19.8% yield — inside the range filter guidance treats as typical.
The same brew, measured against brew water by mistake
- Dose 30 g
- Beverage 500 g entered
- TDS 1.35%
Result: 22.5% — apparently over-extracted, and wrong by the water the grounds kept.
This is why the tool checks plausibility rather than simply returning the arithmetic.
Common mistakes
Entering brew water instead of beverage weight
The grounds retain a large share of the water, so the figure comes out roughly a quarter too high — enough to turn an under-extracted brew into an apparently ideal one.
Reading a difference smaller than the span as a difference
Two brews whose yields differ by less than the width of the span have not been shown to differ. On a small dose the span is wider than most people expect.
Treating a number inside the band as a good cup
The band is a convention derived from panel preference. Plenty of coffee inside it tastes unpleasant and plenty outside it is excellent.
Comparing an espresso reading against a filter target
They are different measurements on different beverages with different reliability. Compare espresso readings only against your own other espresso readings.
What the result does not mean
- Extraction yield does not measure quality. It measures how much of the coffee dissolved, which explains what you tasted rather than judging it.
- This is not a laboratory measurement. A handheld refractometer, an uncooled sample and an uncalibrated instrument each move the figure more than most people expect.
What this does not account for
- Assumes the refractometer was calibrated and the sample cooled to its stated range. Both are common sources of inconsistent readings.
- The allowances for each measurement start as assumptions about a kitchen scale and a hand-held refractometer. They are not the specification of any instrument, and the span they give is a worst case in which all three errors fall the same way. It is not a statistical confidence interval.
- The span covers weighing and reading only. It knows nothing about a calibration that is off, or about oil, gas and fine particles in an unfiltered sample, which published espresso studies remove by filtering, spinning or diluting before they measure.
- Filter bands come from widely used brewing guidance derived from tasting panels, not from a law of nature.
- Espresso is deliberately not banded: its readings are the least reliable a refractometer makes, and published ranges for it are contested.
Where to go next
Learn the technique
Background reading
Equipment involved
Drinks and methods
Practise this in a lesson
Become a Barista · Level 9
The task: Enter a dose and yield you use, and try three TDS values across the tool's range to see how extraction yield responds. Then hold TDS constant and change the yield by five grams.
Become a Coffee Expert · Level 8
The task: Enter a filter recipe — for example 15 g of coffee and a beverage weight of about 220 g — and find the TDS that gives 20 per cent extraction. Then keep that TDS and halve the beverage weight, as if you had brewed with half the water.