Pour Over Coffee: Four Variables, Only One Real Dial

Every pour-over guide hands you the same four knobs — ratio, grind, water temperature, pour — and a number for each. 1:16. Medium-fine. 205F. Three pours, thirty seconds apart. The implication is that missing any of them ruins the cup equally.

They do not. And they are not even the same kind of control. One is arithmetic you can do before tasting anything, one is the dial you actually turn, one is a gate you either clear or do not, and one is a mixing input with no published sensory standard behind it. Sorting them by type rather than importance is what stops a pour-over routine feeling like a ritual you might be performing wrong.

Four variables, four different jobs

The frame: two of these you can reason about before tasting, one you can only find by tasting, and one you mostly just need to not get badly wrong.

Ratio: the only one you can predict

Brew ratio is the one variable you can work out on paper. Add more coffee to the same water and the drink gets proportionally stronger: you have not changed how much of each particle dissolves, only how many are in the bed.

The Specialty Coffee Association's home brewer standard pins this down. Every certification test in SCA Standard 310-2021 uses a fixed brew ratio of 55 g of coffee per 1.000 kg of water (§7.1.5), with the target defined as a brew strength of 1.15 to 1.55% dissolved solids at that ratio.

55 g/kg works out to 1:18.2 — weaker than the 1:16 usually quoted as gospel. That is a laboratory constant for certifying machines rather than a taste recommendation, but it does mean the famous number is convention, not specification.

Our coffee ratio calculator will do the arithmetic for any dose. Then stop moving it: a ratio that changes every morning makes every other variable unreadable.

Grind: the dial the standard itself turns

Here is the detail that reorders the whole list. To bring a brewer into the target window, the SCA does not adjust the ratio, the temperature or the water flow. It adjusts the grind, and only the grind, within the particle limits of Table 2 (§7.4).

That table is specific. For a conical basket the final distribution must be less than 12% by mass above 850 µm, at least 68% between 425 and 850 µm, and 20% or less below 425 µm.

At your counter that reduces to two moves, and they are the only two:

• Finer — more surface area, slower drawdown, more extraction. Reach for it when the cup is sour, thin, or finishes abruptly.
• Coarser — less surface area, faster drawdown, less extraction. Reach for it when the cup is bitter, drying, or harsh.

Move one step at a time and taste in between. Settings do not transfer between grinders, which is why our grind size chart is built around texture and drawdown rather than numbers.

Temperature: a gate, not a dial

SCA 310-2021 §6.2 is strict here: the slurry must reach 90 to 96 °C by the moment a third of the water has been sprayed onto the bed, and stay there for the rest of the cycle. A brewer that cannot is non-compliant.

That reads like precision. Then UC Davis measured it. Batali, Ristenpart and Guinard brewed drip coffee at 87, 90 and 93 °C, adjusting grind and brew time so every batch landed on the same brew strength and extraction yield, then put the results in front of a trained descriptive panel. Brew strength had the largest sensory effect and extraction yield a smaller but real one — while temperature had no appreciable impact.

Read it this way: temperature is a gate to clear, not a dial to fine-tune. Where the brew ends up matters far more than the route it took.

In practice that is three lines of instruction:

• Boil, then pause. The familiar thirty-second wait is a convention rather than a measured rule, but with an ordinary kettle and a room-temperature dripper it lands you inside the band.
• Treat a variable-temperature kettle as a convenience, not a flavour upgrade.

The pour: a mixing input with no sensory standard

The pour attracts the most choreography and carries the least published evidence as a taste control. What it demonstrably does is mix. Physicists at the University of Pennsylvania imaged the inside of a brewing cone and described a cycle nobody had documented: the falling jet erodes a crater in the bed, grounds pile up at its edge, the wall collapses in a miniature avalanche, and it starts again — continuous mixing for as long as you pour. Two useful things fall out of that.

What the imaging supports:

• A coherent, laminar stream. This is the real argument for a gooseneck spout, and it is why the researchers frame the result as reaching a given strength with fewer beans.
• A little height. More height drives more agitation, and more agitation keeps the bed mixing.

What it does not support:

• Pouring from very high. Past a point the stream breaks into droplets and drags air into the bed instead of digging into it.
• Choreography as a flavour claim. The imaging used transparent silica gel rather than coffee, the confirmation runs measured dissolved solids, and nobody tasted anything.

For the spout question on its own, gooseneck vs regular kettle is the longer answer.

What to change when it tastes like that

Grind and ratio between them account for nearly every complaint on that list. That is the same ranking, arrived at from the other end.

What we could not source

The canon carries three numbers everyone repeats: 1:16, a thirty-second bloom, and a two-and-a-half to three-and-a-half minute brew. We could not trace any of them to a published standard or a peer-reviewed measurement. SCA 310-2021 specifies temperature, a test brew ratio, water chemistry and a grind distribution, and says nothing about bloom time or pouring pattern.

That does not make them wrong — defaults thousands of people converge on are usually defensible. It does mean that if you are treating a thirty-second bloom as a rule you keep failing, you are being harder on yourself than the evidence is.