How to Dial In Espresso: The 18 g In, 36 g Out Method

Dialling in is the process of adjusting grind size until a fixed dose of coffee yields a fixed weight of espresso in a fixed time. The recipe to start from is 18 g in, 36 g out, in 25 to 30 seconds, which is a 1:2 brew ratio. Everything else on this page is detail around that one sentence, and the single most important rule is that you change grind alone until the time lands, then stop touching it.

That top pick is not a courtesy placement. You cannot dial in anything without a grinder that adjusts finely enough inside the espresso range, because the entire method is built on making small, repeatable changes to a single variable. A grinder with twelve coarse steps across its whole range gives you perhaps two settings that fall anywhere near espresso, and the gap between them is a ten second jump in shot time. There is no adjustment available between "far too fast" and "stalled", so the shot you want sits in a gap the machine cannot reach. The Baratza Encore ESP at $199.95 exists precisely because of this: it puts twenty micro-adjustments inside the espresso range, which is the difference between dialling in and guessing. The second requirement is a scale reading to 0.1 g at $31.99, because a dose you have not weighed makes every other measurement on the page meaningless.

What does dialling in actually mean?

Espresso forces hot water through a compressed bed of coffee at roughly nine bar of pressure. The coffee resists, the water pushes, and how quickly the water gets through is decided almost entirely by how finely the coffee is ground. Grind finer and you create more surface area and smaller gaps, so the water moves slower and dissolves more. Grind coarser and the water moves faster and dissolves less.

Extraction yield is the percentage of the dry coffee that ends up dissolved in the drink, and the conventional target window is 18 to 22 percent. Below it the shot reads sour, thin and salty because the sweet compounds have not come out yet. Above it the shot reads bitter, drying and ashy because compounds you did not want have come out too. Dialling in is the act of steering the extraction into that window using the only lever that meaningfully controls it.

You cannot measure extraction yield at home without a refractometer, and you do not need one. Shot time is a proxy for it that is accurate enough to be useful, provided the dose and the yield are held constant. That is the whole reason the method fixes two numbers and varies one: with 18 g in and 36 g out held still, time tells you what the grind is doing. Change the dose as well and time tells you nothing, because you have moved the target and the arrow at the same time.

What are the only three variables you should change?

Grind size, dose and brew temperature. That is the complete list for a home machine, and they are not equal partners. Grind size is the primary control and does the overwhelming majority of the work. Dose is a secondary adjustment worth roughly half a gram in either direction. Brew temperature is a fine trim, usually one or two degrees, and on many machines it is not adjustable at all.

Everything else people list as a variable is either a consequence of these three or is not a variable in the first place. Shot time is a consequence of grind. Pressure is fixed by the pump and the over-pressure valve on almost every home machine, and yes, the box says fifteen bar, but the valve regulates it down to roughly nine at the puck. Tamp pressure is not a useful variable at all past about twenty pounds. Pre-infusion is a real variable on machines that offer it, and it belongs to a later stage of the hobby than a first dial-in.

The reason to keep the list this short is arithmetic. With one variable you learn something definite from every shot. With two, a result that moves in the wrong direction leaves you unable to say which change caused it, so you have spent coffee to learn nothing. Work through the dose and yield calculator before you start and write down the target, because a target you did not write down quietly moves during a session.

Why do you change grind first and everything else second?

Because grind is the only variable with real authority over flow rate. Moving the dose from 18 g to 18.5 g changes shot time by a second or two. Moving the grind three steps finer can change it by ten. If the shot is running at seventeen seconds, dose and temperature cannot rescue it, and reaching for them first is how people end up with a notebook full of changes and a shot no better than the one they started with.

There is also a practical ordering argument. Grind is the one adjustment you have to make again every time the coffee changes, because beans differ in density, roast level and age. Dose and temperature, once you find settings you like, tend to stay put for months. Learning to move the thing that changes constantly, and to leave the things that do not, is most of what separates a fast dial-in from a slow one.

When you do move grind, move it decisively while you are far from target and gently once you are close. From seventeen seconds, three or four steps finer is a sensible jump. From twenty six seconds, one step is plenty. The grind size chart shows where espresso sits relative to every other brew method, which is worth a look if you are unsure whether your grinder even reaches espresso fineness.

What does a sour shot mean, and what does a bitter shot mean?

Sour means under-extracted. The water got through too quickly, and the acids in coffee dissolve before the sugars do, so what you are tasting is the front half of an extraction that never finished. It presents as sharp, thin, sometimes salty, and it makes the sides of your tongue contract. The fix is to grind finer, which slows the water down and lets the rest of the extraction happen.

Bitter means over-extracted. The water sat in the bed too long and kept dissolving past the point where the pleasant compounds ran out. It presents as ashy, drying, lingering, with a burnt edge at the back of the throat. The fix is to grind coarser. People confuse bitterness with strength constantly, and they are unrelated: a very strong shot can be sweet, and a weak one can be viciously bitter.

The table below is the working version of that logic, extended to the failure modes that are not simply sour or bitter. Read the symptom, make the one change, pull again.

Two rows in that table are not grind problems and are the ones most often misdiagnosed as grind problems. Spraying and one-sided flow are channelling, which is water finding a low-resistance path through the bed. Grinding finer to fix channelling makes it worse, because a finer grind raises the resistance everywhere except in the channel, so more of the flow goes down the channel. The extraction troubleshooting chart separates the two failure families in more detail.

How long should a shot take, and what if the time is right but the taste is not?

25 to 30 seconds from the moment the pump starts to the moment you stop it at 36 g. Some machines have a few seconds of pre-infusion at low pressure before the pump ramps up, and it is conventional to include that in the total. What matters far more than which convention you pick is that you use the same one every time, because you are comparing shots against each other, not against a standard.

Now the hard case: the time is right and the shot still tastes wrong. This is where most people get stuck, and there are only four explanations worth checking, in this order.

First, channelling. A channelled shot can hit exactly twenty seven seconds while tasting simultaneously sour and bitter, because part of the bed over-extracted and part barely extracted at all. The taste is the average of two failures rather than a single balanced result. Distribution is the fix, not grind. Run the dry grounds through a WDT distribution tool at $32.69 to break up clumps, then tamp level with a calibrated spring tamper at $42.99. The puck prep guide works through the whole sequence.

Second, the basket. A stamped basket from the box has uneven hole spacing and an inconsistent open area, so the bed drains faster in some places than others no matter how well you prepare it. A precision basket such as the IMS 18 g Baristapro at $36.92 has laser-cut holes with even spacing, and it is the cheapest change that measurably improves evenness on almost any machine.

Third, the coffee itself. Espresso wants beans roughly seven to twenty one days off roast. Younger than that and carbon dioxide from roasting disrupts the flow through the bed, which produces erratic timing and a foamy, hollow shot. Older than about six weeks and the coffee has staled, which reads as flat and cardboard-like regardless of what you do to the recipe. A forgiving blend such as Lavazza Super Crema at $21.49 is the right thing to learn on, precisely because it tolerates a dose or grind error while you build the habit.

Fourth, the ratio. 1:2 is a starting point, not a law. A dark roast built for milk often tastes better pulled shorter, at 32 to 34 g out, which raises the strength and lowers the extraction. A light single origin often wants 40 to 45 g out, which is a lungo and pushes the extraction higher. The brew ratio calculator will convert any dose and yield pair into the ratio so you can record it properly rather than remember it vaguely.

How many shots does dialling in a new bag take?

Three to six for a bean you have used before, five to ten for something unfamiliar. Light roasts take longer than dark ones because they are denser and resist extraction more, so the useful grind window is narrower. Here is what a real session looks like, written out shot by shot, with a fixed dose and a fixed yield throughout.

Three things in that log are worth naming. The dose never moves, which is what makes the time column readable. The early adjustments are large and the late ones are small, because the relationship between grind steps and seconds is not linear and it steepens as you approach the stall point. And shot five exists on purpose: the session did not end when the shot got good, it ended when the shot got worse, which is the only way to know that shot four was the peak rather than a stop on the way up. Most guides quietly omit that step and leave readers permanently unsure whether one more turn would have helped.

Write the winning setting down. Grinder setting, dose, yield, time, bean and roast date, five numbers on a card taped to the machine. Next bag from the same roaster will land within a step or two of it, which turns a ten shot dial-in into a three shot one.

When do you change dose or temperature instead of grind?

Once the timing is stable and repeatable across three consecutive shots, and the taste is still off in the same direction every time. That last clause matters. If shot to shot results are still jumping around, the problem is consistency, not recipe, and adding another variable will bury the signal completely.

Dose is the adjustment for body and strength. Raising it to 18.5 or 19 g in the same basket packs the bed tighter, raises resistance and produces a thicker, more concentrated shot at the same yield. Lowering it to 17.5 g does the reverse. The limit is the basket: an 18 g basket overfilled to 20 g leaves no headroom under the shower screen, the puck gets compressed by the group itself, and you get the sideways spray that looks exactly like a channelling problem. Move dose in half gram steps and expect to nudge grind afterwards, since dose and grind both change resistance.

Brew temperature runs 90 to 96 C, which is 194 to 205 F. Lighter roasts want the top of that range because they are harder to extract, and darker roasts want the bottom because they extract easily and turn bitter when pushed. If your machine has a PID, adjust by one or two degrees at a time and give it a few minutes to settle before judging, because the reading stabilises long before the metal does. If your machine has no temperature control, temperature is not one of your variables, and you have two rather than three.

Yield is technically a fourth lever, and the honest way to describe it is as changing the recipe rather than dialling in the current one. Pulling to 32 g instead of 36 g at the same dose gives you a ristretto-leaning shot: stronger, less extracted, sweeter with dark roasts and often sour with light ones. Pulling to 45 g gives you a lungo: weaker, more extracted, more likely to expose bitterness. Both are legitimate. Just record which one you are doing, because a shot pulled to a different yield is not comparable with the ones before it.

What should you do before you accept that the grinder is the problem?

Check three things first, because all three are cheaper than a grinder and all three produce symptoms that look like grinder symptoms. Are you weighing the dose on a 0.1 g scale, or filling by eye and by the grinder timer? A timed dose drifts as beans age and as the hopper empties, which produces exactly the random shot-to-shot variation people blame on burrs. Are you distributing the grounds before tamping? Are you using a precision basket rather than the stamped one that came in the box?

If all three are in place and shots are still inconsistent, or if you find yourself stuck between two settings where one runs at twenty seconds and the next runs at forty, the grinder genuinely is the constraint. That is a real and common outcome, especially with grinders whose adjustment is stepped and whose steps were designed for filter coffee. The espresso grinder roundup covers the tiers where that gap closes, and the practical threshold is lower than most people expect: a grinder with genuine espresso-range micro-adjustment starts at roughly $199.95, which is less than the difference between two mid-range machines.

Whatever you buy, the method does not change. Eighteen grams in, thirty six grams out, 25 to 30 seconds, one variable at a time, and a written record of the setting that worked. That routine is what makes espresso repeatable, and repeatability, not equipment, is what separates a good morning shot from a lucky one.