Every grinder lies to you. Not maliciously — the numbers and click-stops are real, in the sense that they correspond to physical positions of the burrs. They just do not correspond to anything universal. My grinder's "18" is your grinder's "12" is somebody else's "medium-fine." The number is internal documentation that nobody outside of your kitchen can read.
This is fine, once you know it. The actual question — the one your tongue is trying to answer when you sip a brew that came out wrong — is not "what number should I have used." It is "what particle distribution did I just produce, and how did it interact with water at this temperature for this contact time." Once you can think in those terms, the number on the dial becomes a placeholder, the way a variable name is a placeholder for whatever value happens to be living at that memory address.
This piece is the technical version. If you want the personal-experience version — the three grinders I bought before settling on the one I kept — that lives over on HexRoast Field Notes. This one is about the physics.
/ 01What "grind size" actually is
When you grind coffee, you are not producing a single particle size. You are producing a distribution. A normal-curve-shaped one, with most particles clustering near a target size and outliers on both ends — fines (very small) and boulders (very large). The shape of that curve is what determines how your coffee will extract.
A good burr grinder produces a tight distribution: most particles within a narrow band, few outliers. A bad grinder (or a blade grinder) produces a wide, lumpy distribution: lots of dust, lots of chunks, very little in between.
This matters because extraction speed is roughly inversely proportional to particle size. Tiny particles extract fast. Large particles extract slow. When your distribution is wide, you have particles extracting at different speeds simultaneously, in the same brew. By the time the chunks have given up enough flavor to taste right, the dust has over-extracted into bitterness. There is no brew time that satisfies both populations.
/ 02The four flavor zones of extraction
Specialty coffee researchers have spent decades mapping how extraction percentage maps to taste. The numbers below are simplified, but they are the right ballpark:
Grind size is the most direct lever you have for moving across these zones. Coarser grind = slower extraction = lower final percentage. Finer grind = faster extraction = higher final percentage. Everything else (water temp, contact time, water volume, agitation) is a smaller knob on the same instrument.
/ 03Why the same setting tastes different on different beans
Even with a perfect grinder, two beans ground at the same setting will taste different. The reason is bean density. Lighter roasts are denser — they have lost less mass to water evaporation during roasting. Denser beans break differently under the burrs. The same physical gap between burrs produces a slightly coarser grind on a light roast than on a dark one.
This is why "I just dial in my grinder once" does not work for anyone serious about coffee. Each bean wants its own dial-in. A light Ethiopian needs a finer setting than a dark Sumatran on the same grinder, even though both are being brewed in the same V60.
The good news: once you have dialed in two or three beans across the roast spectrum, you can interpolate for new ones. A medium roast lands roughly between your light and dark settings. A natural-process bean usually wants slightly coarser than a washed bean of similar roast level (the fruit-forward flavors come out fast and over-extract easily).
/ 04Dialing in by taste, not by number
The right workflow for any new bean, on any grinder, looks like this:
- Brew at your default setting. Whatever your grinder's "medium pour-over" is. Time the brew. Note total contact time.
- Taste with intent. Is it sour and thin, or bitter and harsh? These are the two failure modes you are calibrating against.
- If sour and thin: grind finer. One click finer is usually right. Two if the sourness was severe.
- If bitter and harsh: grind coarser. Same logic.
- If balanced but flat: change brew temperature, not grind. Grind moves you across the extraction zones; temperature changes how those compounds taste once extracted.
- Adjust one variable at a time. Otherwise you cannot tell which knob fixed it.
Most beans dial in within three brews. Some difficult ones — light Ethiopian naturals, processed-method beans, very old beans — can take five or six.
/ 05Common dial-in mistakes
// 05.1Adjusting too much at once
Going from "10" to "15" because the cup tasted slightly off is too aggressive. One click of an espresso grinder can change extraction by 1–2%. One click on a hand grinder is more like 0.5–1%. Move in small steps. Patience is a brewing skill.
// 05.2Confusing acidity with sourness
Acidity is a desirable bright quality — citrus, stone fruit, berry notes. Sourness is unbalanced and aggressive — it makes you wince. A great Kenyan coffee is acidic. An under-extracted Kenyan coffee is sour. The difference matters because the fix is opposite: more acidity might mean grinding coarser to preserve brightness; sourness means grinding finer to balance it with sweetness.
// 05.3Forgetting to clean between bean changes
Every grinder retains some grounds in the burrs and chute between bags. Switching from a dark roast to a light one without purging means your first cup is contaminated. Run 5–10 grams of the new bean through and discard before brewing for real. Yes, it feels wasteful. It also fixes a problem most home brewers do not realize they have.
/ 06The minimum-viable grinder thesis
You cannot dial in a brew that the grinder cannot produce consistently. The cheap-grinder failure mode is that even at the same setting, particle distribution changes from grind to grind. You do not have a moving target — you have a target that is itself moving while you aim at it.
The threshold where this stops being true sits at about $80–100 for a hand grinder and $250–300 for an electric. Below that, you can dial in by feel as much as you want; the grinder is going to vary on you. Above that, you have a stable platform you can actually calibrate against.
If you take only one thing from this piece: the grinder is the most important piece of equipment in your setup, full stop. Better than the kettle, better than the brewer, better than the beans. A great bean ruined by an unstable grinder will taste worse than a mediocre bean ground precisely. The math does not care what the marketing on your espresso machine says.
/ 07What this means in practice
Buy the best burr grinder you can afford. Ignore the number settings as labels — treat them as relative position markers only. Learn to dial in each bean by taste. Adjust one click at a time. Clean the burrs when you switch beans. Trust your tongue more than your spreadsheet.
The math underneath is real, but the calibration loop is mostly sensory. Two months of attentive dialing-in across five or six different beans will teach you more about extraction than any amount of reading. Including this piece. Especially this piece.
We grind every bag the day before it ships.
If you'd rather not memorize extraction physics every morning, the easiest fix is starting with beans that haven't been sitting on a shelf for six months. The waitlist is the first to know when each new roast is ready.
Join the waitlist →