Recipe notebooks and dialling in systematically

Most people brew coffee, notice it is not quite right, change something, and forget what they changed.

A record turns that into a process that converges.

What to record

Per brew, six or seven items.

Coffee. Name, roaster, roast date.

Days since roast. Calculate it. Coffee behaves differently at day five and day twenty-five, and this explains a lot of unexplained variation.

Dose, in grams.

Grind setting. Whatever your grinder's notation is.

Water, weight and temperature.

Time, total brew time or shot time.

Result, in a few words. Sour, bitter, flat, balanced, thin, sweet, harsh.

And the change you would make next time.

That last item is what makes it a process rather than a diary.

The systematic approach to a new coffee

Rather than guessing repeatedly.

Start from your standard recipe for that method. Note the result.

Change one variable, in a meaningful step. Note the result.

Small changes produce ambiguous results. Move enough to taste a difference.

Continue in the same direction until it is clearly worse, then come back.

Deliberately overshooting is the fastest route to understanding the range, and most people stop before they find the boundary.

Then adjust a second variable if needed.

Three or four brews typically settles a new coffee. Guessing takes a week.

Which variable to change first

Grind, always. It has the largest effect and it is the intended adjustment.

Then ratio, if the strength is wrong but the balance is fine.

Then temperature, if you cannot get where you want with grind — particularly useful when your grinder's steps are coarse.

Then time and agitation, which are usually consequences of the above rather than independent choices.

Reading your own records

After a few dozen entries, patterns appear that no guide could have told you.

That your grinder needs two clicks finer for washed Ethiopians than for Brazilians.

That a bag needs a click finer around day eighteen.

That your pour over is consistently better at ninety-four than ninety-two.

That the coffee you thought you disliked was actually being brewed too coarse.

These are specific to your equipment, your water and your palate, which is exactly why they cannot be looked up.

Building standard recipes

The output of the process.

Eventually you have a starting recipe per method: dose, grind, ratio, temperature, time.

New coffees start from that and need small adjustments rather than exploration.

Write the standard recipes somewhere permanent, separately from the running log.

When to start recording again

You do not need to log every cup forever.

Resume when: you buy a new grinder or brewer; you change water; you move somewhere at a different altitude; a coffee behaves unexpectedly; or something has gradually got worse and you cannot say why.

That last case is common. Equipment drifts, burrs wear, scale accumulates, and a slow decline is invisible without a reference point.

The format

A paper notebook by the grinder is the format people actually use.

Apps exist and are more searchable, and the friction of unlocking a phone means fewer entries get made.

Whatever you will actually use is the right choice. A half-completed spreadsheet is worth less than a scruffy notebook that is up to date.

The espresso version

More variables and more benefit.

Record dose in, yield out, time, grind setting, brew temperature, and the taste.

Espresso changes noticeably as a bag ages, and the record makes that visible rather than mysterious.

It also means that when you return to a coffee you liked six months ago, you have the settings.

Blind tasting yourself

Worth doing occasionally.

Brew the same coffee two ways, have someone else pour them into unmarked cups, and taste.

Expectation affects perception substantially. People reliably prefer the cup they believe was made the more careful way, whether or not it was.

A few blind comparisons will tell you which of your adjustments actually matter and which are ritual.

Some are ritual, and there is nothing wrong with that as long as you know which is which.