Espresso Shot Troubleshooter

How it works

Every espresso shot gives you two signals: how long it ran and what it tasted like. Those two signals encode almost everything you need to know about extraction.

Time tells you about flow. Fast means under-extracted by default. Slow means over-extracted.
Taste tells you about result. Sour means under-extracted. Bitter means over-extracted. Balanced means stop.
Together they tell you the cause. Sour and fast = grind. Sour and slow = channeling. Bitter and fast = dose. Bitter and slow = grind.

The troubleshooter encodes the same logic experienced baristas use. It picks the highest-impact next step so you change one variable at a time instead of guessing.

How to use the troubleshooter

Pull a shot at your current setting and target yield.
Pick the dominant taste note — sour, balanced, or bitter.
Pick the time band the shot landed in.
Read the recommendation and apply that single change before the next pull.
Re-pull and re-taste. If the change moved the cup the right way, lock it in.

For target yields and time bands by ratio, use the espresso ratio calculator. For the full method behind the recommendations, read the 3-shot dial-in guide.

The nine-cell map

The full grid the troubleshooter walks. Useful as a reference once you've used it a few times.

Taste	Fast (under 22 s)	On target (22–32 s)	Slow (over 32 s)
Sour	Grind finer	Grind 1 click finer	Fix distribution
Balanced	Grind 1 click finer	Lock it in	Grind 1 click coarser
Bitter	Reduce dose 0.5 g	Grind coarser or drop temp	Grind coarser

When the troubleshooter is wrong

Beans are too fresh or too old. Coffee under 5 days or over 4 weeks off roast behaves differently. Adjust for freshness first — see resting coffee beans.
Water is the issue. Very soft water under-extracts; very hard water over-extracts. If shots are uniformly off regardless of grind, try bottled water at 50–150 ppm.
The basket is the issue. A worn or undersized basket channels at any dose. If a fresh basket fixes the problem, that was the variable.

What this troubleshooter does not do

It doesn't track shots over time. Every visit is a fresh one-shot decision. To see patterns — bag-age trends, basket comparisons, recurring channeling — log shots in Beany. The app applies the same decision logic and remembers what happened so you don't solve the same problem twice.

Frequently asked questions

How accurate is the recommendation?

The troubleshooter encodes the same decision tree most home baristas use after years of dialing in. It is correct in the common case, but it cannot see your machine, basket, or beans. Treat it as a strong first move, not a final verdict — re-taste before locking in the change.

What if my shot was both sour and bitter?

That combination almost always means channeling. The troubleshooter routes sour-and-slow shots to a distribution fix rather than a grind change. If a shot tastes sour AND bitter at the same time, default to fixing distribution before adjusting grind.

Should I really change grind by 2 clicks at a time?

It depends on the grinder. On most home grinders, 1 to 3 clicks (or notches) is the safe range for a noticeable shift. If your grinder has very fine adjustment, scale the recommendation up. If it has coarse steps, scale down. The principle holds: small, observable changes.

When should I stop dialing in?

When the same setting produces the same balanced shot twice in a row, you are dialed in. If you keep adjusting and the cup keeps changing despite identical inputs, the variable is no longer grind — it is technique. Stop dialing in and focus on distribution consistency.

Does this work for any espresso machine?

Yes. The decision logic is machine-agnostic. A lever machine, a pump machine, a single-boiler, a dual-boiler, and even a manual press all respond to the same dose / yield / time / taste loop. The number of clicks may differ, but the direction (finer / coarser / fix prep) is universal.

How is this different from just trying things?

The troubleshooter encodes the order of operations. Random experimentation tends to repeat the same mistakes — changing two variables at once, grinding finer when the issue was distribution, or chasing the grinder when the beans are stale. A fixed decision tree skips those traps.