Decaf grows up
Short answer: decaf was bad for two reasons — the removal processes damaged the coffee, and nobody sent good green coffee to be decaffeinated in the first place. Both have improved, and the second mattered more. What most bars have not changed is how they make it: decaf beans roast, grind and extract differently, so a decaf run on the house setting is wrong before it starts.
Decaf is usually the least examined line on a coffee menu: listed last, sold without enthusiasm, made from whatever is in the second hopper on the house recipe. The commercial argument for taking it seriously is stronger than the prejudice against it — and the prejudice is a memory of something that has changed.
Why was decaf bad, and what actually changed?
Two causes, independent of each other, only one of them technical.
The processes were crude. Early industrial decaffeination was aggressive — long exposure, high heat, water and solvents removing far more than caffeine. Aromatics and sugars went with it, and no skill at the bar recovers what has been taken out. Processes have improved: shorter contact times, better selectivity for caffeine, and mechanisms designed to return the flavour-bearing fraction to the bean rather than discard it.
Nobody sent good coffee to be decaffeinated. The larger reason, and commercial rather than technical. Decaffeination costs money per kilo, adds handling and loses weight; spending that on a high-scoring lot to sell to a customer assumed not to care looked like burning margin. So the coffee routed to decaffeination was already the least interesting: low-grade, defect-heavy, sometimes old crop. The process subtracted from a poor starting point, the cup confirmed expectations, and that justified sending poor coffee next time.
The loop broke when enough buyers began asking for a decaf worth drinking. Good green coffee now goes to be decaffeinated, and that has done more for the cup than the process improvements did. Ask about the green coffee first and the process second.
How do the three process families work?
All decaffeination happens on green coffee, before roasting, and all of it faces one problem: caffeine is soluble, but so is much of what makes coffee taste of anything. The families differ in what the caffeine is drawn into and whether the rest is put back.
Solvent-based methods. The beans are softened with water or steam so the caffeine can move, then brought into contact with a food-grade solvent that binds caffeine strongly and other soluble compounds weakly. In direct versions the solvent contacts the beans; in indirect versions the beans are soaked in water, the water is treated with the solvent, and the water — still carrying the flavour compounds — goes back to the beans. The permitted solvents and their limits are regulated wherever coffee is sold, and because those solvents are highly volatile and roasting takes the bean far above their boiling points, residues in the finished cup are negligible. Done well the method is selective, and the cup holds its origin character better than a poorly executed water process does.
Safety and perception are different questions, though. A customer asking whether their decaf was made with a solvent is not asking a scientific question, and "it is permitted and regulated" is a true answer that does not land well. Know which process your decaf used and say so plainly. Your roaster knows; ask.
Water-based methods. The beans are soaked in hot water, which pulls out the caffeine along with a great deal of everything else. The loaded water passes through a filtration medium that captures caffeine while letting most of the larger flavour molecules through, and that flavour-bearing water goes back to the beans, which reabsorb it. Later batches use water already saturated with coffee solubles, so on balance only caffeine leaves. Gentle, no solvent, and it sells well on the label. The cup is rounder and softer, with some loss at the top end — the brightest aromatics are hardest to put back.
CO2-based methods. Carbon dioxide held under high pressure behaves as a selective solvent: it dissolves caffeine readily and most flavour compounds poorly. Moistened beans are exposed to it, the caffeine-loaded CO2 is drawn off and stripped in a separate vessel, and the CO2 is recycled. Selectivity is the strength — less collateral extraction, so less to put back — and cup clarity is generally good. The plant is expensive, which has historically meant large volumes.
The best predictor of cup quality is not the process, though. It is the green coffee going in.
What does decaffeination do to the bean that matters at the bar?
Decaffeination is a physical process as well as a chemical one, and the bean it produces behaves differently.
It is more fragile. Soaking, pressure and heat alter the cell structure, and the bean comes out more porous and more brittle — which affects grinding, and how fast it stales once ground.
It looks darker than it is. Decaffeinated green coffee is typically brown rather than green, and browns further and faster in the roaster. Colour, the roaster's most immediate cue, becomes unreliable: the bean reads as developed before it is, and first crack is often quieter.
It roasts differently. Lower moisture and altered sugars mean heat moves through it differently and browning runs ahead of development, so decaf takes less energy to reach a given colour and is easier to scorch — and how a bean was roasted changes what it does in the basket, per how roast profile changes extraction.
It grinds differently. A more brittle bean shatters rather than fractures cleanly, producing more fines at any given burr gap. The resulting bed is finer and more restrictive than the setting suggests, and the fines fraction governs flow and channelling more than average particle size does — see particle size distribution explained. Decaf is usually less dense too, so a volumetric dose delivers less coffee by weight.
It extracts differently. More porous, more fines, more surface area: solubles come out faster. On a recipe built for the house espresso, decaf runs long on extraction and short on structure, arriving hollow or harsh depending on which way the fines push the flow. The relationships involved are the ordinary ones, covered in the complete guide to espresso extraction variables.
So: a decaf pulled on the house grind setting and the house recipe will not be right. Most bars do exactly that, which is why most bar decaf is poor — and it has nothing to do with the coffee having been decaffeinated. It is the same error as running a new single origin on yesterday's setting. Dialling decaf in separately usually means going a step coarser and often dropping the dose slightly, but the setting has to be found rather than predicted, with the caveats in can you dial in grind size by taste alone.
How do you actually run two coffees at one bar?
A second coffee needs a second grind setting, and a single grinder holds one.
Purge and re-dial for each decaf order. The default, and it does not work. Changing the setting, clearing the chamber and finding the setting again takes minutes and wastes coffee, so nobody does it during service. The setting never moves, the decaf comes out on the espresso recipe, and the drink is poor.
A dedicated second grinder. The clean answer: two hoppers, two settings, no changeover, decaf dialled in and left alone. The costs are counter space and one more thing to clean and calibrate. The parallel logic for splitting production across two machines is in one machine or two: splitting coffee production, and where the grinder goes is a workflow question before a purchasing one — maximising barista workflow.
A single-dose grinder for decaf. A good middle path at low volume. No hopper: beans are weighed in per shot from a sealed container, which suits an item ordered a few times a day and keeps it fresh — hopper staling is a real cost at low throughput (keeping coffee fresh in the hopper). Slower per drink, and it needs a scale, but it takes less counter than a second full grinder.
Pre-ground decaf. The pragmatic compromise, with a real quality cost. It comes at a fixed distribution you cannot adjust, and decaf's porosity means it stales faster than pre-ground caffeinated coffee would — see whole bean vs pre-ground: what happens after grinding. The ceiling is lower and there is no getting round that. But it is consistent, takes no counter and needs no changeover, and a competently made drink from pre-ground beats a badly made one from whole bean. For a bar selling a handful of decafs a day it is often the right answer.
What is the commercial case?
Caffeine is the constraint on when a customer will accept a coffee, which makes it a trading-hours constraint.
From mid-afternoon onward, a growing share of customers decline a coffee they would otherwise want, because of what it will do to their evening. The same happens at the table after a meal — the guest who would have ordered an espresso orders nothing, and the dessert often goes with it. That is the argument in coffee in restaurants: the after-meal espresso, and decaf is the line that converts the refusal. It also underwrites the late-evening menu: coffee-based desserts as a menu opportunity and coffee cocktails and the evening daypart both depend on a coffee a customer will drink at ten o'clock.
Beyond the clock, some customers avoid caffeine for medical or personal reasons, and they order coffee-shaped drinks somewhere. Whether an individual should consume caffeine is a question for a qualified professional, not for a menu; the operator's part is narrower — have something good, and know what it is.
A loyalty argument follows. A customer who orders decaf orders it every time, has no alternative within the category, notices when it is bad, and remembers who makes it properly.
Decaf is not caffeine-free
Decaffeination removes most of the caffeine, not all of it. A small residual amount remains, and it accumulates across several cups. A customer avoiding caffeine entirely should be told that rather than reassured, and a bar that says "no caffeine" is making a claim it cannot support. "Decaffeinated — a small amount of caffeine remains" is accurate and takes one line. There is a distinct middle ground as well: reduced-caffeine coffees and half-caff blends, covered in low caffeine and half-caff: a genuine category.
How should decaf appear on the menu?
As a drink, not as an accommodation.
Do not apologise for it. Staff language shapes expectation more than menu design does. "We do have decaf" invites low expectations; naming it the way the espresso is named does not. A good decaf properly dialled in is worth a sentence of description in the same register as everything else — and staff who mention it when offering coffee late in the day will sell it, while staff who wait to be asked will not.
Where it sits says what you think of it. Decaf as a footnote, or available only on request, tells the customer the bar considers it a concession — and means the customers who would order it never learn it exists. The line never earns its keep, and the lack of sales then confirms the decision not to invest in it: the same closed loop that kept decaf bad in the first place.
Where the case against sits
A bar selling very little decaf should not buy a second grinder for it. It sits idle, occupies counter a busy bar needs for something else, and the coffee in it stales between orders. Pre-ground handled well, or a single-dose grinder if volume justifies it, is proportionate. Buying a grinder to signal seriousness about a line selling a few cups a day is a poor use of money and counter alike.
Good decaf costs more — in green terms, and in attention: a second setting to maintain, a second thing to taste, a second stock line to keep fresh at low turnover.
A bar unwilling to dial it in separately should not claim to take decaf seriously. A good decaf through the house recipe is a bad drink made from expensive coffee, which is worse than a modest decaf made competently.
Adding the line changes nothing on its own. The daypart argument above depends on the decaf being good. Listing it without changing how it is ground, stored and made does not extend the trading day — it gives the customers who order it a reason to stop.
Frequently asked questions
Is solvent-decaffeinated coffee safe to serve? The permitted solvent processes are widely used and regulated, and residues in the finished cup are negligible — those solvents are volatile and roasting takes the bean well past their boiling points. Perception is a separate matter from safety: know which process your decaf used, and answer plainly if asked.
Which decaffeination process tastes best? Process matters less than the green coffee. CO2 and well-executed solvent methods tend to preserve origin character; water methods give a rounder cup with some loss at the top end. A good lot through a serviceable process beats a poor lot through the gentlest one.
Do we need to dial decaf in separately? Yes. It is more brittle, makes more fines at the same setting, is usually less dense and extracts faster. Running it on the house recipe is why bar decaf has a bad reputation.
Is decaf caffeine-free? No. A small residual amount remains, and a customer avoiding caffeine completely should be told that rather than reassured.
This article is part of the Vea Group knowledge base. Vea Group S.p.A. is an Italian designer and manufacturer of professional and premium coffee machines, with heritage dating to 1919 and production in Chignolo d'Isola, Italy, and Suzhou, China.