Yes. The same coffee, ground the same way and brewed to the same recipe, tastes different on two different waters, and the number that decides how different is alkalinity rather than hardness.
Brewed coffee is water almost all the way down. Even at the very top of the mineral range the trade counts as acceptable, a liter of brewing water carries about a quarter of a gram of dissolved solids and is 99.975 percent solvent by weight. That quarter gram does the work. It changes how much acid comes out of the grounds, and it changes how much of that acid you still taste by the time the cup reaches you. Hardness counts the calcium and magnesium in the water. Alkalinity counts something else entirely, and alkalinity is the one that moves the cup.
What actually changes when you brew the same coffee on two different waters?
Everything except the coffee. You buy a bag, brew it at home and get something thin and dull, then drink the same roast at a cafe forty minutes away and find it bright and full. The grinder gets blamed first, then the brewer, then the roast date. The ingredient that changed by the largest volume is the one that never gets measured.
Put numbers on it. The Specialty Coffee Association of America's standard, Water for Brewing Specialty Coffee, revised on 21 November 2009, sets the target for total dissolved solids in brewing water at 150 mg/L, and counts anything from 75 to 250 mg/L as acceptable. That is the entire mineral budget a cup of coffee gets to spend, and most of the disagreement about brewing water is an argument about how to spend it.
Water is not a neutral carrier that shows up, collects flavor and leaves. It is a solvent with things already dissolved in it, and those things take part in the extraction. Dissolved calcium and magnesium carry two positive charges each, and in water they attach to the negatively charged flavor compounds in coffee and help pull them into solution. Bicarbonate does the opposite kind of work: it is a base, and it reacts with the acids the brew has just extracted. One set of minerals helps things out of the bean. Another set changes what those things taste like once they are out.
That is why water sits underneath every other brewing variable rather than beside them. If you change your ratio of coffee to water, you are changing how much solvent meets how much coffee. If you grind finer to pull more out, you are changing how fast that solvent can get at the grounds. Both of those assume the solvent itself is a constant. In a single kitchen over a single week, it roughly is. Between two cities, two buildings on different filtration, or one building before and after a service call on the filter, it is not.
The practical version: ratio and grind are the two variables most people adjust, and both of them are being measured against a third variable most people have never seen a number for.
What is alkalinity, and how is it different from hardness?
They are two separate measurements of two separate things, and they get swapped for each other constantly. Barista Hustle draws the line cleanly: "GH is general hardness, and measures the total amount of calcium and magnesium in the water. KH stands for Karbonathärte, which is the German for Carbonate Hardness." General hardness is a headcount of two metals. Carbonate hardness, which in water-report language usually appears as total alkalinity, measures how much acid the water can absorb before its pH starts to move. It is mostly bicarbonate.
Alkalinity is also not pH, and the confusion between those two is worth stopping on. pH tells you where the water sits on the acid-to-base scale at this moment. Alkalinity tells you how stubbornly it will stay there when you add acid to it. Scott Rao puts it in one line: "A solution can be very alkaline but have low alkalinity, and vice versa." A water can read a perfectly agreeable pH 7.4 on a test strip and have almost no buffering capacity, or read the same 7.4 and have a great deal of it. Those two waters will not brew the same coffee, and the strip cannot tell them apart.
Barista Hustle settles one more thing worth settling: "Since calcium and magnesium both have two positive charges, they behave in a very similar way, so can be considered interchangeable when we talk about hardness." For the purpose of a hardness number, a milligram of magnesium and a milligram of calcium are doing the same job.
Hardness still matters, mostly because it is what plates the inside of your equipment. We put the hardness buckets, and the conversion between milligrams per liter and grains per gallon, in our piece on how often to descale a commercial coffee brewer, because that is the number that sets a maintenance calendar. It is not the number that sets the flavor.
Why does alkalinity decide whether the cup reads bright or flat?
Because coffee is an acidic drink, and alkalinity is a measure of how much of that acidity the water will cancel before you ever taste it. Rao is blunt about the ranking: "Alkalinity is the single-most important factor in how water will affect coffee flavor." And then the mechanism, in his words: "Simply put, higher alkalinity makes coffee less acidic. Lower alkalinity makes coffee more acidic."
Barista Hustle describes the same thing from the drinker's side. "The alkalinity buffers out acidity in the coffee, and makes it taste less sour. Too much alkalinity will make the coffee taste flat." That is the whole span, in one sentence. A little buffering is balance. It is what keeps a sharp Kenyan from reading as sour. More of it is the reason a coffee you know to be lively arrives tasting like it has had a blanket thrown over it.
The chemistry underneath was set out by Hendon, Colonna-Dashwood and Colonna-Dashwood, published in the Journal of Agricultural and Food Chemistry in 2014 (62(21):4947–4950). On the role of bicarbonate they write: "Bicarbonate/carbonate interacts strongly with 5 (pKa = 2.66), neutralizing some of this less desirable acid." Compound 5 in their numbering is one of the harsher acids in the brew, so a measure of neutralizing is genuinely useful. The trouble is that bicarbonate does not stop at the acid you wanted gone.
Rao gives a working number for where he likes to land: "My personal preference for lightly roasted, well-developed beans is an alkalinity level of 30—40ppm." To put that beside something from the same page of the same subject, 40 mg/L of alkalinity is about a quarter of the 150 mg/L of total dissolved solids the 2009 standard asks for. The buffering fraction is small, it is a minority of what is already a tiny mineral load, and it is still the thing most responsible for whether the cup reads bright or dull.
One clarification that saves a lot of argument: soft water and low-alkalinity water are not the same claim. Softness is a statement about calcium and magnesium. Our own position on softened water has not changed, and it is that soft water "often leads to under-extraction and pale, flat coffees." Both can be true at once, because they are two different axes, and a water can be soft and well buffered or hard and barely buffered. That is precisely why one number will not cover it.
Does it matter whether the minerals are calcium or magnesium?
Here the honest answer is that four credible sources do not agree, and I am going to leave that on the page rather than pick the version that reads best.
Hendon's 2014 paper found magnesium the stronger extractor and said so: "Mg2+-rich water is paramount if increasing extraction yield is important." That is a clear finding, and it is a finding about yield. The same paper's next sentence answers a different question: "If the motivation is to achieve the best balance of flavors for a given lighter roast coffee, then both Ca2+ and Mg2+ do a comparable job, with Mg2+ having the added feature of preventing scale formation." Higher yield, comparable balance. Those are two different goals, and the paper is careful about which one it is answering.
Bratthäll, Figueira and Nording went at it again in 2024, in Heliyon (10(5):e26625), with gas chromatography and NMR. Their finding: "Experimental data implied that dissolved magnesium and calcium chloride salts did not affect the extraction of the acids." At drinking-water concentrations, the split people build recipes around did not show up in their measurements of acid extraction.
Jonathan Gagné, writing at Coffee ad Astra in 2018, named the gap directly. Magnesium, he wrote, is "believed to extract slightly different flavors, but to my knowledge there are not yet any lab tests to back this up." Rao's 2023 piece lands in the same place, noting that there is little consensus on the optimal level or on the ideal balance between calcium and magnesium.
So: one paper says magnesium wins on yield and ties on flavor balance, a later paper finds no differential effect on acid extraction at the concentrations we actually drink, and two of the most careful writers in the trade say the sensory evidence has not been produced. We have never run a controlled magnesium-against-calcium comparison on our own cupping table, so I have no result of my own to add to that list. What I can say is that the disagreement sits on a question far downstream of the one most people are actually facing, which is that they do not know their alkalinity at all.
What does Pensacola's water actually say?
We went and pulled our own utility's numbers to work our own water as the example, and the two figures that decide how a coffee tastes are not in the report. That is not an oversight by anyone in particular. The same hole is almost certainly in your own utility's report, and it is the most useful thing I can tell you about water.
Our water comes from the Emerald Coast Utilities Authority, which draws from the Sand-and-Gravel Aquifer. ECUA's 2021 Water Quality Report publishes a full system-wide averages table: sodium at 4.690 ppm, fluoride at 0.432 ppm, nitrate at 1.451 ppm, figures given to three decimal places. It does not carry hardness. It does not carry alkalinity. No pH, no total dissolved solids, no calcium, no magnesium. The 2024 report does not carry hardness or alkalinity either.
The reason is regulatory rather than mysterious. An annual water quality report is a compliance document for contaminants, and hardness and alkalinity are not regulated contaminants. Sodium gets three decimal places because somebody has to report it. Alkalinity gets nothing, because no rule requires it. A utility that treats water beautifully and a utility that treats it badly will both produce a report with that same hole in it.
There is one thing ECUA does publish that bears directly on the cup. Its 2024 report states the treatment plainly: "Calcium Hydroxide (lime) is added for pH adjustment; Phosphoric Acid is added for corrosion control in the distribution system and Chlorine is added for disinfection." Lime is calcium. That is a mineral decision, made deliberately, on behalf of the pipes, and every coffee brewed in this county inherits it. How much, and what it does to the finished number, the report does not say, so neither will I. Worth noting too that NAS Pensacola runs its own system, so some addresses in this area are not on ECUA water at all.
Which brings me to what we actually do about it, and what that costs. We tell every office coffee account the same three things: rinse the basket daily, wash the carafe, descale monthly. Gulf Coast tap water is hard enough that monthly is not cautious, it is arithmetic. What monthly costs is roughly half an hour of somebody's Monday, on every machine, every month, indefinitely, and I pay for that half hour because I have never once measured what is in the water that is doing the scaling. A descale calendar is what you buy when you do not have a number. It works, and it is not free: half an hour a month is thirty hours over five years, on every machine, and not one of those hours tells me what is in the water.
So what do you actually do about it?
Find out first, then decide whether to treat. Those are two decisions, they cost different amounts, and the first is the one that gets skipped.
Finding out has three routes. Ask your utility directly for hardness and alkalinity, since the absence of those lines from the annual report does not mean the lab has not measured them. Buy an aquarium-style GH and KH drop test, which will give you both numbers for the price of a bag of coffee and is the fastest honest answer available to a home brewer. Or send a sample to a water lab, which is what a cafe planning a build should do, because the equipment decision downstream of that number is a four-figure one.
Once you have the numbers, here is the target set the 2009 SCAA standard publishes, and it remains the most widely used reference in the trade: total dissolved solids 150 mg/L, acceptable 75 to 250. Total alkalinity 40 mg/L. Calcium hardness 68 mg/L, acceptable 17 to 85. pH 7.0, acceptable 6.5 to 7.5. Sodium 10 mg/L. The standard also explains how to read its own table: "The Target is the most desirable point in the Acceptable Range, although falling within the range is considered meeting the standard."
Two things in that table deserve a second look. The first is total alkalinity, where the published acceptable range reads "At or near 40 mg/L." That is a target restated as a range, not a range, and it is the one parameter the document declines to bracket. Rao's stated preference of 30 to 40 ppm for light roasts is, in effect, somebody filling in the bracket the standard left open. The second is sodium at 10 mg/L, which is the figure that makes softened water a problem. A conventional softener does not remove hardness so much as trade it, swapping calcium and magnesium for sodium, and the standard's sodium ceiling is low enough that a softener can carry you past it while your hardness number looks like it finally behaved.
It is also worth being exact about what that document is and is not. It is the SCAA standard, published by the Specialty Coffee Association of America and revised on 21 November 2009, before the merger that produced today's SCA. The SCA's current published-standards listing names nine published standards and four more in development, SCA-110, SCA-131, SCA-315 and SCA-352, and none of them is a water standard. Its certification programs cover brewers, grinders, espresso machines and training venues. So a mineral packet can be formulated to hit the 2009 SCAA targets, and that is a real and useful thing to formulate toward. Nothing is approved against them, because the standard describes no approval mechanism.
Then the treatment decision itself. For a home brewer, the cheap and reliable move is not a mineral kit, it is knowing your two numbers and choosing a bottled or filtered water that lands near the targets when your tap does not. For a cafe, the decision is bigger and it is made at the plumbing, which is why we wrote separately about the best water for espresso machines and what it does for flavor and longevity. An espresso machine is a boiler holding heated water under pressure, and it is far less forgiving of the wrong water than a brewer you can descale on a Monday morning.
And here is the part I have to say plainly about us. I will sit in a church basement and tell a pastor that water is the thing he is going to address last and should address first, and if he asks me what our own alkalinity is, I do not have a number for him. ECUA does not publish one and I have never paid a lab for one. On the variable I am most confident about, I am working from a descale schedule rather than a measurement, which is exactly the trade I just told you not to make.
The coffee is the ingredient you chose, and the water is the one that showed up. Every test you run on the coffee, every grind adjustment and every ratio tweak, is being read through a solvent you have never measured, and it will keep quietly setting the ceiling on all of them until you find out what is in it.
Planning a cafe or a church coffee bar? Put water on the build budget as its own line item, next to the espresso machine and the grinder, rather than discovering it after the plumbing is in and the machine is scaling. That is one of the first things we cost out in a coffee shop consult, because it is the cheapest decision to make early and one of the most expensive to make late.
