Ask which clay holds heat better and you will get an answer with real physics behind it and no numbers attached. Earthenware retains, porcelain lets go. Both halves of that are broadly true and nearly useless in a kitchen, because the two mugs you are actually choosing between are not the same weight, not the same wall thickness, not the same width across the top and not the same glaze, and every one of those differences moves the temperature more than the clay does.
Between two mugs of the same volume and the same wall thickness, earthenware holds coffee marginally longer than porcelain, and the gap is small enough that you would need a thermometer to find it. Weight, wall thickness, the width of the opening and whether you warmed the mug first all matter more. Put the mug on a kitchen scale and look at the mouth. A heavy warmed earthenware mug beats a thin cold porcelain one by a wide margin, a heavy warmed porcelain mug beats a thin cold earthenware one by about the same margin, and if what you want is hot coffee forty minutes from now, the answer is not a mug at all.
The reason the material question gets asked so often is that material is the thing written on the label. Weight is not, mouth diameter is not, and wall thickness is not. So people reach for the property they can read instead of the properties that decide the outcome, which is a reasonable thing to do and gives the wrong answer.
There are four questions here, not one
"Keeps coffee hot" bundles together four different pieces of behaviour, and a mug can be good at one and bad at another.
- How much heat the mug steals the moment you pour. This is decided almost entirely by how much the mug weighs.
- How fast the drink loses heat after that. This is decided mostly by the open top, not by the wall.
- How hot the outside gets in your hand. This is a comfort question and it runs opposite to the second one.
- How hot you actually want it. This is the one that turns the whole question around, and it comes last for a reason.
Separate them and the material argument shrinks to something small and honest.
What the two words actually name
Earthenware and porcelain are firing temperatures, and everything else follows from the porosity that results.
Earthenware is fired low, broadly in the region of 1,000 to 1,150°C. The clay body never fully vitrifies, so it stays porous, absorbs water, and depends on a glaze to hold liquid at all. It is also weaker, which is why earthenware is nearly always made thick.
Porcelain is fired high, broadly 1,250 to 1,400°C. The body vitrifies, absorption falls to almost nothing, and the fired material is strong enough to be made thin and still survive a household. Thinness is the whole commercial point of porcelain. Stoneware sits between the two and is closer to porcelain on the properties that matter here.
Two things follow that are worth more than the definitions. None of these words is regulated for tableware the way a hallmark is regulated for silver, so a mug sold as stoneware is often a glazed earthenware body under a heavier glaze. And the honest test needs no vocabulary at all. Hold the mug. Thick and heavy for its size means it behaves like earthenware whatever the box says, and thin and light and ringing when tapped means it behaves like porcelain.
The arithmetic of the first sip
Fired clay stores roughly a fifth to a quarter as much heat, gram for gram and degree for degree, as water does. That single ratio decides the first sip, and it is why several of the espresso and milk-drink recipes in Anette Moldvaer's Coffee Obsession open by telling you to warm the cup before anything else happens.
Run it. Take 300 g of coffee at 85°C, a kitchen at 20°C, and clay at a quarter of water's heat capacity.
A 350 g thick mug behaves, thermally, like adding about 90 g of cold water to your coffee. The mixture settles at roughly 70°C. You have lost about fifteen degrees before the cup reaches your mouth.
A 180 g thin mug behaves like adding about 45 g. The mixture settles nearer 76°C, so the loss is closer to eight or nine degrees.
That is arithmetic from published material values, not a measurement we have taken, and a real mug will land somewhere either side of it depending on its glaze and how the wall is profiled. The direction is not in doubt, though, and neither is the size. Fifteen degrees is not a subtlety. It is most of the difference between a cup you want and a cup you reheat.
Now rinse either mug with hot water for a few seconds and pour it away. Most of that loss disappears, because you have already put the heat into the ceramic that the coffee was about to donate. Four seconds of hot water buys more than most equipment upgrades, and it buys the most in exactly the case where the material argument says you should be safest, the heavy mug.
So on the first sip, the light mug wins. Ten minutes later the heavy one is ahead, because it is now a warm object sitting against the drink instead of a cold one pulling from it. Which of those you care about depends entirely on how fast you drink, and nobody selling mugs will ask you that.
The heat leaves through the top
Here is the part that makes the earthenware-versus-porcelain question smaller than it looks.
Evaporation is the dominant loss path for an open cup of hot coffee, and it is not close. Turning one gram of water into vapour takes roughly 2,300 joules. Cooling one gram of liquid water by one degree takes about 4.2 joules. So evaporating a single gram off the surface removes about as much heat as cooling 550 grams of coffee by a full degree. The steam you can see coming off the cup is the loss, happening in front of you.
Convection and radiation from that same free surface add to it. Both scale with area, and area scales with the square of the diameter. A 9 cm mouth presents about one and a half times the surface of a 7.3 cm mouth. Two mugs holding the same volume, one wide and shallow and one narrow and tall, are losing heat through the top at rates that differ by about fifty percent, and the clay has said nothing yet. Moldvaer's observation that a wide bowl cools fast is this fact, stated from the other end.
The wall is the second path, and this is where porosity finally does something measurable. The pores in an earthenware body are full of air, and air is a poor conductor, so a porous body conducts slightly less heat than a vitrified one of the same thickness. Earthenware also tends to be thicker, because it has to be. Both effects push the same way. Both are small next to the open top.
Which produces a conclusion no ceramics shop wants to reach: a saucer laid over the mouth of a thin porcelain mug will outperform the finest heavy earthenware mug sitting open. We sell ceramics and we would rather you knew that than bought the wrong thing from us.
The porous clay paradox
The same porosity credited with holding heat is used, elsewhere in the world, to throw heat away.
A Spanish botijo and an Indian matka are unglazed earthenware water vessels that keep their contents cool in hot weather. Water seeps slowly through the porous wall and evaporates on the outside, carrying the latent heat away with it. Identical material, opposite job. The variable that decides which behaviour you get is the glaze, not the clay.
That gives a clean rule for buying. A fully glazed earthenware mug is a sealed vessel with a mildly insulating body, and it behaves the way the marketing says. A partly glazed or unglazed one is a slow evaporator that also drinks your coffee. Kenneth Davids describes where that absorbency ends up. Over open fires in parts of Latin America and Indonesia, unglazed clay bowls are still used to roast, and they go glossy with age, because the oils driven off the beans settle into the surface and stay. In a roasting bowl that is character and nobody minds. In a drinking vessel it is last week's coffee, held in the body of the object, coming back at you.
Porosity carries a piece of history with it. Davids also relays a guess of Ian Bersten's about where roasting proper started. The usual answer is Yemen. Bersten's case for Damascus rests on one difference in equipment: Syrian workshops could work metal, and a metal pan goes hotter than a clay bowl. Davids calls it speculation and so do we. It is a clean illustration of what low-fired clay is, though: a material that will not take you above a certain temperature and will not hold you there.
Unglazed clay is not a mistake and the traditions using it are not confused. The Ethiopian jebena is earthenware, and café de olla is brewed in a clay pot with the clay claimed as an ingredient. Those are brewing vessels rather than drinking ones, and we have no measurement that supports or refutes the flavour claim.
The temperature you actually want
This is the part that turns the question over.
Jessica Easto sets out the Specialty Coffee Association's cupping protocol in How to Taste Coffee, and the temperatures in it are specific. Tasting does not begin until the sample has fallen to about 71°C. Acidity, body and balance are judged between roughly 71 and 60°C. Sweetness, uniformity and cleanliness are held back until the coffee approaches about 38°C. Assessment stops at about 21°C. The protocol exists because flavour is a moving function of temperature, and the people who taste coffee professionally do not evaluate anything at the temperature most people drink at.
Easto goes to the physiology underneath it. Sugar registers most strongly on a tongue held somewhere between about 22 and 33°C, a figure she takes from Talavera and colleagues, whose 2006 paper in Cellular and Molecular Life Sciences reviews how temperature moves the response of taste receptors. That is a review of mechanism rather than a trial, and the practical implication is modest and firm: very hot liquid dulls your ability to taste it.
Put that against the arithmetic above. Coffee poured at 85°C and drunk immediately is being drunk well above the range in which you can detect much of what is in it. A mug that holds it at 80°C for longer is holding it in the part of the curve where you are least sensitive. So the real reason to want mass in a mug is not to keep coffee hot. It is to stop the last third of the cup falling out of the range you can taste, which is a floor problem, and the material argument has been sold to everyone as a ceiling problem.
That reframing changes the recommendation for slow drinkers and leaves it alone for fast ones. It is also the strongest argument available for a smaller mug, which is a separate piece of work: the drinks got worse as the vessels got bigger.
What a mug will not do
No mug, in either material, will keep filter coffee at drinking temperature for an hour. The open top defeats every clay ever fired, and anything sold on that promise is either lidded, double-walled or overstating itself.
Neither material will change the flavour of the coffee through a sound glaze. A glaze is a glass. The body behind it never touches the drink, so claims that a particular clay imparts character to a fully glazed mug are claims about something the coffee cannot reach.
Earthenware will not go through a dishwasher indefinitely without consequence. The glaze and the body contract at different rates, and repeated heating and cooling is what turns that mismatch into the fine network of cracks called crazing. Some buyers want that network and some send the mug back. Either way it is a fit between glaze and body, not a manufacturing failure, and low-fired bodies show it sooner.
And weight will not rescue a cold start. A heavy mug you did not warm is the worst option on this page, worse than the thin one you also did not warm, which is a genuinely counter-intuitive result and the single most useful thing in this article.
The rim, which is not a heat question at all
Your mouth meets the mug before it meets the coffee, and the shape of that edge changes what the coffee seems to be.
Come off a thin edge and the drink arrives as a narrow band on a small patch of tongue, which is read as brighter and cleaner. Come off a thick or rolled edge and it lands wider and slower, which is read as rounder and softer, with less apparent acidity. Neither is better. They suit different coffees, and a light roast in a heavy rolled rim is being made to sound duller than it is.
This matters to the material choice because porcelain can be fired thin and stay strong and earthenware cannot. Choosing earthenware is largely choosing a thick rim, whether or not you intended to. Interior colour is a separate variable with its own consequences, covered in why a white interior is a reference field and a dark one is a decision to stop looking.
Four buyers, four answers
You drink 300 ml of filter coffee slowly, over twenty minutes. Heavy, fully glazed, narrow mouth, warmed before you pour. Earthenware or stoneware. The mass is working for you here and the narrow mouth is doing more than the clay.
You drink fast and you taste attentively. Thin porcelain, unwarmed, and let it fall. You will be inside the range Easto describes within two or three minutes instead of eight, and you will taste more of what you paid for.
You drink espresso or milk drinks. Warm the cup, without exception. The drink is small and the mug outweighs it several times over, which makes this the case where preheating stops being an improvement and starts being the difference between the drink and a failure. Two vessels, two jobs covers why one cup cannot cover both ends of the range.
You want it hot at a desk for forty minutes. Neither. Buy a lidded double-walled vessel and accept that you are optimising for something other than flavour, because you are, and there is nothing wrong with that.
The station
Five things, and the item that would do the most for the problem in the title is one we do not sell.
- A mug you have weighed. Between 300 and 400 g for a slow drinker, under 250 g for a fast one. This is ours, and the reason to buy it from anyone is that you like looking at it, not that its clay outperforms another clay.
- A second, smaller vessel for espresso and short milk drinks, because volume decides which drink you have made.
- A saucer or a lid that actually fits the mug. Most of the heat leaves through the top and almost nobody addresses the top.
- A kettle you can pour from twice, once to warm the mug and once to brew.
- A wooden or cork coaster. An unglazed foot ring will mark a finished surface over years of daily use, and a heavy mug set down hard on stone chips at the foot before it chips anywhere else.
Two we do not sell and recommend anyway: a vacuum flask, warmed before filling, for anyone who brews more than they drink at once, and a cheap probe thermometer, which will settle this argument in your own kitchen in one morning for less than the cost of a mug. As of August 2026 a basic instant-read probe is around $15, which is the cheapest piece of evidence in coffee.
What this analysis cannot see
We have not measured any of it. The heat capacities, conductivities and firing ranges used above are published values for classes of material, not measurements of any mug, and the property that most decides how a specific mug behaves is the one nobody publishes: the glaze thickness and the way the wall is profiled from rim to base. Two mugs sold as the same material and the same volume can differ more from each other than the categories differ on average.
There is a second gap and it is larger. Nobody has run the experiment that would settle the taste half of this: the same coffee, in two mugs of different material and matched weight and geometry, tasted blind at matched temperatures. Until that exists, the claim that material affects flavour through a sound glaze has no evidence for it, and the claim that it does not has only mechanism on its side. Mechanism is decent footing. It is not the same thing as having looked.
The honest close is that this is a smaller decision than it feels like while you are making it. A mug is an object you will pick up several thousand times, and the case for choosing carefully has almost nothing to do with thermodynamics and almost everything to do with the fact that you will be holding it every morning for a decade. That is a real reason. It is just not a heat-retention reason, and we would rather say so than sell you the physics.