You waited eighteen hours. You cut into it while it was still ticking, and the inside is a pale, sticky, slightly translucent paste that sticks to the knife and rolls into a ball when you press it. Sourdough dense and gummy in the middle is the single most-searched sourdough failure, and the frustrating part is that two completely opposite mistakes produce a crumb that looks superficially similar.
The good news is that they do not look similar for long. Under-proofed and over-proofed loaves fail in opposite directions, and once you know which direction you are looking at, the fix is obvious and usually only takes one variable.
Before the crumb, look at the loaf. Its silhouette tells you more than the inside does.
A third possibility sits behind both of those, and it accounts for a genuinely large share of “gummy” complaints: the loaf was fine and you cut it too early. We will get to that, because nobody wants to hear it and everybody needs to.
The short answer
Under-proofed sourdough is tall, tight and heavy. It springs violently in the oven, tears open along a side or a shoulder rather than along the score, and shows a dense, pasty band near the base with a big irregular tunnel under the top crust. The crumb feels wet and clay-like and does not spring back when pressed.
Over-proofed sourdough is flat, wide and pale. The score spreads open dully instead of lifting into an ear, there is little oven spring, and the crumb is fragile and shreddy with small uneven holes or a collapsed area. It smells sharply sour.
If it is tall and gummy, ferment longer. If it is flat and gummy, ferment less, or keep the same time and drop the dough temperature. And if the loaf looked good and the crumb was sticky, wait four hours next time before you cut it.
The two failures point in opposite directions
Fermentation builds gas and, past a point, dismantles the structure holding it. Those two processes overlap, and everything about proofing is about catching the loaf in the window where you have enough of the first and not too much of the second.
Under-proofed dough has an intact, strong gluten network with too little gas in it. So it holds together beautifully, resists spreading, and stands tall. Then it hits 465°F (240°C) of oven heat, the trapped gas expands fast, the surface has no give, and the loaf bursts out wherever it is weakest. That is the aggressive tear people mistake for good oven spring.
Over-proofed dough has plenty of gas and a gluten network that acid and protease activity have partly taken apart. It cannot hold what it has. It slumps on the peel, spreads instead of rising, and the crumb structure that survives is thin-walled and fragile.
Both give you gumminess, but for different reasons. In the under case, part of the dough never developed cell structure and stayed as a compressed paste. In the over case, the walls between the cells collapsed and welded together during the bake.
I would push back gently on the way this is usually taught. Most guides describe over-proofing as the common beginner error. In practice I think under-proofing is far more common, particularly in kitchens under 70°F (21°C), because people follow clock times from recipes written in warmer conditions and are also nervous about “going too far”.
Reading the crumb properly
Cut the loaf across the middle, not near the end. Look at the whole cross-section rather than hunting for one big hole.
What under-proofed looks like
A tall, domed loaf, sometimes with a dramatic split up one side where the crust gave way. Inside, the top third has an open, irregular structure, often with one large cavern immediately under the crust where the gas rushed upward and collected. Below that the crumb tightens progressively, and the bottom quarter to third is dense, close-textured and noticeably darker or more translucent.
Press it. Under-proofed crumb compresses and stays compressed, like plasticine. It tastes bland and slightly sweet, because the fermentation never produced much acid or the aromatic compounds that come with it.
The score usually opened, and opened hard, but the ear may be jagged rather than clean.
What over-proofed looks like
Wide, low, spread past its own footprint. The score line is a dull seam rather than a lifted flap. The crust is often paler than you expect, because the yeast consumed the residual sugars that would otherwise have browned in the Maillard reaction, which is one of the more reliable visual tells.
Inside, the holes are small and irregular, sometimes with a collapsed band just below the top crust where the structure fell during or just after baking. The crumb shreds and tears rather than slicing cleanly, and the whole thing may feel damp without the clay-like density of the under-proofed loaf.
It smells and tastes sharply sour, sometimes with a vinegary edge.
The full diagnostic table
| What you see | Under-proofed | Over-proofed | Other cause |
|---|---|---|---|
| Loaf shape | Tall, domed, holds its footprint | Flat, wide, spread | Weak shaping tension |
| Oven spring | Violent, tears at a shoulder or side | Almost none | Cold dough or no steam |
| Score behaviour | Opens hard, jagged ear | Spreads dully, no ear | Wrong angle or blunt blade |
| Crust colour | Deeply browned, sometimes very dark | Pale, resists browning | Oven temperature |
| Crumb near base | Dense, pasty, darker band | Even but fragile | Cut while warm |
| Crumb near top | One large cavern under crust | Collapsed band | Trapped shaping bubble |
| Hole size | Small and tight, uneven | Small, thin-walled, ragged | Over-degassed at shaping |
| Texture pressed | Compresses and stays flat | Shreds and tears | Insufficient bake time |
| Flavour | Mild, slightly sweet, doughy | Sharply sour, vinegary | Long cold retard, not a fault |
| Dough at shaping | Firm, snappy, holds a boule | Slack, sticky, tears | Hydration beyond flour capacity |
Use the shape and the score first. They are the two most reliable signals, and they are visible before you cut anything.
The cause everyone forgets: you cut it hot
I would put money on this being a third of all gummy-crumb reports.
A loaf leaving a 450°F (230°C) oven has an interior around 208°F (98°C) and is still actively cooking. Starch that gelatinised during the bake begins to retrograde, which is the process of amylose and amylopectin chains recrystallising into a firmer structure, and moisture continues migrating from the wetter centre outward. Neither finishes quickly.
Cut into it at twenty minutes and you get exactly what you would expect: sticky, pasty, a smeared knife. That crumb would have set into something perfectly good if left alone.
How long is long enough? For a standard 900g (2lb) loaf, two hours minimum on a wire rack, four if you can bear it. For high hydration above 80%, closer to four. For anything with a significant rye percentage, the traditional advice is overnight, and up to 24 hours for a proper rye sourdough, because rye’s pentosans hold water very differently from wheat gluten.
The rack matters. Sitting a hot loaf on a solid surface traps steam underneath and gives you a soggy base, which then gets diagnosed as under-fermentation.
Set a timer and walk away. If you struggle with this (most of us do), a plain countdown timer running in a browser tab is a surprisingly effective commitment device.
Internal temperature: a floor, not a verdict
The standard target for lean sourdough is 205-210°F (96-99°C). Below about 200°F (93°C) the starch gelatinisation is genuinely incomplete and the loaf will be gummy no matter what else you did.
Here is the part that gets skipped. A badly under-fermented loaf will reach 208°F (98°C) perfectly happily and still slice like putty, because gumminess in that case is not about heat penetration, it is about gas cell structure that never formed. Temperature confirms the loaf is cooked. It says nothing about whether it is good.
Some bakers go further and pull enriched or very wet loaves at 208-210°F (98-99°C) while taking lean loaves to 205°F (96°C), and others insist everything should hit 210°F (99°C). The differences here are small and the disagreement is real. What matters more is total bake time: a loaf that hits temperature in 28 minutes because your oven runs hot is under-baked in every way that counts except the thermometer, so give it the full time and check colour as well.
Probe into the side of the loaf towards the centre rather than through the base, and take the reading before the loaf has been sitting out for ten minutes.
The three structural causes that sit underneath
Fermentation timing gets the blame for problems that were actually baked in earlier.
Hydration beyond what your flour can carry
An 85% hydration dough made with 11.5% protein all-purpose flour is not a challenge, it is a category error. The flour cannot form a gluten network strong enough to hold that much free water, so the dough behaves like batter, spreads in the banneton, and bakes into something wet and dense regardless of your timing.
If your dough is soup at shaping, drop hydration by five points and bake the same recipe again. 75% instead of 80%. It is not a retreat. The relationship between hydration, flour protein and handling is covered properly in hydration explained in grams, including the levain arithmetic that makes most people’s “75%” doughs wetter than they think.
Gluten that was never developed
Folds are not decorative. A dough that got one half-hearted set of stretch and folds at 30 minutes and then sat for five hours has an underdeveloped network, and no amount of correct timing rescues it. Three or four sets in the first half of bulk, spaced 30-45 minutes apart, is the standard for good reason. If you are unsure which technique suits your dough, the differences between stretch and fold, coil fold and slap and fold matter more at high hydration than at 68%.
The subtler version of this problem is folding too late. Folds in the last third of bulk degas dough that has already done its work, and you will see it as a tighter, denser crumb even though your total bulk time was correct.
Too much whole grain, too soon
Whole wheat and rye do three things at once: bran particles physically interrupt the gluten network, the extra enzyme activity speeds fermentation and softens the dough, and the bran absorbs water slowly so the dough tightens then loosens over time.
Jumping from 10% whole grain to 50% without changing anything else usually produces a denser loaf that also over-ferments on your old schedule. Move in 10% increments, add a little more water each time (roughly 2-3% more for each 10% whole grain, though this varies by mill), and shorten bulk by around 15-20%.
Walking one failed bake back through the timeline
Say Marta bakes most Saturdays. Her recipe, in baker’s percentages: 500g strong white flour (100%), 375g water (75%), 100g of 100% hydration levain (20%), 10g salt (2%). Total dough 985g.
Her timeline last Saturday:
- 8:00am, levain built at 1:5:5, kitchen 70°F (21°C)
- 2:00pm, levain used (six hours, looked domed)
- 2:15pm, mix, dough temperature measured at 73°F (23°C)
- Folds at 2:45, 3:15, 3:45
- 6:15pm, end of bulk, “four hours as the recipe said”
- 6:30pm, shaped, into banneton, into the fridge
- Sunday 8:00am, baked from cold, 20 minutes covered at 450°F (230°C), 22 minutes uncovered at 425°F (220°C)
The loaf came out tall, split along the left shoulder, and the bottom third was dense and pasty.
Work backwards. The recipe said four hours of bulk. Recipes that say four hours almost always assume 78°F (26°C) dough. Marta’s dough was at 73°F (23°C), five degrees cooler. Using the rule of thumb that fermentation rate roughly doubles per 15-18°F (8-10°C), five degrees Fahrenheit costs her something in the region of 20-25% of her fermentation rate, which means her four hours did roughly the work of three hours at the recipe’s assumed temperature.
Check that against the dough itself. Marta did not measure rise, but she remembers the dough felt firm and “snappy” at 6:15pm and held a tight ball when shaped. That is an under-fermented dough. A dough at the end of a proper bulk feels airy, slightly domed, jiggly when you shake the bowl, and shows bubbles at the edges where it meets the container.
The fix for next Saturday is one change: extend bulk to 5 hours 15 minutes at the same 73°F (23°C), or warm the dough to 78°F (26°C) and keep four hours. Not both. Change one thing so you learn something.
Marta could also add an aliquot jar, which is the single most useful upgrade available here. Take about 30g of dough at the start of bulk, press it into a small straight-sided jar, mark the level, and let it sit next to the main dough. When the sample has risen by your target percentage, bulk is done. For a 75% hydration white loaf with a cold retard to follow, 50-75% rise is a common target, though bakers argue about this constantly and some experienced people bulk to 100% and get excellent results with strong flour.
That argument is worth taking seriously rather than resolving. Target rise depends on flour strength, hydration, whether you are retarding, and how much structure your shaping adds. Pick a number, bake three loaves at it, and adjust from your own results.
If you would rather not do the temperature arithmetic every time, the Sourdough Proofing Scheduler takes your measured dough temperature and gives you a bulk window and a set of clock times for folds, shaping and the fridge. It runs in your browser, stores nothing, and needs no sign-up, which mostly means you can open it with wet hands and no faff. For planning an oven around other things you are baking that day, the Bread Bake Timeline Scheduler handles the whole sequence.
How to judge the end of bulk without guessing
Three methods, in the order I would trust them.
- Percentage rise, measured. Aliquot jar as described above, or a straight-sided container with the starting level marked. Target 50-75% for a standard white sourdough at 75% hydration going into a cold retard. Higher whole grain, aim lower. This is the only method that gives you a number you can repeat.
- Visual and tactile cues in the main dough. It should be domed at the edges rather than flat, show bubbles on the surface and along the sides of the container, feel airy and lighter than it looks, and jiggle as one mass when you shake the bowl. It should not be a soup, and it should not still feel tight and elastic.
- The poke test, used at final proof rather than bulk. Flour a finger, press about 1cm into the shaped dough, and watch. Springs straight back fully: needs longer. Springs back slowly and leaves a shallow dent: ready. Does not spring back at all and the dough sighs: over-proofed. The poke test is genuinely contested, and it should be, because it is heavily influenced by hydration, flour and how hard you press. Treat it as a confirming signal, not a decision.
Time is not on that list, and that is deliberate. Clock times are useful for planning your day and nearly useless for judging fermentation, because they cannot see your kitchen. Use them the way the proofing scheduler does: as a prediction of roughly when to start checking, not as the instruction itself.
What goes wrong when you try to fix it
Overcorrection is the classic. You had a dense loaf, so you extend bulk by two hours, and now you have a flat one. Adjust in increments of 30-45 minutes, not hours.
Changing three things at once is the other one. New flour, longer bulk and higher hydration in the same bake tells you nothing about which change helped.
Then there is the retard confusion. Fermentation does not stop in the fridge, it slows, and how much it slows depends on how quickly your dough actually gets cold. A 900g shaped loaf in a domestic fridge can take two to three hours to reach 39°F (4°C), and it ferments the whole way down. If you bulk to 75% rise and then retard for 18 hours, you may well be over-proofed by morning even though your bulk was textbook. Shorten bulk when you retard long, or shorten the retard.
One more, and it is the one that costs people months: assuming the starter is at fault. A starter that doubles reliably in six hours is fine. If yours does not, that is a separate problem with its own diagnostic order, but do not go rebuilding a healthy culture because a loaf came out dense.
Finally, salt. Leaving it out or halving it speeds fermentation noticeably and weakens the gluten, and people do forget it. 2% of flour weight is the standard, which is 10g for 500g of flour. If a bake goes strangely fast and slack for no reason you can identify, check whether the salt went in.
Rescuing the loaf you already baked
You cannot un-bake it, but you have options better than the bin.
Return the whole loaf to a 350°F (180°C) oven, uncut, for 10-15 minutes. This drives off some moisture and firms the crumb. Then cool it fully on a rack. This works best on loaves that were slightly under-baked rather than badly under-fermented.
Slice thin and toast. Toasting drives the same moisture out and the density becomes an asset rather than a fault. Gummy sourdough makes excellent croutons, breadcrumbs, panzanella, French toast and bread pudding, and a dense sour crumb holds up better in all of those than a light open one would.
What you should not do is wrap it while warm. That traps every bit of the moisture you are trying to lose, and by morning the crust will be leathery and the crumb worse.
Frequently asked questions
Is my sourdough gummy because it is under-proofed or over-proofed?
Look at the shape before the crumb. Under-proofed loaves are tall and tight with a violent tear along one side and a dense pasty band near the base. Over-proofed loaves are flat and wide, spread past the score line, and have a fragile crumb that shreds when you cut it. Tall and gummy means under. Flat and gummy means over. Crust colour helps too: over-fermented loaves resist browning.
Why does my sourdough have a dense gummy band at the bottom?
That band is usually under-fermented dough that had not developed enough gas structure before it hit the oven, so it compressed under the weight of the rising loaf above it. It can also come from cutting the loaf while still warm, or from a dough that was degassed by folding or shaping too late in bulk. Extend bulk by 30-45 minutes before you change anything else.
How long should sourdough cool before cutting?
At least two hours for a standard 900g loaf, and four is better. Starch retrogradation and moisture redistribution continue long after the loaf leaves the oven, so a warm interior reads as sticky and gummy even when the fermentation was perfect. High-hydration and rye loaves need longer, with rye traditionally rested overnight or up to 24 hours. Cool on a wire rack so steam does not collect under the base.
What internal temperature should sourdough be when baked?
Most bakers pull lean sourdough somewhere between 205°F and 210°F (96°C to 99°C). That confirms the starch has gelatinised and the loaf is cooked through. It does not confirm the loaf is good, because a badly under-fermented loaf will hit the same temperature and still slice gummy. Treat temperature as a floor rather than a verdict, and check crust colour and total bake time as well.
Can you fix a gummy sourdough loaf after baking?
You cannot re-ferment it, but you can improve it. Return the uncut loaf to a 350°F (180°C) oven for 10 to 15 minutes to drive off moisture, then cool it completely on a rack before slicing. Failing that, slice thin and toast, or turn it into croutons, breadcrumbs, French toast or a bread pudding, where density stops being a problem and becomes useful.
Why is my sourdough flat and spread out?
Most often over-fermentation, where the gluten network has been degraded by acid and enzyme activity until it can no longer hold gas. The tells are a dull score that spread rather than opened, a pale crust and a sharply sour smell. Weak shaping tension and hydration beyond what your flour can carry produce a similar shape, so check those next. A long cold retard on top of a full bulk is a common combination behind it.
Where to point your next bake
Look at the silhouette first, then the score, then the crumb. Tall and torn means give it longer. Flat and dull means pull it earlier or run the dough cooler. Good-looking loaf, sticky slices means you cut it hot.
Then change one variable, write down the dough temperature and the percentage rise, and bake it again. Three loaves with one variable moving will teach you more than thirty loaves of copying recipes, because your kitchen, your flour and your starter are not the ones the recipe was written for.
Most people fixing this find their answer inside two bakes. It is usually 45 more minutes of bulk, or four more hours of cooling.