Every recipe says something like "bulk ferment 4–5 hours." That number was true in the kitchen of the person who wrote it, on the day they wrote it, with their starter.
Fermentation speed roughly doubles for every 17°F / 9°C. A dough that finishes bulk in 4 hours at 78°F needs closer to 8 at 68°F. Add in how vigorous your starter is, how much of it you used, what flour you're on, and whether your kitchen drifts overnight — and the honest version of "4–5 hours" is "somewhere between 3 and 9."
So the clock is a prompt to go and look. The dough is the instruction. Learning to read it is the difference between following recipes and baking bread.
1 · Volume. The primary signal, and the only one that's close to measurable. How much has it grown from where it started? Not "has it doubled" — how much. More on the target in § 03.
2 · The edges. Look where the dough meets the side of the container. Early in bulk it slumps away and sits flat. As it ferments it climbs and the edge becomes domed — a convex curve rising to meet the wall. When it goes past its best, that dome sinks back and the edge flattens or dips.
3 · The jiggle. Shake the container gently. Under-fermented dough is dense and barely moves. Ready dough wobbles as one soft mass, like set jelly. Over-fermented dough sloshes — the movement is loose and liquid rather than springy.
4 · The surface. Domed, slightly glossy, with some bubbles visible on top and through the sides. A flat, matte, bubble-free surface means it needs more time. A surface covered in large broken bubbles that are collapsing as you watch means you're late.
This is where most advice goes wrong by giving one number. The right target depends entirely on what happens after bulk:
| What follows bulk | Stop bulk at | Why |
|---|---|---|
| Overnight cold retard | 30–50% | Fermentation keeps running for hours as the dough cools |
| Shape and bake same day | 60–75% | Only the short proof left to finish it |
| High whole grain | Lower end | Ferments faster; the gluten degrades sooner |
| Enriched dough | Higher end | Fat and sugar slow the yeast down |
"Doubled" — the number most recipes give — is almost always too far for a dough heading into the fridge. That single mismatch is behind a large share of flat, over-proofed loaves.
Eyeballing "50% risen" in a wide bowl is genuinely hard, because a bowl's cross-section changes with height — a 50% volume increase might only look like a couple of centimetres.
The fix costs nothing. Right after mixing, take a small piece of dough — 30–40 g — and press it into a straight-sided container: a shot glass, a small jar, a narrow tumbler. Mark the level with a rubber band. That sample sits next to the main bowl, at the same temperature, fermenting at the same rate.
Because the sides are straight, height is volume. When the aliquot has risen 50%, so has your dough — and you can read it off precisely instead of squinting into a bowl.
| Mistake | Why it misleads | Instead |
|---|---|---|
| Going by the clock | The recipe's kitchen isn't yours | Volume and edge shape |
| Judging by bubbles alone | A slack over-fermented dough is very bubbly | Bubbles plus dome plus jiggle |
| Using a wide bowl | Big volume change, small visible change | Straight-sided container, or an aliquot |
| Waiting for doubling before a retard | The fridge adds hours of fermentation | Stop at 30–50% |
| The poke test during bulk | It's a final proof test, not a bulk one | Save it for after shaping |
Notes from the counter
If there's one page on this site I'd want a beginner to read twice, it's this one. Shaping is a motor skill you'll pick up in twenty loaves. Scoring takes ten. This is the one that takes a season, because you're learning to read a continuous process from a handful of ambiguous signals, and you only get one data point per bake.
Which is the actual argument for the Calibration Loaf, and for baking the same thing repeatedly early on. Every variable you hold still is one less thing muddying the signal you're trying to learn.
The aliquot jar feels like cheating and isn't. Professional bakeries do the equivalent with temperature probes and rise charts. Give yourself an instrument rather than relying on judgement you haven't built yet — the judgement comes anyway, faster, once you can see what 50% actually looks like.