How I Finally Nailed High-Hydration Sourdough After 30 Failures
Batch number 12 is the one I remember most vividly — not because it was good, but because it went directly from the proofing basket into the bin, still in the parchment. It had the structural integrity of warm hummus. My husband, who has been impressively patient through this entire ordeal, called it "artisan soup." He wasn't wrong.
I'd been chasing that open crumb — the kind you see in the photos where light practically shines through the holes — for the better part of a year. I was watching the videos, reading the posts, buying the books. My starter was healthy. My technique was improving. But hydration was the thing I kept getting catastrophically wrong.
So I started keeping a log. An obsessive, slightly unhinged baking diary with columns for flour weight, water weight, temperature, fold schedule, bake time, and a 1–10 crumb score alongside my notes. Thirty-one entries. Here's what actually changed things.
The Hydration Number Is A Lie (Sort Of)
When I first read that a "high hydration" sourdough sits around 75–80%, I treated that like an instruction. Start there. See what happens. What happened was soup-dough, repeatedly, because I hadn't accounted for the one variable nobody leads with: flour absorption.
Here's the thing. A 75% hydration dough made with a strong bread flour — something like a 13–14% protein content — behaves completely differently than 75% hydration dough made with an all-purpose flour sitting at 10–11% protein. The protein absorbs the water. More protein, more absorption, firmer dough at the same percentage. I was using a supermarket bread flour I'd grabbed because "bread flour" seemed right. It wasn't high enough protein to handle what I was throwing at it.
My fix: I sourced a high-protein flour — I landed on one at around 13.8% protein — and immediately, at the same 78% hydration I'd been using, my dough was workable. Not easy, but workable. That one substitution saved the entire project. If you're stuck in soup territory and you haven't checked your flour's protein content, check it. It's listed on the bag, usually under "Nutrition Facts" per 100g as grams of protein. 13g or above per 100g? You're in the right zone.
The Starter Percentage Finally Made Sense
I was also using too much starter. The standard range is 15–20% of your flour weight. I was at 25%, because more starter meant more rise, right? More fermentation activity? Faster results?
What it actually meant: my dough over-fermented before I could develop any real structure. The gluten was getting eaten alive before it had a chance to form the network I needed. The crumb was gummy and dense in all the wrong ways, not airy and open. I dropped to 17% and gave my dough more time at cooler temperatures to ferment slowly. The difference in structure was immediate and obvious.
The other thing this taught me: if you're converting recipes or scaling — which I was, constantly, because I wanted to test in smaller batches — baker's percentages are genuinely useful here. Everything is expressed as a percentage of total flour weight. So in a 500g flour recipe:
- 78% hydration = 390g water
- 17% starter = 85g starter
- 2% salt = 10g salt
Once I started using a kitchen calculator and just working from percentages instead of copying weights directly, scaling stopped introducing errors. I was baking 300g test loaves — much easier to sacrifice a 300g failure than a full kilo batch — and the math scaled cleanly.
Temperature Is A Conversion Problem Too
Nobody tells you that "room temperature" in a sourdough recipe is a variable, not a constant. The author who wrote "bulk ferment 4–5 hours" probably had a kitchen around 24–26°C. My kitchen in winter sits at 19°C. That's not close enough. At cooler temperatures, fermentation slows dramatically — you might need 7–8 hours to achieve what the recipe describes in 4.
I bought a probe thermometer and started measuring dough temperature instead of ambient temperature. The target most bakers aim for is a final dough temperature of around 24–26°C after mixing. If your water is cold and your flour is cold, your dough starts cold and ferments slowly. I started warming my water to compensate — and this is actually where a cooking converter becomes genuinely useful. Target dough temperature calculation is a real formula:
Desired water temperature = (Target dough temp × 3) − flour temp − ambient temp − friction factor
For hand-mixing, friction factor is roughly 0. So if I want 25°C dough, my flour is at 19°C, and my kitchen is 19°C: I need water at (75 − 19 − 19) = 37°C. That's warm but not hot. I started hitting that mark and my bulk ferment times stabilized. The recipe timelines started making sense for the first time.
The Ingredient Substitutions That Actually Worked
At some point around batch 18, I ran out of my good flour mid-experiment and had to make do with what I had. I ended up doing a blend: 80% bread flour and 20% whole wheat. My log notes say "desperate substitution, expected disaster."
It wasn't a disaster. It was the best crumb I'd produced to that point. The whole wheat flour added some extra bran and germ, which created nucleation sites for bubbles in the crumb — or at least that's my working theory. What I know for certain is it tasted better and the open structure improved. I've kept the 20% whole wheat addition ever since.
The caveat: whole wheat absorbs more water than white flour. If you're substituting in whole wheat, you generally need to increase hydration by about 5% to compensate. So my effective recipe shifted from 78% to 83% hydration to account for the absorption difference. This is the kind of adjustment a good baking converter will flag — when you're swapping flours, you're not just swapping flavour, you're changing how much water the mix can absorb.
A few other substitutions I tested and logged:
Honey instead of a small amount of sugar (0.5% of flour weight): Slight improvement to crust colour, minimal impact on crumb. Honey is hygroscopic, so it holds moisture — crust stayed crispier for longer after baking. Substitution ratio is 1:1 by weight, but reduce liquids by about 20% of the honey added (since honey is roughly 80% sugar, 20% water). Barely matters at this small scale, but worth knowing if you scale up.
Olive oil (2% of flour weight): Added to a couple of batches to see if it helped extensibility. It did — dough was easier to shape. Crumb was marginally tighter, which I didn't love. I dropped it.
Cold butter (lamination attempt): One of my more chaotic experiments. Croissant technique, sourdough dough. I won't say more except that batch 22 does not appear in the log's scoring column.
What Batch 31 Looked Like
The loaf that finally looked like the photos came together around 6am on a Saturday. 500g high-protein bread flour, 100g whole wheat, 83% hydration (498g water), 17% starter (102g), 2% salt (12g). Autolyse 45 minutes before adding starter and salt. Four stretch-and-folds at 30-minute intervals. Total bulk ferment 6.5 hours at 22°C ambient. Overnight retard in the fridge, baked in a Dutch oven at 250°C for 20 minutes covered, then 220°C for another 25 minutes uncovered.
I cut into it after two hours of cooling — excruciating — and the crumb had actual holes. Irregular, gorgeous, exactly what I'd been chasing. I took four photos before I tasted it. It tasted like sourdough bread is supposed to taste: complex, slightly tangy, with a crust that shattered when I pressed it.
The thing I want to say to anyone in the middle of their own string of soup-doughs: the failures weren't wasted. Every bad loaf generated a data point. The logging was annoying and felt excessive, but it's what let me isolate variables — flour protein, starter percentage, temperature, bulk time — and change one thing at a time. You can't improve what you don't measure.
And if you're scaling a recipe, substituting flours, or trying to reverse-engineer a hydration from a recipe that doesn't give you percentages — do the math properly. Baker's percentages exist because they work. A cooking unit converter with baker's percentage support is genuinely the second-most-useful tool in my bread kitchen. The first is the scale. But you already knew that.