Building a starter from scratch in 7–14 days

Building a sourdough starter from flour and water is one of the most quietly satisfying things in baking. You start with two ingredients that don't do anything interesting on their own, and a couple of weeks later you have a self-sustaining microbial culture that can leaven bread for the rest of your life. No commercial yeast, no special equipment: just flour, water, time, and a small amount of patience for the first ten days.

This article walks through what's actually happening at each stage, why most people get confused around day 3, and how to know when your starter is genuinely ready to bake with. The mechanics are simple. The timing is what people misjudge.

What you're trying to grow

A mature sourdough starter is a balanced ecosystem of two kinds of microorganism: wild yeasts (mostly *Saccharomyces* and *Candida* species) and lactic acid bacteria (mostly *Lactobacillus*). Both are already present on the flour itself, on your hands, and floating around in the air of any kitchen where bread, fruit, or fermented food has ever been. You don't need to buy them or seed them; they show up on their own.

What you have to do is create the conditions in which those microbes can multiply, while the unhelpful microbes (mold, undesirable bacteria) lose. The rules of the game are: keep it slightly acidic, keep it consistently fed, keep it warm enough but not hot.

What you need

• Flour. Whole-grain rye works fastest. Whole-grain wheat is second-best. Plain white flour works but takes longer because it has fewer microbes and fewer minerals on the bran. A 50/50 blend of whole rye and white flour is a reliable starting mix.
• Water. Tap water is fine in most places. If your tap water is heavily chlorinated, leave a glass of it uncovered overnight before using; the chlorine will dissipate. Otherwise no special treatment needed.
• A clear glass jar. A 500 ml / 16 oz jar with a loose-fitting lid. Glass so you can see what's happening; loose lid so the gas can escape. A rubber band around the jar to mark the starter's level is the cheapest, most useful tool you'll buy.
• A digital scale. Reading in 1-gram increments. Volume measurements (cups, spoons) are too imprecise to make this work reliably.

The basic schedule

Day 1: combine 50 g of flour and 50 g of water in your jar. Stir until smooth. Mark the level with a rubber band. Cover loosely. Leave at room temperature (20–24 °C / 68–75 °F is ideal).

Days 2 onward: once a day, discard about half of what's in the jar. Add fresh flour and water in equal weights (50 g of each) and stir to combine. Mark the new level. Cover. Wait.

That's the whole protocol. Discard, feed, wait, repeat: once a day for about a week, then twice a day for the second week if needed. The technique doesn't change. What changes is what your starter does in response.

What you'll see, day by day

Days 1–2: nothing

The mixture sits there. Maybe a few small bubbles by day 2. This is normal. You're at the very beginning of microbe colonisation.

Days 3–4: the false start

This is where most beginners get excited and most get confused. Around day 3 or 4, the mixture often rises dramatically, sometimes doubling or tripling. Lots of bubbles. It smells aggressive, sometimes sour, sometimes acrid. People assume the starter is ready and try to bake with it.

It is not ready. What's happening on day 3–4 is not a sourdough culture. It's a population of bacteria called *Leuconostoc* and similar species that thrive in fresh flour-and-water at neutral pH. They produce a lot of gas quickly and then die off as the environment becomes acidic. The yeast and *Lactobacillus* that you actually want haven't taken hold yet. Bread baked with day-3 starter doesn't rise; the microbes responsible for the bubbles are already gone.

Days 5–7: the dead zone

After the day-3 burst dies down, the starter often goes quiet. Few bubbles, little rise, sometimes a strange smell. This is the most discouraging stage. Don't give up. The *Leuconostoc* are dying off and the wild yeast and *Lactobacillus* are slowly establishing. Keep feeding once a day on schedule.

Days 7–10: the real activity

Sometime in this window, the starter starts behaving differently. After a feeding, it rises predictably (1.5–2× volume within 6–12 hours). The smell becomes pleasant: yogurt-like, slightly bready, mildly sour. The bubbles are evenly distributed, not just at the surface.

At this point you have something that's *probably* a working starter. Confirm by switching to twice-daily feedings and watching the timing.

Days 10–14: confirming readiness

Switch to feeding every 12 hours. Use a 1:1:1 ratio (equal weights of starter, flour, water). The starter is ready to bake with when, after a feeding, it reliably doubles in volume within 6–8 hours at room temperature, falls back, and the cycle repeats predictably the next feeding.

If your starter doubles in 6 hours one day and barely moves the next, it's not stable yet; keep feeding twice daily for another 3–5 days. Stability is more important than peak height.

How to know when to actually bake

The float test is the most-cited check, and it works: drop a small spoonful of fed, peaked starter into a glass of water. If it floats, it has enough gas trapped to leaven bread. If it sinks, it's either too young, past peak, or not active enough.

But the float test alone isn't enough; a starter can pass the float test on day 5 (during the false start) and still produce a brick. The better signal is predictability over a week. If your starter has hit peak within roughly the same time window for 5–7 consecutive feedings, it's ready.

Common mistakes that kill the timeline

• Skipping feedings. A starter that goes 36+ hours without food during the first two weeks gets badly out of balance. Set a daily reminder.
• Too cold a kitchen. Below 18 °C / 64 °F, microbe growth slows so much that you can wait three weeks instead of two. Find a warmer spot: top of the fridge, near a heating vent (but not on it), or in a turned-off oven with the light on.
• Switching flours mid-build. Pick one flour or one blend and stick with it for the first two weeks. The microbe population partly adapts to whatever you're feeding it; switching flours mid-build resets some of the progress.
• Discarding too little. If you don't discard enough each feeding, the new flour can't keep the acidity in the right zone. The 1:1:1 ratio (50% discarded) is the conservative starting point. After day 7, ratios like 1:5:5 work better (covered in the next article).

What 'discarding' actually means

The discarded starter doesn't have to go in the bin. After day 7 or so, it's perfectly edible. It's just unfed sourdough. Use it in pancakes, crackers, or pizza dough. During the first 7 days, when the culture is still establishing, discard it: the microbe balance isn't reliable yet for eating or baking.

After day 14

A 14-day-old starter is functional but young. It will keep getting better (more flavor, more reliable timing, more resilience to neglect) for months. A starter that's a year old is genuinely a different organism from one that's two weeks old, even if it looks the same. Use yours, feed it, and it'll keep improving long after you've stopped paying close attention.