Cultures · 8 min read · 18 September 2026

Kombucha shelf life: setting a best-before date you can defend

The date on the label is a claim you make about a product that is still alive when it leaves you. The question is not when kombucha goes off. It is how long it stays the drink you sold.

Most producers pick their first shelf life the way everyone does, by reading another brand's bottle. Then a wholesale buyer asks for the number on a listing form, an auditor asks what it is based on, and a pallet comes back gassy in week ten. This guide is about what carries on inside the bottle, the two numbers that decide the date, and how to get real evidence out of the samples you are already keeping.

What keeps happening after you fill

A raw, unpasteurised kombucha leaves the building with a live culture and whatever sugar is left in it. Five things carry on, all of them faster when it is warm.

Alcohol goes up. The yeast keeps working on the residual sugar, so the reading you take at filling is the lowest that product will ever give you.
Pressure goes up. The same fermentation makes CO2 inside a sealed package. Fizz turns into foam, foam turns into a gusher, and past a point the package itself is what fails.
The acidity keeps climbing. pH drifts down slowly, and long before anything is unsafe the drink tastes sharper than the one your customer bought last month.
It changes how it looks. Sediment settles out and a thin new pellicle can form in the bottle. Normal in a live product, and still the most photographed thing in this category.
The aroma fades. Fruit and tea notes flatten while the acid stays, so the balance moves even when nothing has gone wrong.

A filtered or pasteurised product behaves differently, and your own numbers will show it. Everything below assumes the live version, because that is the one where the date is hard to defend.

The two numbers that decide the date

Taste, clarity and fizz decide whether somebody buys the second bottle. Two numbers decide whether the date itself holds up.

The alcohol on the last day. Whatever line you sell under, the reading that counts is not the one from filling day. It is the one from the last day of the date you printed, on product stored the way the label says to store it.
The package on the last day. A bottle or a can has a pressure it is rated for, and a live product keeps making pressure. The date has to sit inside what your package, your fill level and your closure can take at the warmest temperature the product will really see.

The date is the shorter of those two, minus room to spare. Everything else is quality. It can cost you the listing, but it is not what makes the number defensible.

The date and the storage line are one claim

A best-before on its own says very little. The same batch that holds for four months cold can be out of specification in six weeks in a warm store room. The storage line on your label is not decoration next to the date. It is the condition the date is true under.

So decide what you are claiming before you test anything. Does the product go out refrigerated and stay refrigerated, through your van, the customer's back room and their shelf? Or are you claiming an ambient shelf life, which is a much harder claim for a live product and needs the numbers to carry it? If your route or your customers break the cold chain routinely, the honest date is the one that survives the broken version.

A shelf-life study you can actually run

You are probably already putting a bottle or two from every batch on a shelf. A retention sample nobody opens is an archive. Opened on a schedule and written down, it is a study.

Take one normal batch. Not your best one. A batch that represents what you actually sell, in the package you actually use.
Set aside enough bottles. Around a dozen per storage condition, so you can open one at every checkpoint and still have a spare when a reading looks wrong.
Store them two ways. The condition you intend to print, and one deliberately worse, for example room temperature. Treat the warm set as a stress test rather than a clock. A live culture does not scale neatly with temperature, so it tells you what a bad week does, not what six months do.
Fix the checkpoints before you start. Day 0, then 14, 30, 60, 90, and every 30 after that. Put them in the calendar on the day you fill, because a study that depends on remembering is a study that stops at day 30.
Measure the same things every time. pH on a meter you calibrated that day, alcohol by a method you can defend, carbonation or pressure, and two lines of tasting note with a photo of the bottle.
Keep going past the day you were hoping for. The most useful checkpoint is the first one that fails, because it is the only one that tells you where your margin actually is.
Then set the date short of it. If the product still holds at day 120 cold and crosses the line at 150, printing 120 prints the edge. Ninety leaves room for the warm van, the bottle at the back of the shelf, and the batch that ferments a little faster than this one did.

Run it again when something real changes: a new recipe version, more residual sugar, a different fruit addition, another bottle or closure, a new filler. And once a year on a normal batch anyway, because a process drifts even when nobody changes anything.

What a buyer asks, and what an auditor asks

These are two different conversations, and it helps to know which one you are in.

A buyer asks about the date itself. How many days the product has, how much of that life has to be left when it arrives (many wholesale customers set a minimum, often a fixed share of the total), where the date is printed, and what the storage line says. They are not asking how you ferment. They are asking whether you hit the number you gave them.
An auditor may ask what is behind it. What the date is based on, where that study is written down, whether the storage condition you tested is the one on the label, how the oldest stock gets shipped first, and what you did the time a batch came back.

Getting the first conversation wrong costs you a listing. The second one is the reason to keep the study and not just the conclusion.

Living with the date once it is printed

The date rides with the lot. A best-before is not a lot number and does not replace one. Two batches can carry the same date and still have to stay separable.
Oldest date leaves first. Picking by date rather than by whatever is nearest the door is the whole difference between a tidy tail and a pile of stock nobody can sell.
A new shelf life applies forward. If a study takes you from 90 days to 120, that is for what you fill after it. Bottles already labelled keep the date on them. Re-dating printed stock is not a shortcut, it is a different and much more serious problem.
Decide the tail in advance. At what point does stock get pushed, given away as samples, or written off? Writing that rule down once is easier than deciding it in a hurry with 300 bottles and three weeks left.

Frequently asked questions

Can I use the shelf life from a similar kombucha?

As a starting point to test, yes. As an answer, no. Their residual sugar, package, fill temperature and cold chain are not yours, and the number you print is a claim about your product. Borrow it as a hypothesis, then open your own bottles on a schedule and find out whether it holds.

Do I need a laboratory for this?

Not to start. A pH meter you calibrate, an alcohol method you can defend, a way to check carbonation and your own senses, all written down with dates, are already a study. A lab is worth paying for at the points where a number decides something, especially the last day of the shelf life, or when a customer or an authority wants figures that did not come from you.

Best before or use by?

They are different claims. A use-by date is a safety date, used for foods that go from fine to dangerous quickly. A best-before is about quality. Most kombucha carries a best-before, but which one applies to your product in your market is a question for your food authority, and it is worth asking once rather than inheriting the answer from another brand's label.

BrauMo keeps the date with the lot instead of in your head: the shelf life lives on the recipe, every lot takes its own best-before from its own packaging date, stock is picked oldest date first, and you get told before a lot reaches its date. Retention samples are logged with a code and a place, and a reading you take from one later goes on the same batch as the readings from brew day, so the evidence behind the date sits with the batch it came from. See how batches, lots and dates fit together →