Probiotic CFU: Guaranteed at Expiration, Not at Manufacture

Pick up any probiotic on the shelf at a grocery store or a Whole Foods and read the front of the bottle. The number will be big — 30 billion CFU, 50 billion CFU, 100 billion CFU. It is the largest text on the label. It is meant to be the reason you pick it up.

Then look at the fine print on the back — the Supplement Facts panel, the "manufactured for" line, sometimes a footnote in a lighter font under the ingredient block. Somewhere in there, most probiotic bottles will say some version of "guarantees X billion CFU at time of manufacture" or "potency verified at production."

Those two lines are the entire story of the probiotic industry, and almost nobody reads the second one.

Probiotics are living microorganisms. They die during shipping. They die on the pharmacy shelf. They die in your bathroom cabinet. They die at rates that depend on the strain, the encapsulation, the storage temperature, the humidity, and the time on the shelf. The 50-billion-CFU number the manufacturer measured at the end of the production line is not the number that reaches your gut. In many products it is not even close.

Here is what "at manufacture" versus "at expiration" actually means, what the CFU decay curves look like for real product formats, and the single line on the label that separates a probiotic dosed to work from one dosed to look good on the shelf.

What "CFU at manufacture" actually measures

CFU stands for colony-forming units. It is the standard way to quantify a live-bacteria population — you take a sample of the product, dilute it, plate it on selective agar, and count the colonies that grow. Each colony represents at least one viable, reproducing organism at the moment of plating.

"CFU at time of manufacture" means that the manufacturer plated a sample from the batch at the point the finished product left the production line, before the bottles were sealed, boxed, palletized, shipped, warehoused, and sold. That is a real number. It reflects a real quality-control check. What it does not reflect is the number of viable organisms in the capsule you are about to swallow, six to eighteen months later, after that capsule has been in trucks, warehouses, and store shelves that may or may not have been temperature-controlled.

The gap between manufacture and consumption matters because live bacteria do not sit inertly in a capsule the way a dry vitamin does. They are metabolically slowed by the low water activity of the capsule matrix, but they are not paused. Every day on the shelf, some fraction of the population dies. The death rate depends on strain, format, and storage — and it is almost always faster than the label wants you to notice.

The decay curves are steep

The published stability data for common probiotic strains is not flattering to the industry.

Lactobacillus rhamnosus GG — one of the most-studied strains and the one many premium probiotics feature — loses 30-60% of viability in twelve months at room temperature (25°C, 60% relative humidity) in a standard hard-shell vegetarian capsule. A bottle labeled "50 billion CFU at manufacture" that sits on a warm distributor shelf for four months, then in a warehouse for two months, then in your medicine cabinet for six months, delivers somewhere between 15 and 25 billion CFU by the last capsule in the bottle. The label is not lying. The label is simply not measuring what you care about.

Bifidobacterium species — Bifidobacterium longum, B. infantis, B. lactis — are worse. Bifidobacteria are anaerobes with lower thermal tolerance than most lactobacilli, and they die faster on the shelf. Room-temperature decay rates of 50-80% over twelve months are typical in standard capsule formats. A Bifidobacterium-heavy blend labeled "100 billion CFU" at manufacture may reach the consumer at 20-50 billion CFU on the front-of-bottle capsules and closer to 10-20 billion CFU on the back-of-bottle capsules.

Refrigerated strains decay more slowly. A room-temperature-storage claim on a Bifidobacterium product is a red flag on its own — most strains that survive room-temperature storage well are the ones best evolved for it (spore-forming strains like Bacillus coagulans, or freeze-dried heat-tolerant Lactobacillus reuteri variants). The "no refrigeration needed" convenience of shelf-stable formats is real for a subset of strains and misleading for others.

Blended multi-strain products decay unevenly. The bottle says "50 billion CFU total, 12 strains." The strain that dominates the CFU count at manufacture is often not the strain that dominates the count at expiration, because the strains die at different rates. The consumer thinks they are taking a diverse population and are increasingly taking whichever strain happens to be most stable in the capsule matrix.

Why brands use "at manufacture" numbers

The convention exists for one reason: it makes the front-of-bottle number larger. A manufacturer that guarantees potency at expiration has to overformulate at manufacture — putting in 100 billion CFU on the production line so that 50 billion CFU is still viable when the customer opens the bottle. The one-hundred-billion-CFU spec is more expensive to manufacture but honest to advertise. The alternative — putting in 50 billion CFU and labeling it 50 billion CFU at manufacture — is cheaper and lets the front label say the same thing.

The customer sees identical front labels. Two bottles, both "50 billion CFU," both around $30. One has 50 billion CFU when you swallow the first capsule and something like 30-40 billion CFU by the last capsule. The other has maybe 25 billion CFU on the first capsule and 8-12 billion CFU by the last. The label discipline that separates them lives in the fine print, and most consumers never look at it.

The industry has known this for at least two decades. The Council for Responsible Nutrition (CRN) has published guidance recommending expiration-date labeling. The World Gastroenterology Organisation's practice guidelines specify that the number reported on a probiotic should be the minimum viable count at end of shelf life. The FDA does not require any of this. As long as the label is not affirmatively false — and "50 billion CFU at time of manufacture" is not false — the manufacturer is compliant.

The result is a market where the honest brands have a smaller front-of-bottle number than the dishonest ones for the same dose delivered.

What "guaranteed through expiration" actually looks like

The single line on the label that separates a real probiotic dose from a decaying one is some version of:

"Guarantees [X] billion CFU through the expiration date" "Potency guaranteed to expiration" "[X] billion CFU at end of shelf life"

Any of those means the same thing: whatever number is printed on the front is what you get out of the last capsule in the bottle, not the first. It is the honest number. If it appears on the label at all, it is nearly always in small print near the ingredient panel or the manufacturer address block, and it is one of the strongest single signals of a brand that respects the buyer.

The absence of that line — or the presence of "at time of manufacture" or "potency verified at production" — tells you the number on the front is a snapshot, not a floor. The actual dose the consumer receives is unknown and depends on the supply chain the bottle traveled through.

Reputable brands that use expiration-date labeling typically overformulate by 50-200% at manufacture to guarantee the front number at expiration. That is the cost of honesty in this category, and it is why an "expiration-guaranteed 50 billion CFU" bottle usually costs more than an "at-manufacture 50 billion CFU" bottle. You are not paying for a bigger number. You are paying for the number to still be there when you swallow it.

The strain-identification issue rides alongside

A CFU number alone is only half of a real dose. The other half is which strain, at which dose. Every probiotic strain that has ever produced a clinical outcome in a peer-reviewed trial has been identified down to the genus, species, and strain designation — for example, Lactobacillus rhamnosus GG (the "GG" is the strain identifier and matters for the clinical evidence), or Bifidobacterium longum 35624 (the strain in the branded IBS product Align), or Saccharomyces boulardii CNCM I-745 (the strain used in the antibiotic-associated-diarrhea trials).

A label that says "Lactobacillus rhamnosus, 10 billion CFU" without a strain designation is telling you that the product contains some strain of L. rhamnosus, not necessarily the one that has published clinical outcomes for the indication you are buying it for. The strain matters because two strains of the same species can behave completely differently — one may colonize the gut and produce measurable clinical effect, and another may pass through without doing anything the trial data would predict.

A probiotic label that discloses strain designation, guarantees CFU through expiration, and gives a per-strain CFU count within a blend is doing the three things the honest version of this category requires. Most retail probiotics do none of the three.

What the honest label looks like

Read down a bottle. The features that separate a real dose from a decaying one:

Strain-level identification. Every strain listed with a genus, species, and strain code (e.g., "Lactobacillus acidophilus NCFM," "Bifidobacterium lactis Bi-07"), not just "Lactobacillus acidophilus."

Per-strain CFU count. In a multi-strain blend, each strain has its own CFU number. "50 billion CFU total, split across 12 strains" tells you almost nothing. "10 billion CFU L. rhamnosus GG, 10 billion CFU B. lactis Bi-07, 5 billion CFU L. acidophilus NCFM..." tells you what you are actually taking.

Expiration-date CFU guarantee. The number is guaranteed at the end of shelf life, not at production. The line will use phrasing like "guarantees X billion CFU through expiration date" or "potency guaranteed to expiration."

Storage instructions that make thermodynamic sense. A product that includes Bifidobacterium strains and does not require refrigeration is a product that is either using a spore-forming or specifically stabilized variant (which should be disclosed) or is going to lose most of its Bifidobacterium load on the shelf. A room-temperature stability claim without technical justification is a red flag.

Third-party CFU verification. A brand that ships product with a certificate of analysis showing CFU at manufacture and at a post-shelf-life stability point is doing the work the label claim depends on.

The absence of any of these does not make the product fake. It makes the product unquantified — a bottle that may or may not deliver a therapeutic dose, and the buyer has no way to tell from the label.

What we do

We do not sell probiotics yet. When we do — and probiotics are on our roadmap — the label will guarantee CFU through the expiration date, disclose strain designation to the level the clinical literature uses, list per-strain CFU counts in any blend, and post the stability certificate of analysis for each batch. The number on the front will be the number in the last capsule of the bottle, not the first.

That is the version of this category that respects the buyer, and it is the version we would be embarrassed not to ship. It is also the version that costs more to make, which is why the shelf is full of the other version.

The next time you are in the supplement aisle looking at probiotics, turn the bottle over and find the CFU line. Read what comes after the number. That line is the whole product.