Why Your Yoghurt Set Is Failing: A Production Manager's Guide to Diagnosing Fermentation Problems

The batch that will not behave

Few problems in a dairy plant are as frustrating as a fermentation that stops cooperating. The recipe has not changed. The supplier has not changed. The operator has not changed. And yet the pH is crawling, the set is weak, and the shift supervisor is asking whether to hold the batch or dump it.

The instinctive response is to blame the culture. Cultures are the most visible variable in the process, they arrive in a box with a batch number, and changing supplier feels like decisive action. In our experience, culture quality is the cause far less often than it is the suspect. More commonly, the culture is the first thing that shows the strain of a problem that started somewhere else, in the milk, in the cleaning cycle, in the cold chain, or in the tank temperature profile.

First: get the curve, not the opinion

Before anything else, you need acidification data. Not a start pH and an end pH, but a curve: pH or titratable acidity logged at regular intervals through the fermentation, together with the vat temperature at each point. If you are not logging this, start today, because without it every conversation about fermentation performance is a conversation about memory.

The shape of the curve tells you a great deal. A curve that starts normally and then flattens mid-fermentation points to a different problem than one that never gets going at all. A curve that varies with the tank tells you the answer is in the equipment, not the culture. A curve that varies with the milk tanker tells you the answer arrived on a truck.

Compare the failing batch against your best batches from the same product. If you do not have a reference curve, build one from the next three good runs. This single habit resolves a surprising share of fermentation disputes with suppliers, because it converts an argument into evidence.

Second: ask what changed in the milk

Milk is not a constant. Composition moves with season, feed, stage of lactation, herd changes and blending between suppliers. Total solids and, in particular, the protein content drive gel strength; a drop in protein produces a weaker set even when acidification looks perfect on paper.

Standardisation is the other common culprit. If solids are being adjusted by powder addition or by membrane concentration, small errors in that step change the buffering capacity of the milk. Higher buffering means acid production has to work harder to move the pH, which reads on your curve as a slow culture when in fact the culture is performing exactly as it always has.

Practical check: pull composition data for the milk used in the failing batches and compare it against the good ones. If protein, total solids or the solids-not-fat figure has moved, you have your first suspect and it is not the culture.

Third: rule out inhibitors

Three things stop a culture that is otherwise healthy: antibiotic residues, sanitiser carryover, and bacteriophage. They present differently and it is worth learning to tell them apart.

Antibiotic residue usually produces a near-total failure to acidify, affecting whichever batches used the contaminated milk, and it is straightforward to screen for at intake. If your plant does not test every delivery, this is the cheapest control you will ever install.

Sanitiser carryover, typically chlorine or quaternary ammonium compounds left in a line or a tank after cleaning, tends to be tank-specific and intermittent. If failures cluster on one vessel or one shift, look hard at the CIP rinse step, the final rinse water, and whether anyone has changed a chemical concentration recently.

Bacteriophage is the one that production teams most often miss, and it is the most important to catch. Phage are viruses that infect and kill lactic acid bacteria. The classic pattern is a fermentation that starts normally and then stalls, worsening over successive batches, often affecting some strains while others in the same blend continue working. Phage build up in an environment, in whey aerosols, drains, and residues on equipment, which is why the problem gets progressively worse rather than appearing suddenly. The controls are strain rotation, strict separation of whey from the fermentation area, air management and rigorous cleaning. If your acidification has been drifting slowly worse over weeks rather than failing overnight, treat phage as the leading hypothesis until you have excluded it.

Fourth: look at how the culture is handled, not just which culture it is

Frozen and freeze-dried cultures are living material and they are sensitive to the journey. Temperature excursions in transit, a freezer that is opened repeatedly through a shift, a partially used sachet resealed and stored warm, or thawing at ambient rather than to the supplier's instruction: all of these reduce viable cell counts before the culture ever reaches the vat.

Inoculation accuracy matters just as much. If dosing is done by eye or by a scoop rather than by weight, the actual inoculation rate can move considerably between operators. A lower inoculation rate produces a slower curve, which then gets recorded as a culture problem.

In the Gulf specifically, the cold chain deserves particular scrutiny. Cultures typically travel a long way to reach a regional plant, and the last stages of that journey, airport handling, customs clearance, local transport, are the stages with the least visibility. If you cannot demonstrate an unbroken temperature record from the supplier's freezer to yours, that gap is a genuine risk, and it is one of the reasons regional producers are increasingly asking for locally held stock and local technical support.

Fifth: control time and temperature properly

Set-point temperature is not the same as actual temperature. Large vats develop gradients, and a probe in one position may not represent the mass. If two tanks with the same set point behave differently, map the temperature at several points and depths in each before concluding anything about the culture.

Then look at what happens after fermentation. Cooling too slowly allows over-acidification and a sour, shrinking product. Cooling too abruptly, or pumping and agitating a set gel, breaks the protein network and produces syneresis, the watery layer on top that customers notice immediately. Many complaints that arrive labelled as fermentation problems are in fact post-fermentation handling problems, and no change of culture will fix them.

What this actually costs

It is worth putting the commercial frame around this, because it changes how the problem gets prioritised. A failed fermentation is not only the value of the discarded milk. It is the tank hour lost, the labour on that shift, the cleaning cycle you now have to repeat, the delivery you cannot fulfil, and the customer confidence you spend the next quarter rebuilding. In a plant running near capacity, lost tank time is often the largest single item, and it is the one least likely to appear in the batch write-off report.

This is also why switching culture supplier is an expensive way to run an experiment. A supplier change brings requalification, a period of recipe adjustment, and the possibility that you have simply moved the same underlying problem to a new box. Diagnose first.

A workable order of investigation

If you are in the middle of this problem now, work through it in this order: log the acidification curve and compare against a known good batch; pull milk composition and standardisation records for the affected days; screen for antibiotic residues and review the CIP records on the affected vessels; check whether failures cluster by tank, shift or strain, which points towards phage or equipment; audit culture cold chain and inoculation practice; and map tank temperature and post-fermentation cooling. Most plants find their answer in the first three steps.

Where an external view helps

Plants that live with a fermentation problem for months usually do so not because the team lacks skill, but because the team is close to the process and the plant cannot afford to stop and investigate. An outside technologist brings a structured protocol, no attachment to the existing explanation, and pattern recognition from other factories that have had the same problem for a different reason.

Caseum & Co has spent over two decades in dairy production across Europe, the Middle East and Southeast Asia, building factories, launching cheese and cultured product lines, and solving exactly this class of problem. We work confidentially, we work to a defined scope, and we hand back a plant that your own team can run.

If your fermentation is not behaving, we will review your acidification data and your process with you in a 30-minute call at no cost, and tell you honestly whether you need us. Write to caseum.co@gmail.com or use the contact form at caseumco.com.