Condiment development: designing a storage and shelf life study

What to take away

• A shelf life figure has to come from your own product, in your own container, tested over time. It cannot be borrowed.
• Store samples the way customers will, including an opened set, not only sealed jars.
• Decide the failure criteria before the study starts, or you will move the goalposts.
• Safety limits and quality limits are different questions. Only the first is settled by your process authority.

Design the study before you make the samples

Write down four things first: how many jars, from how many batches, stored under which conditions, and tested at which intervals. Then write the criteria that will count as failure. Color change beyond a stated point, separation that does not redisperse, aroma loss, texture change, seal loss. Deciding this in advance is what separates a study from a series of opinions.

Use at least two production batches. A single batch tells you about that batch, and batch to batch variation is exactly the thing a shelf life claim has to survive.

Conditions worth including

Condition	Why include it
Ambient warehouse, sealed	The base case for a shelf stable product
Warm and humid	Represents a bad summer in transit or storage
Cold, sealed	Reveals separation and crystallisation
Light exposed	Shows color loss in clear packaging
Opened, refrigerated	The condition your label instructs
Opened, room temperature	The condition customers actually use

That last row is uncomfortable and necessary. People leave condiments on the counter. Knowing what happens is better than assuming nobody does it.

Pull enough samples. A study that runs out of jars at month four cannot be extended, and going back for more from a later batch destroys the comparison. Work out the number of test points, multiply by the number of conditions, multiply again by the number of batches, and then add a reserve for the tests you did not think of.

What to measure, and what to leave to a laboratory

Sensory scoring, color, viscosity and separation you can track yourself with a consistent method. Microbiological and chemical questions belong to a laboratory, and any safety relevant limit belongs to your process authority. Do not convert a quality observation into a safety conclusion. A jar that still looks fine is not evidence that a process is adequate.

Where you need to check what is already known about a hazard in a product like yours, the USDA food safety research projects database is searchable by category, commodity and funding source, covering bacterial pathogens, chemical contaminants, natural toxins and more. It is a reasonable way to see what has been studied before you design your own tests.

Storage conditions in the building itself

Once the product exists, where it is stored becomes an operational question rather than a study. Layout, temperature control, separation of raw and finished goods, and pest exclusion are design problems. AMS runs a wholesale market and facility design team that provides technical assistance on constructing or remodeling facilities including incubator kitchens and food hubs, following building codes, zoning and fire regulations. If you are fitting out a space rather than renting one, that is a source of help most small makers do not know exists.

For the consumer facing version of storage advice, see our blend storage rules, and for where this study sits in the wider sequence, see the development process guide.

Common questions

How long should a study run? At least as long as the shelf life you intend to claim, plus a margin.

Can I use accelerated testing? It gives early signals and does not replace real time data for a claim.

Who decides the final number? You do, on evidence, with any safety relevant limits set by your process authority.

What if a jar fails early? Stop, find the cause, and restart the study. A shortened study is not a shorter answer.