Still the product you signed off
Sensory shelf life is how long a product still resembles the one that was approved. It is a date you measure, not one you inherit from the lab analysis.
Asked lot after lot, the same question becomes sensory quality control of suppliers: whether what arrives this week is what was signed off. Both forms of the question need the same two things: a stable reference and a criterion written down before you start.
The expiry consumers can taste
A product can be perfectly safe and have stopped being acceptable long before: the aroma fades, a rancid note appears, the texture changes. A best-before date is set from microbiological and physicochemical criteria, and neither of them sees that moment.
A panel does, provided it carries the same scale from the first session of the study to the last one, months later. How that scale is trained and checked is explained in full in SensoExpert Lab.
Extending a date without measuring the sensory side is the fast route to a withdrawal over consumer complaints that no microbiological parameter had anticipated.
Four decisions, and the first one governs
A shelf-life study is set up before the first sample is served. These are the four decisions, in the order they have to be taken; leave the first one until the end and what you have is a table of scores with no verdict.
You decide what counts as failure
Which attribute, how much deviation and against which reference: that is what separates “still the same product” from “no longer the same product”. Without that criterion written down, the study cannot conclude.
We keep a stable reference
We freeze or hold product in conditions that halt its evolution, so there is something to compare against. Comparing today’s product with a memory of how it used to be is not a measurement.
The panel measures over time
In one session the panel assesses samples of different ages, coded and in balanced order. That is what is distinctive about the design: the trend shows within a single session, not across sessions months apart.
You place the cut-off
With the curve for each attribute in front of you, you mark when the failure criterion is crossed. That is the number that goes on the label and into the conversation with retail.
What the study gives you, and what you decide
- The curve
- How each attribute evolves through storage, rather than one pass-or-fail verdict. Each is tracked separately because they do not all move at the same pace, and the first one to cross the criterion is the one that sets the date.
- The date and its criterion
- The point where the failure criterion is crossed, written down next to the criterion that defines it. A date with no criterion cannot be defended to a retailer or to an auditor.
- The control threshold
- That same criterion turned into the limit you then apply to every batch. It is the piece that connects the study to day-to-day quality control.
The same question, lot after lot
With a reference profile established and that threshold agreed, every incoming batch is measured against it. The deviation stops being an impression in a meeting room and becomes an attribute, a number and a limit that was written down in advance. A calibrated panel is what turns “this batch tastes odd” into something you can raise and correct.
This is the recurring form of the question you started with: not when the product expires, but whether this week’s is still the one that was approved. When the question is a one-off — validating a new supplier, a new plant — all you need is whether it differs from the control, and the design for that is shorter.
A shelf-life study does not start by measuring. It starts by deciding what counts as failure.
The study tells you when that threshold is crossed, and how confidently. It does not tell you where to put it: that is decided by whoever knows the product and what its label promises.
Which date do you have to defend?
Tell us the product, the storage conditions and the date you work with today. We will tell you how the study is set up and how soon it gives you the first usable result.