Where Did the Difference Enter?

A difference does not identify its source

Two people encounter what appears to be the same product and respond differently. One notices bitterness. One does not. One finds a package intuitive. One struggles with it. One tries a product once and returns. Another incorporates it into a routine.

The difference is real. But the observed difference does not tell us where it entered the system.

It is tempting to jump directly from outcome to explanation: the product is too bitter; the user is unusually sensitive; the message failed; the segment is wrong; the measure is noisy. Any of those explanations may be correct. None follows automatically from the difference itself.

This is a diagnostic question we keep returning to: where did the difference enter?

A source-of-difference map

A useful map begins before perception and continues beyond it.

First, the product-generated stimuli may differ. Formulation, material properties, temperature, delivery, packaging, instructions, device performance or storage may have changed what was available to detect.

Second, exposure and sampling may differ. People do not take identical sips, bites, puffs, touches or glances. Contact time, sniffing, oral processing, handling, pacing, sequence and adaptation can change the sensory information reaching receptors even when the product specification is unchanged.

Third, the person and state may differ. Sensitivity, anatomy, saliva, age, prior exposure, fatigue, hunger, illness, medication, thermal state and many other biological conditions can alter what is detected or how strongly it is represented. Martha Bajec’s work on thermal taste is a direct example: temperature changes can evoke taste-like sensations in some people, and thermal responsiveness is associated with broader differences in oral sensation.

Fourth, attention and expectation may differ. Labels, brands, prior experience, social information, price, instructions and goals can direct sampling and alter interpretation before or during contact.

Fifth, perception and meaning may differ. The same sensory pattern can be organized, named and valued differently. Familiarity, culture, memory and purpose can change what a signal means without changing the signal itself.

Sixth, behaviour and constraint may differ. A person can like a product and still not choose it. Availability, cost, ritual, effort, competing goals, social acceptability and habit can intervene between experience and action.

Finally, measurement and analysis may introduce or conceal the difference. Instructions, scale anchors, response formats, timing, task order, sample composition and aggregation all shape what becomes visible in the data.

The stages interact

This is not a claim that every difference has one clean location. The stages interact.

A label can change expectation, expectation can alter sampling, sampling can change the sensory input, and the resulting experience can alter later behaviour. Physiological state can change pleasantness, which changes attention and pace. Repeated use can change both the person and the meaning of the product.

The point of the map is not to force a complex system into boxes. It is to stop an outcome from being treated as though it already contains its own explanation.

What evidence would distinguish the explanations?

The next step is to turn plausible explanations into discriminating evidence.

If the product is suspected, characterize the relevant physical and sensory properties under controlled conditions. If exposure differs, measure or standardize sampling and use behaviour. If physiology is plausible, identify the state variables or stable individual differences that could alter response. If expectation matters, manipulate information independently of the product. If behaviour is the outcome, measure behaviour rather than asking liking to impersonate it. If measurement is suspect, compare tasks, anchors, order, timing or analytic choices.

The strongest design is not the one with the most measures. It is the smallest credible design capable of separating the explanations that would lead to different decisions.

Why the question matters commercially

Where the difference entered determines what can reasonably be changed.

Reformulation is useful when the evidence locates the problem in product-generated stimuli. It is expensive and potentially counterproductive when the problem lies in delivery, instructions, population, state, context, use sequence, measurement or the gap between liking and adoption.

A difference in consumer response is therefore the beginning of the investigation—not the end of it.

Find the source before deciding what to change.