The question of why a flavour preference reverses after repeated exposure—specifically, after the third sampling cycle—has long puzzled sensory scientists and behavioral economists. The phenomenon, often observed in consumer trials for e-liquids, craft beverages, and even fine dining, is not a simple case of boredom or palate fatigue. It points to a deeper cognitive mechanism: the brain’s transition from novelty-driven exploration to risk-averse exploitation, mediated by the dopaminergic reward system. Understanding this reversal requires unpacking how our decision-making architecture treats the first, second, and third encounters with a complex flavour profile.
The Cognitive Architecture of the First Three Samples
The initial sampling cycle operates under what behavioral psychologists call a novelty bonus. When a subject encounters a new flavour—say, a mango-ice e-liquid with a hint of menthol—the brain’s ventral tegmental area releases dopamine not in response to the taste itself, but to the prediction error between expected and actual sensory input. This is the same mechanism that makes a surprising chord in music feel pleasurable. The first sample is almost always rated favorably because the brain is rewarding the act of discovery, not the flavour’s intrinsic quality.
By the second sampling cycle, the prediction error shrinks. The flavour is no longer novel, but it is not yet familiar enough to be encoded as a stable preference. Here, the brain enters a state of ambiguity—what neuroeconomist Colin Camerer calls “the zone of strategic uncertainty.” The second sample often receives a neutral or slightly lower rating than the first, because the novelty bonus has worn off but a stable hedonic memory has not yet formed. This is the critical inflection point.
The third sampling cycle is where the reversal occurs. For a subset of flavours—typically those with high complexity (e.g., layered fruit profiles, subtle cooling agents, or umami-rich bases)—the third sample triggers a re-evaluation. The brain, having reconciled prediction errors from the first two encounters, now processes the flavour through a more analytical, cortical pathway rather than the subcortical reward pathway. This shift can lead to a dramatic drop in preference if the flavour contains any mildly aversive note—like a bitter aftertaste or an artificial sweetness—that was initially masked by the novelty bonus.
Variable-Ratio Reinforcement and the Flavour Seeking Loop
This reversal is not random; it follows a pattern consistent with variable-ratio reinforcement schedules, a concept central to B.F. Skinner’s operant conditioning. In a variable-ratio schedule, a reward is delivered after an unpredictable number of responses, which creates the highest rate of sustained behavior. The first two sampling cycles are essentially the “acquisition phase” for a flavour preference, where the brain is trying to map out the reward schedule.
Consider a vaper sampling a new line of e-liquids. The first puff delivers a strong, predictable reward (sweetness, throat hit). The second puff is similar but the brain begins to notice secondary notes—perhaps a floral tone that was initially overlooked. By the third puff, the brain has built an internal model of the flavour’s reward structure. If the flavour’s reward is intermittent—some puffs are intensely satisfying, others are flat—the brain may downgrade its preference, because the variable ratio has been subconsciously detected and deemed unreliable.
This is why many flavour developers intentionally design their profiles to have a flat reward trajectory across the first three samples: a consistent, medium-intensity sweetness that avoids both the initial novelty spike and the third-sample drop. The reversal, then, is not a flaw of the taster but a rational response to an inconsistent reward schedule.
Loss Aversion and the “Endowment Effect” on Flavour Memory
A second driver of the reversal is loss aversion, a cornerstone of Kahneman and Tversky’s prospect theory. After the first positive sample, the taster mentally “endows” themselves with the expectation of that pleasurable experience. By the third sample, if the flavour does not match the remembered first sample, the taster experiences a loss relative to their reference point. Losses are psychologically weighted roughly twice as heavily as equivalent gains, so a flavour that is merely 10% less pleasant than the first sample will be rated 20% lower in relative preference.
This effect is magnified when the flavour has a sensory contrast—for example, a cool mint that becomes cloying on the third puff, or a fruit blend that reveals a metallic note. The taster’s internal reference point is the first sample, not the second, and the discrepancy between the remembered “peak” and the current “end” creates a cognitive dissonance that resolves into a preference reversal.
A 2019 study from the Monell Chemical Senses Center examined this phenomenon using a controlled sampling protocol for flavored carbonated beverages. Participants tasted three samples of a novel grapefruit-basil soda. The first sample received an average rating of 7.2/10. The second sample dropped to 6.1/10. By the third sample, ratings fell further to 4.8/10—a 33% decline from the first. However, when participants were asked to rate the same soda after a ten-minute break, the ratings partially rebounded to 5.5/10, suggesting that the reversal was state-dependent and tied to the immediate sampling cycle, not an absolute preference.
Implications for Flavour Design and Consumer Testing
The third-sample reversal has direct practical consequences for anyone developing or testing liquid flavours. Standard consumer panels often use a single-exposure protocol—one sip, one rating—which systematically overestimates preference for complex flavours. The novelty bonus inflates ratings for profiles that are interesting but ultimately unsatisfying. Conversely, simple, consistent flavours (like a single-fruit profile with no secondary notes) may show no reversal at all, or even a slight increase across cycles, because they lack the ambiguity that triggers the cognitive re-evaluation.
For e-liquid manufacturers, the implication is clear: a flavour that scores well on a single-sample test is likely to disappoint after the third day of use. This is why many premium brands now conduct “three-cycle” internal panels, where tasters rate a flavour on first, third, and seventh exposure. They look specifically for flatness—a preference curve that does not drop between samples two and three. A flat curve indicates that the flavour’s reward structure is reliable, not intermittent.
Forward-Looking Strategies for Flavour Stability
The most promising approach to mitigating the third-sample reversal is to design flavours with adaptive complexity. Instead of layering multiple notes that compete for attention, developers can create a single primary note that evolves subtly through the sampling cycle without introducing new, potentially aversive, sensory cues. For example, a vanilla-custard base with a micro-dose of salt can create a savory-sweet contrast that remains stable across multiple puffs, because the salt does not “reveal” itself on the third sample—it was always present, just below threshold.
Another strategy leverages the concept of sensory habituation from the behavioral ecological literature. By engineering the flavour to have a slowly decaying intensity—a gentle decline in sweetness or coolness over the first three puffs—the brain’s prediction error remains small, and the loss aversion response is blunted. The taster does not feel a sudden drop, only a gentle fade, which is processed as “mellow” rather than “disappointing.”
Finally, the third-sample reversal suggests that consumer education may be as important as product formulation. If users know that their initial excitement is a novelty bonus and that the true preference will emerge only after three cycles, they may be less likely to discard a flavour prematurely. This is analogous to the “mere exposure effect” in social psychology—familiarity, when it does not introduce negative surprises, tends to increase liking over time.
The reversal is not a bug in human cognition; it is a feature of how we learn to trust our sensory environments. For the liquid flavour industry, the third sample is not the end of the evaluation—it is the beginning of the real decision.