In flavour retail, the difference between a bundle that gets reordered and one that quietly lapses often comes down to how the offer is worded. A recent split-test across a mid-sized e-liquid catalogue found that reframing a multi-bottle bundle as a way to avoid losing a discount lifted repeat orders by roughly 11% over an equivalent bundle framed as a gain. Why should a phrasing change move a replenishment decision that is, on its face, purely practical?
The Asymmetry at the Heart of the Question
The result is less surprising once you place it against prospect theory. Kahneman and Tversky's central finding is that losses loom larger than equivalent gains — the pain of giving up something is felt more acutely than the pleasure of acquiring the same thing. In a replenishment context, the "loss" is not money already spent; it is a benefit the customer is about to forfeit. A gain-framed bundle says: buy three, get 15% off. A loss-framed bundle says: your 15% bundle rate expires in four days — reorder to keep it. The arithmetic is identical. The psychology is not.
What makes flavour retail a particularly clean setting for observing this is that reorders are habitual and low-deliberation. The customer already knows what they like; the decision is largely about when and whether to restock, not what to choose. That is exactly the kind of routine decision where a small shift in framing can tip behaviour, because there is no strong preference to override — only inertia to interrupt.
Variable-Ratio Reinforcement and the Flavour Catalogue
Behavioural psychology offers a second lens that flavour retailers rarely name but constantly exploit: variable-ratio reinforcement. B.F. Skinner's schedules of reinforcement showed that behaviour rewarded on an unpredictable ratio of responses is more persistent than behaviour rewarded on a fixed schedule. The classic illustration is the slot-machine-like persistence of pigeons pecking a lever that pays out irregularly.
Flavour catalogues map onto this structure almost too neatly. New releases arrive on no fixed cadence. Limited batches sell out. A flavour a customer loved in March may be gone by June, replaced by something adjacent but not identical. This irregularity is why enthusiasts check back, sample, and discuss — and it is why a well-designed reorder prompt can work: it arrives at a moment of genuine uncertainty about availability.
The ethical line here matters. Variable-ratio structures can build genuine engagement with a craft (tasting, comparing, discovering) or they can manufacture compulsion. The 11% lift is interesting precisely because it appears to operate on loss aversion around a known benefit, not on novelty-seeking. That is a healthier mechanism: it reminds people of value they already hold rather than dangling an unpredictable reward.
Loss Framing Without Manipulation
The temptation, once a team sees an 11% number, is to escalate. Countdown timers. "Last chance" banners on every page. Expiring "rewards" that were never really earned. This is where loss framing curdles into dark-pattern territory, and where the academic literature is cautionary. Loss aversion is real, but it is also easy to overdraw — customers who feel repeatedly manipulated develop what researchers describe as reactance: a motivated pushback against perceived pressure. The short-term lift becomes a long-term churn problem.
The distinguishing test is whether the loss being framed is real. A bundle rate that genuinely expires is a legitimate loss frame. A discount that resets the moment you leave the page is a fabricated one. In flavour retail, the legitimate version is straightforward: if a customer has an established bundle price, telling them it lapses is honest information delivered in the frame that matters to them. The manipulation risk rises when the frame is used to invent scarcity that does not exist.
A Concrete Illustration
Consider a hypothetical but representative test. A shop offers a "Build Your Own Six" bundle at a per-bottle discount. Group A sees: Save 18% when you bundle six. Group B sees: Your 18% bundle rate is held for 72 hours — after that, bottles revert to list price. Both groups face the same economics. Group B, across a large enough sample, reorders at a meaningfully higher rate — in the range of the 11% figure — because the message is anchored on what they stand to lose rather than what they stand to gain. The effect concentrates among customers with an existing bundle history, which is consistent with loss aversion operating on an established reference point.
Decision-Making Under Uncertainty and the Reorder Moment
There is a third layer worth naming: the reorder decision is made under genuine uncertainty about the future. Will my favourite still be in stock? Will the formula change? Will the price hold? Research on decision-making under uncertainty suggests that people manage this by anchoring on a reference state — usually the status quo — and evaluating changes relative to it. Loss framing works partly because it makes the reference state explicit: this is what you have; here is what you lose by not acting.
For flavour retailers, that suggests a design principle. The most effective reorder communication is not a generic "buy again" nudge but a specific reminder of the customer's established terms — the bundle they built, the price they locked, the cadence they've settled into. The message treats the customer's history as the reference point and frames the prompt as protecting it. That is a fundamentally different posture from acquisition marketing, which is all gain frames and novelty.
Where This Points Next
The forward-looking question is not whether loss framing works — the evidence, both experimental and now commercial, says it does — but how to deploy it in a way that compounds trust rather than erodes it. Three directions seem worth pursuing.
First, personalise the reference point. The 11% lift likely understates what is achievable when the loss frame is tied to a specific customer's actual bundle history rather than a generic offer. The more concrete the "what you'd lose," the stronger the effect — and the more honest it is.
Second, separate the frame from the pressure. Loss framing does not require urgency theatre. A calm, accurate statement of terms that will lapse is both more ethical and, on the reactance evidence, more durable than a flashing timer. The goal is to inform, not to corner.
Third, measure the decay curve, not just the lift. The interesting metric is not the 11% at first reorder but whether loss-framed customers remain customers at month twelve. If the frame builds a habit of protecting a valued arrangement, retention should hold. If it merely extracts one extra order, the number will fade — and that fading will tell you the frame was doing manipulative rather than informational work.
Flavour retail sits at an unusually honest intersection of craft and behavioural design. The customers who reorder are not passive; they have preferences, routines, and a reasonably good sense of when they are being played. Loss framing, used well, respects that: it tells people plainly what they have and what they'd give up. Used badly, it becomes the thing the literature warns against. The 11% is a signal worth following — provided the shop is willing to follow it toward clarity rather than pressure.