Flavour reordering is one of the least glamorous decisions a customer makes and one of the most consequential for a liquid flavour shop's repeat business. The question worth sitting with is not whether to reward loyal reorderers, but how many rungs to put on the ladder: two skill tiers or five. The answer, drawn from research on goal gradients and decision-making under uncertainty, points toward fewer tiers, more clearly earned — and it changes how you design the badge system itself.
The Goal-Gradient Problem With Five Tiers
Behavioral research has a long record of the "goal-gradient effect," first documented in the 1930s by Clark Hull and later applied to consumer loyalty by Ran Kivetz and colleagues in a 2006 Journal of Marketing Research study. Their finding, replicated many times since: effort and engagement accelerate as people approach a goal, and the acceleration is steepest when the remaining distance is short and legible. A customer with four of five stamps filled behaves differently from one with two of five, and both behave differently from a customer with one of two.
The problem with a five-tier badge system is that it creates long stretches where the customer is, psychologically speaking, in the middle of nowhere. Tier two of five and tier three of five feel nearly identical to a reorderer who is buying the same bottle of mango-nicotine blend every three weeks. The distance to the next badge is neither short enough to trigger the sprint nor novel enough to register as progress. The middle tiers become friction: they add visual complexity to the account page without adding motivational pull.
Two tiers, by contrast, keep the customer in one of two legible states — earned or not-yet-earned — and the not-yet-earned state is always within plausible reach. The goal gradient does its work because the goal is never far.
Why the Middle Tiers Fail Specifically for Reorders
Reordering is a low-attention behavior. A customer who has settled on a flavour is not comparison-shopping each cycle; they are executing a habit. Badge systems that require the customer to notice subtle tier distinctions compete with that habit and usually lose. Five tiers assume a level of engagement with the loyalty program that the reorder behavior itself does not generate. Two tiers assume only that the customer can tell whether they've crossed a line — which is roughly the level of attention a habitual buyer has available.
Variable Reinforcement and the Case for Earned Scarcity
B.F. Skinner's work on schedules of reinforcement established that variable-ratio schedules — where a reward follows an unpredictable number of responses — produce the most persistent behavior. This is commonly invoked to explain why intermittent rewards hold attention, and it is often misapplied to loyalty design as an argument for randomized perks.
The more useful reading for a flavour shop is narrower. Variability works when the underlying behavior is already established and the reward is a surprise layered on top. It works poorly as the primary structure, because a customer who cannot predict whether a reorder will be rewarded learns to attend to the reward rather than the product. That attentional shift is corrosive in a category where the product experience — taste, consistency, throat hit — is the actual retention mechanism.
Two skill tiers, earned deterministically, do something different: they make the reward predictable enough to be trusted and scarce enough to be valued. The "skill" framing matters here. A badge labeled for a specific behavior — say, a flavour-rotation badge for customers who have tried three distinct profiles, or a consistency badge for a set number of reorders — tells the customer what they did, not just that they spent money. That is closer to a competence signal than a spending signal, and competence signals tend to survive the novelty phase that kills most loyalty programs.
Loss Aversion and the Cost of Demotion
Daniel Kahneman and Amos Tversky's prospect theory established that losses loom larger than equivalent gains. In a tiered system, this asymmetry cuts in a specific direction: the threat of dropping a tier is a more powerful motivator than the promise of climbing one, and it is also a more reliable source of resentment.
Five-tier systems almost always include a demotion mechanic, whether explicit (spend thresholds that reset) or implicit (inactivity windows). Each demotion event is a small loss, and the customer experiences it as such — disproportionate to its actual value. A customer who loses a mid-tier badge after a slow month does not simply return to baseline; they carry a grievance that can outweigh the accumulated gains from every badge they ever earned.
Two tiers can be structured to avoid this entirely. The first tier is a threshold that, once crossed, is permanent — a genuine status rather than a rolling window. The second tier can be a recurring recognition that resets without penalty, framed as a seasonal distinction rather than a rank. The design goal is to eliminate the demotion event, not to soften it.
A Concrete Illustration
Consider a shop that runs a five-tier program with quarterly spend thresholds. A regular customer who reorders every three weeks clears tier three comfortably in most quarters but occasionally dips to tier two during travel or a stretch of trying competitor products. Over two years, that customer experiences roughly three demotions. Each one is minor in dollar terms — a small discount difference — but each one is a loss, and the cumulative effect is a customer who now thinks of the program as something that can be taken away. The same customer under a two-tier system with a permanent first tier and a non-punitive seasonal second tier has no demotion events to remember, and the badge retains its meaning because it was never revoked.
Design Implications for a Liquid Flavour Shop
The practical translation is not "fewer badges because simpler is better." It is that each badge should be doing identifiable motivational work, and most five-tier systems have two or three tiers doing none. Before adding a tier, ask what behavior it is meant to elicit and whether the goal gradient at that distance is steep enough to matter. If the answer is vague, the tier is decoration.
Three design commitments follow. First, anchor each tier to a behavior the customer can name — a rotation milestone, a consistency streak, a review contribution — rather than to raw spend, which the customer cannot easily track and which invites comparison with other customers. Second, make the first tier permanent once earned. The loss-aversion literature is unambiguous about the cost of revocation, and there is no offsetting benefit that justifies it for a habitual reorderer. Third, keep the second tier's criteria visible and within a single reorder cycle's reach, so the goal gradient stays active rather than decaying into background noise.
The forward-looking question for shop operators is not how to add more recognition but how to make the recognition that exists carry weight. As the category matures and loyalty mechanics become table stakes, the shops that retain reorderers will be the ones whose badges mean something specific and cannot be quietly withdrawn. Two tiers, earned and held, do more work than five that flicker.