Flavour sachet bundles — the multi-pack configurations that let a buyer mix, say, mango, menthol, blue raspberry, and tobacco in a single order — have a curious ceiling. Retail data from online vape and flavour retailers repeatedly show that once a bundle climbs past four distinct options, the conversion rate falls off a cliff, even when the per-unit price keeps dropping. The question worth examining is why four functions as a soft boundary when seven should, by any simple economic logic, offer a better deal and therefore a better sell.
The Arithmetic That Should Work and Doesn't
The rational-actor model predicts a straightforward relationship: more variety at a lower unit cost should increase purchase intent. A seven-flavour bundle priced at, hypothetically, $2.10 per sachet beats a four-flavour bundle at $2.60 per sachet on both dimensions a buyer claims to care about. Yet the four-option configuration outsells it, and not by a trivial margin.
This is not a flavour-industry quirk. It mirrors findings from jam-tasting experiments that have circulated in behavioral economics for two decades. In the best-known version, shoppers in an upscale grocery store encountered a display of either six or twenty-four jam varieties. The larger display attracted more browsing attention — roughly 60 percent of passersby stopped, versus 40 percent for the smaller one — but converted to purchase at roughly one-tenth the rate. The six-jam table sold more. The lesson was not that variety is bad but that variety past a certain point imposes a cost the buyer pays in cognitive currency rather than dollars.
For sachet bundles, the threshold appears to sit lower than six. Four is where the curve bends.
Why Four Sits at the Edge of Working Memory
The Chunking Constraint
The most parsimonious explanation draws on George Miller's 1956 work on the "magical number seven, plus or minus two." Miller argued that immediate memory holds roughly seven discrete items, but later research refined this: the effective limit is closer to four chunks when those items must be held, compared, and manipulated rather than merely recalled. Nelson Cowan's 2001 review put the working-memory capacity estimate at about four items for most adults under typical conditions.
A four-flavour bundle asks the buyer to hold four distinct flavour profiles in mind, weigh them against each other, and commit. A seven-flavour bundle asks for nearly double that load. The buyer does not experience this as "I am exceeding my working memory capacity." They experience it as vague unease — a sense that they haven't properly considered all the options, that they might regret a choice they can't fully evaluate. That unease suppresses the click.
Anticipated Regret and the Cost of Exclusion
Kahneman and Tversky's work on loss aversion established that losses loom larger than equivalent gains. In a bundle context, every flavour not selected registers as a small loss relative to the flavours selected. A four-option bundle generates a manageable number of foregone alternatives. A seven-option bundle generates a combinatorial explosion of perceived exclusions: not just which flavours you left out, but which combinations you'll never taste.
This is where the sachet format differs from, say, a single-flavour purchase. The bundle implies a curated experience — a rotation, a sampler, a set. The buyer isn't choosing one thing; they're choosing a set and implicitly committing to use all of it. Seven flavours means seven obligations to finish, seven chances to discover a dud, seven small regrets if the menthol turns out to be the only one they actually wanted.
What the Four-Option Ceiling Reveals About Preference
There's a second-order effect that matters for anyone designing or marketing these bundles. When a buyer faces four options, the decision feels tractable enough that they can form a genuine preference — "I want the fruit-forward set" or "I want the menthol-heavy set." When they face seven, the decision shifts from preference expression to optimization, and optimization under uncertainty is aversive.
This connects to Herbert Simon's concept of satisficing. People don't maximize; they find an option that clears an acceptable threshold and stop. A four-option bundle makes satisficing easy — any reasonable mix clears the bar. A seven-option bundle makes satisficing feel irresponsible, because with seven options there's presumably a best one, and the buyer hasn't found it. The result is delay, abandonment, or a retreat to a familiar single-flavour purchase.
Retailers who have experimented with bundle sizing report a consistent pattern: the four-option bundle converts best, the five-option bundle is roughly comparable, and the six- and seven-option bundles underperform despite superior unit economics. Some have tried to rescue the larger bundles with "build your own" interfaces, only to find that the customization burden makes the problem worse rather than better.
Variable-Ratio Reinforcement and the Bundle Habit
There's a reinforcement angle worth noting, though it cuts against the intuitive reading. Flavour sachets are consumables with inherently variable outcomes — each new flavour is a small trial, and the reward (a flavour you love) arrives on an unpredictable schedule. This is the structure B.F. Skinner identified as variable-ratio reinforcement, and it's the same structure that makes slot machines and social media feeds compulsive.
But variable-ratio schedules work best when the response cost is low. Pulling a lever is cheap. Committing to a seven-flavour bundle is not — it's a $15 to $25 outlay and a weeks-long consumption commitment. The reinforcement structure encourages repeat purchase, but only if the initial purchase clears the friction threshold. Four options clear it. Seven do not. The irony is that the larger bundle, which should generate more variable-ratio trials per purchase, generates fewer purchases overall.
Designing Forward: What the Ceiling Suggests
The practical implication is not that bundles should be capped at four forever. It's that the constraint is cognitive, not economic, and cognitive constraints can be engineered around.
One direction: sequential bundling. Offer a four-flavour bundle with a clearly framed "next set" that unlocks after purchase — a subscription ladder rather than a single large commitment. This preserves the low-friction entry while building the variable-ratio habit over time. The buyer doesn't face seven options; they face four, then four more.
Another: themed curation. A seven-flavour bundle framed as "the complete menthol collection" or "the dessert rotation" reframes the choice from which seven to whether this theme appeals. The working-memory load collapses because the buyer evaluates one coherent concept rather than seven discrete items. Early tests of themed bundles suggest they outperform generic large bundles, though rarely match the four-option baseline.
The most forward-looking move may be to stop treating four as a limitation and start treating it as a design parameter. The four-option bundle isn't a compromise; it's a configuration that matches how people actually decide. Retailers who internalize that will stop chasing the seven-option upsell and start investing in the quality of the four-option set — better flavour pairing, clearer tasting notes, smarter rotation logic. The ceiling isn't a wall. It's a signal about where the real design work begins.