Price pack architecture is the deliberate design of a brand's ladder of pack sizes and price points, so that each rung serves a distinct shopper, occasion and channel rather than simply offering more of the same product for proportionally less money. It is the pack-and-mix lever inside revenue growth management, and it is where most consumer goods companies are hunting for growth now that four years of across-the-board price increases have stopped working.
There is an assumption buried in every price pack architecture deck, and almost nobody researches it: that the ladder you design is the ladder the shopper actually sees. It is not. The evidence has been sitting in the Journal of Marketing since 2003, and it is more damaging to the discipline than most brand teams realise, because the distortion is not random. It concentrates precisely on the SKU you care about most.
What a pack-price ladder is supposed to do
A well-built ladder is not a size range. It is a set of jobs, each assigned to a rung.
| Rung | The job it does | The shopper question behind it | What you actually need to research |
|---|---|---|---|
| Entry / trial pack | Buy a first-time trier at low risk | "Is this worth trying at all?" | Trial barrier, not price sensitivity |
| Core pack | Carry the volume and set the price reference | "Is this my normal?" | Reference price and repertoire position |
| Value / bulk pack | Reward planned, stock-up buying | "Am I being rewarded for committing?" | Occasion and storage reality, not unit maths |
| Premium format | Capture a different consumption moment | "Is this for a different situation?" | Occasion distinctiveness |
| Channel-exclusive pack | Prevent direct price comparison across retailers | "Is this the same thing cheaper elsewhere?" | Perceived comparability |
The logic that holds the ladder together is the quantity discount: a larger pack should cost less per unit than a smaller one. That expectation is so ingrained that it barely gets stated. Retail executives assume it, shoppers assume it, and the architecture assumes it.
The assumption nobody tests
The formal name for the ladder breaking is a quantity surcharge. Sprott, Manning and Miyazaki define it exactly: "Quantity surcharges occur when the unit price of a brand's larger package is higher than the unit price of the same brand's smaller package" (Journal of Marketing, 67(3), 34-46, 2003).
It is not a rare anomaly. The authors summarise earlier work finding quantity surcharges in 16% to 34% of supermarket brands available in two or more package sizes, and note a more recent investigation finding a 27% incidence across two U.S. markets. Their own dataset covered exactly 800 brands and 1,247 brand-size pairs. In it:
- 15.8% of all brands included one or more surcharges at the retail level
- 11.2% did at the wholesale level
- At the brand-size pair level, incidence was 11.1% retail and 8.4% wholesale
Read the retail-versus-wholesale gap carefully, because it is the whole story. The surcharge rate is consistently higher at retail than at wholesale. The manufacturer is not creating most of these inversions. The retailer is.
The surcharge is not an accident, and it lands on your hero SKU
The same paper ran an experiment on real grocery price setters, not students. Of 224 contacted, 197 (87.9%) agreed to participate and 161 (71.9%) completed the study. Each was given a realistic pricing scenario with a current price, cost, margin, category average margin, sales volume and competitor prices.
The result is the single most useful number in price pack architecture, and it almost never appears in a PPA deck. Facing lower competitive prices, price setters created quantity surcharges:
- 8.3% of the time when the focal item was a slow-moving SKU
- 79.1% of the time when the focal item was a top-moving SKU
That difference is not noise (chi-square = 39.38, p < .001). A second cut of the same data pointed the same way: when the smaller size had considerably higher sales volume than the larger one, 45.8% of price setters created a surcharge, versus 12.9% when both sizes were low volume.
The mechanism is not malice, and this matters when you take the argument to a buyer. Retailers price to build a low store-price image. The cheapest way to look cheap is to price aggressively on the items shoppers actually see and remember, which are the high-volume ones. The retailer discounts your best-selling size to signal value, leaves the larger pack where it is, and your ladder inverts above the rung with the most volume.
Only 14.6% of the surcharges observed were on temporary price promotion at the time of the study, so this is not a transient promo artefact. It is a structural feature of how the shelf gets priced.
The hero-rung inversion, stated plainly: you architect the ladder on the assumption of a well-behaved quantity discount, and the retailer re-architects it at the exact rung your volume depends on. The better your core pack sells, the more likely the ladder above it is broken.
Why shoppers do not correct the error
If shoppers audited unit prices, a broken ladder would punish itself and the market would fix it. They do not.
Gourville and Koehler, in the Harvard Business School working paper Downsizing Price Increases: A Greater Sensitivity to Price than Quantity in Consumer Markets (March 2004), documented across marketplace data and laboratory studies that consumers respond far more strongly to changes in price than to equivalent changes in quantity. Their illustration is the clean one: shoppers prefer a product presented as "$6 per half-pound" over the same product at "$12 per pound", although the two are identical.
Put the two findings together and you get the honest picture. The shelf price is doing nearly all the communicating. The per-unit figure, which is where your entire value argument lives, is the part shoppers are least sensitive to. A ladder whose value proposition can only be recovered by division is a ladder that does not communicate.
What to research, and what not to bother researching
Most pack research asks the wrong question. It asks people to choose between sizes and prices, which produces a tidy demand curve for a shelf that does not exist.
Research the occasion first, the ladder second. The rungs are only real if they map to genuinely different consumption situations. If your 6-pack and your 12-pack serve the same occasion for the same person, you have not built a ladder, you have built a discount.
| Question you need answered | Method | Where to start |
|---|---|---|
| Which occasions exist, and which pack does each one call for? | Qualitative depth interviews at scale | pricing research interviews |
| What is an acceptable price range for a given pack? | Van Westendorp price sensitivity meter | van Westendorp |
| What is the revenue-maximising price on one rung? | Gabor-Granger | Gabor-Granger method |
| How do size, price, format and brand trade off? | Conjoint analysis | conjoint analysis guide |
| Which subset of packs maximises reach without cannibalising? | TURF | TURF analysis |
| Which pack attributes matter most? | MaxDiff | MaxDiff analysis |
| Will a new pack read as a different product or the same one shrunk? | Packaging concept testing | packaging concept testing |
Two additions that standard PPA research skips, and that follow directly from the evidence above:
- Test the ladder as shoppers meet it, inverted rungs included. Show a realistic shelf in which the larger pack carries a higher unit price, and find out whether anyone notices, whether it changes what they buy, and what they conclude about your brand when it is pointed out. Nobody researches the broken version of their own ladder.
- Ask what was next to it. A pack does not compete against your other packs in the abstract. It competes against whatever the retailer put beside it that week, and against private label, which now sets the reference price in many categories.
Be careful not to confuse this work with a straightforward price increase study, which asks whether one price can move. Price pack architecture asks whether a portfolio still hangs together, and those fail in different ways.
Running pack-price research without a six-figure pricing project
Historically this was consultancy work: a pricing firm, a conjoint specialist, twelve weeks, and a deck. That is why most ladders get set on judgement and a spreadsheet of elasticities borrowed from a syndicated dataset.
Koji is the AI-native customer research platform built for exactly this gap. You write the brief, Koji runs AI-moderated voice or text interviews with several hundred category shoppers, and the thematic analysis comes back automatically. Because the moderator is an AI consultant you configure, every respondent gets the same disciplined probing on why they reached for the size they did, with no interviewer nudging them toward the answer the brand hopes for.
The part that matters for pack research is that you do not have to choose between depth and structure. Koji supports six structured question types alongside open conversation: open_ended, scale, single_choice, multiple_choice, ranking and yes_no. So a single study can rank pack formats, score price acceptability, capture a yes_no on whether the shopper noticed the unit price, and still spend ten minutes finding out what the pack is genuinely for in that household. See the structured questions guide for how the two combine.
Against the legacy stack, the difference is not subtle. Typeform and SurveyMonkey give you a form that cannot ask a follow-up. Qualtrics gives you conjoint modules and a specialist to operate them. UserTesting and dscout give you a handful of sessions and a stack of video to watch. All of them make the occasion question, the one that decides whether your ladder is real, the most expensive thing to answer. Koji makes it the cheapest, and turns it around in hours rather than weeks.
Frequently Asked Questions
What is price pack architecture?
Price pack architecture is the design of a brand's range of pack sizes and price points so that each pack serves a distinct shopper, occasion or channel. It sits inside revenue growth management alongside base pricing, promotion and trade terms, and it is judged on whether the portfolio grows total category value rather than shuffling volume between your own SKUs.
What is a quantity surcharge?
A quantity surcharge occurs when the unit price of a brand's larger package is higher than the unit price of the same brand's smaller package, reversing the quantity discount shoppers expect. Sprott, Manning and Miyazaki found surcharges on 15.8% of brands at the retail level in a sample of 800 brands and 1,247 brand-size pairs, and cite earlier estimates ranging from 16% to 34% of multi-size supermarket brands.
Why would a retailer price my larger pack above my smaller one?
Usually to build a low store-price image. Retailers price aggressively on the items shoppers notice, which are the high-volume ones. In an experiment with 161 grocery price setters, surcharges were created 79.1% of the time when the focal item was a top-moving SKU, compared with 8.3% when it was slow-moving. Your best-selling size gets the sharp price and the ladder inverts above it.
Do shoppers notice when the ladder is broken?
Mostly not. Research on downsizing shows consumers are considerably more sensitive to changes in price than to equivalent changes in quantity, and the per-unit comparison that would expose a surcharge requires arithmetic few shoppers do at the shelf. This is why a value claim that only works after division tends not to land.
Which research method should I use for pack and price decisions?
Use depth interviews to establish which occasions genuinely exist, then Van Westendorp or Gabor-Granger for acceptable and revenue-maximising price points on a single rung, conjoint when size, price and format trade off against each other, and TURF to choose the smallest set of packs that maximises reach. The sequence matters more than the technique: a statistically clean ladder built on occasions that do not exist is still wrong.
How many shoppers do I need for pack-price research?
For directional occasion work, 30 to 50 depth interviews per market usually reaches saturation. For quantified price acceptability you want a few hundred responses per rung so subgroup cuts stay readable. See the survey sample size guide for how to size this properly against the decision you are making.
See your ladder the way shoppers do
The gap in most price pack architecture work is not modelling capacity. It is that nobody has asked several hundred real category shoppers what the pack is for, whether they ever compare unit prices, and what they thought when the big pack turned out to cost more per unit.
Koji runs that study in days. AI-moderated interviews at panel scale, automatic thematic analysis, structured questions where you need numbers, and a one-click report you can take to a range review. No research team required, and no twelve-week consultancy engagement.
Start with the occasion, not the spreadsheet, and your ladder will survive the shelf.
