Price Pack Architecture Research and How to Run It (2026)

Last updated: 23 September 2026

Quick Answer: Run price pack architecture research by reading the pack ladder in scanner data, then testing shopper trade-offs with interviews and choice tasks. Sales data shows which sizes move; shoppers explain why a size gets skipped. A GAO analysis of 2021 to 2023 scanner data found downsized items were under 5 percent of items yet a larger share of sales.

In cereal, the US Government Accountability Office found that 1.1 percent of items were downsized between 2021 and 2023, and those items carried 8.6 percent of category sales. A few pack decisions sat on an outsized share of revenue.

Price pack architecture is not a pricing question with packaging attached. It is about how shoppers perceive quantity, and sales data answers only half of it. The other half decides whether a new size adds a buyer or moves an existing one down a rung.

What Does Price Pack Architecture Decide?

Price pack architecture (PPA) decides which sizes, counts and formats a brand sells, at which price tiers and in which channels, so that every pack has a job. Research tests whether each rung earns its shelf space.

A typical ladder assigns five jobs.

• Entry pack: the smallest commitment, for trial and cash-constrained buyers.
• Core pack: carries most of the volume and anchors the comparison.
• Value or large pack: rewards heavy users and competes on unit price.
• Premium small format: single-serve packs sold on convenience, not unit value.
• Channel packs: club multipacks, convenience singles, e-commerce bundles.

The ladder also carries a consumption effect. The Cochrane review by Hollands and colleagues, used here because it pools only randomized trials, combined 58 studies with 6,603 participants and found that larger portions, packages or tableware increased the amount of food people consumed. A bigger rung can change how fast a household runs out, which sales data later reports as loyalty.

Which PPA Questions Can Scanner Data Answer, and Which Need Shoppers?

Scanner and retailer data answer what happened: velocity by size, gaps between rungs, promotion lift and volume shifts after a change. Why a size was skipped, whether shoppers noticed a downsize, and how an unlaunched pack would split demand all need shoppers.

Rows are ordered from the existing ladder to a proposed one.

For elasticity on an existing ladder, syndicated scanner data from a provider such as NielsenIQ or Circana, or a retailer's loyalty data, beats any interview program.

NIST reports that 19 states and two territories have unit pricing laws or regulations, and only 11 of those jurisdictions make it mandatory. A ladder is read against a posted unit price in New York and often without one elsewhere, so record where each respondent shops.

How Do You Test a Pack Ladder With Interviews and Trade-Off Methods?

Show the ladder as a set, ask shoppers to choose, then probe the choice. Choice tasks estimate how demand splits across rungs; interviews explain the split, including quantity misreadings a model would treat as preference.

A workable sequence runs in five steps.

1. Show the full ladder with competitors on shelf. A size judged alone has no reference point.
2. Run the choice task with a no-buy option. Vary the price at each rung and the multipack count.
3. Probe every choice. Occasion, who the pack is for, storage, how long it lasts, what it replaced.
4. Ask which pack is the better deal before revealing unit prices. The gap between perceived and actual value is a finding.
5. Rerun key comparisons as a downsize. Present the smaller pack as a change to one the shopper already buys.

A University of Sydney scoping review of 47 snack and drink downsizing studies found that 12 of 15 studies lowered actual or intended consumption when a single smaller package was offered instead of a larger one, while multipack results were inconsistent. One study it covered found participants perceived the unit price as higher as packages got smaller, a perception no choice model sees on its own.

Trade-off designs fit because stated-preference bias falls hardest on absolute price levels, as the explainer on why trade-off methods beat direct price questions shows. For a credible range on one new pack rather than a ladder, the Van Westendorp price sensitivity questions are the lighter tool. Where a formal conjoint is needed, Alchemic runs it in the same fieldwork as the moderated interviews, so the utility estimate and the shopper's explanation come from one sample.

Who Belongs in the Sample, and Across Which Markets and Channels?

Sample by rung and by channel, not by the average category buyer. A sample weighted to the middle under-reads both ends of the ladder, where most decisions sit.

Set quotas on four things.

• Current pack bought, so every rung has enough of its own buyers.
• Channel: grocery, club, convenience, e-commerce and quick commerce where they matter.
• Lapsed buyers of a downsized or delisted pack, who show what the change cost.
• Unit-price labeling: mandatory states against the rest.

Entry packs exist for people who buy small and often, and in many markets that shopper is easier to reach on a phone than through an emailed survey link. Alchemic fields interviews natively inside WhatsApp with no link and no app, and by outbound AI phone call, and publishes 57+ languages including Hindi, Tamil and Telugu. Recruitment runs as managed fieldwork or bring your own across 14 markets including the USA and the UK, and in India from metros and Tier 1 through Tier 2 and Tier 3.

What Should a PPA Study Deliver?

A rung-by-rung verdict: keep, reprice, resize, drop or add. Each verdict carries the share of choices the rung won, the buyers it recruits, whose purchases it takes and the verbatims behind any misread.

Copyable PPA Research Brief

• Decision: the ladder change to make, and by when.
• Ladder in scope: every current and proposed size, count and format, with its job.
• Competitor set: the packs beside yours in each channel.
• Scanner read: velocity, rung gaps and promotion history, done before fieldwork.
• Hypotheses per rung: what each pack should win and what it may cannibalize.
• Quotas: current pack, channel, lapsed buyers, unit-price labeling.
• Method: choice task with no-buy, probing, perceived value before unit price.
• Outputs: choice shares by rung, source of volume, themes with verbatims.
• Decision rule: the result that keeps, resizes or drops each rung.

Signaling size on shelf is a separate test of label hierarchy and shelf standout, which Alchemic's packaging testing runs with 50 to 200 shoppers on WhatsApp, web or phone, returning themes with cited verbatims.

Where Does PPA Research Mislead?

It misleads when stated choices become volume forecasts, when a test manufactures awareness, or when evidence crosses market types.

• Stated choice is not volume. Choice shares show relative preference and tend to overstate trial of a new size.
• Pointing at the change creates it. GAO notes research suggesting consumers respond less to downsizing than to an equivalent price increase, possibly because subtle pack changes go unnoticed. Ask about the pack unprompted first.
• Retailers are missing. No shopper study predicts a delisting.
• The size evidence is high-income. Every study in the Cochrane review ran in a high-income country, 81 percent in the USA, so its effects do not transfer automatically to sachet-led markets.
• Text and voice cannot put a pack in a hand. Weight, grip and pour need a physical test such as an in-home usage test.