# Breakfast grains brand tracker

A fictional manufacturer wants to distinguish lack of awareness from weak trial and weak repeat buying before choosing between reach, retail availability and product improvement. The study measures category behavior and separate stages of the brand funnel; it does not infer causal advertising effects from cross-sectional answers.

Interview adults who share responsibility for household grocery buying and bought breakfast grains in the past four weeks. Use an online opt-in sample, targeting 600 completed interviews in each wave. Set host-managed marginal targets for age band and broad settlement type; do not exclude low-awareness buyers merely to improve brand-base sizes. Keep recruitment source, field dates and questionnaire version in host metadata. If brand-user bases are small, report base sizes and uncertainty rather than treating them as population estimates.

The intended median interview length is 10–12 minutes including reflection and repeated brand follow-ups; this is a planning estimate requiring a timed pilot. Use the same category definition, four-week reference period and brand codes across waves. Include buyers for others: shared buying need not be equal, personal consumption may be Never, and brand or importance ratings may be Cannot judge. Show those responses separately from low ratings. Pretest whether breakfast grains excludes baked bread and whether respondents can distinguish brand names. Report awareness among all eligible buyers, recent purchase among eligible buyers and conversion among aware buyers separately. Treat Other verbatims as uncoded until reviewed; never silently merge them with named brands.

The fictional market uses pounds (£). All brands, places, respondent scenarios and study material are invented. The source scripts collect no names, addresses, contact details or payment information. Consent is a research participation decision; the hosting organization supplies its actual privacy notice, retention policy and withdrawal process before fieldwork.

Files: `SPEC.md` is the research contract; `study.qube` and `study.odin` are independent implementations; `fixtures.json` contains synthetic response scenarios. Run `node tools/test_fictional_consumer_studies.mjs` from the repository root for scoped offline parser and interview checks.

Quota admission, invitations, permissions, duplicate detection and verified cross-case counts belong to the host. These scripts request no quota service, payment or incentive. The supplied offline fixtures establish local questionnaire behavior only. Browser/mobile layout, accessibility, timed length, real recruitment and statistical quality need a pilot. Different engines may display different seeded brand orders; stable codes identify answers. Legacy DLL oracle approval and human readability/marketing approval are separate from local checks.