Occasion-Based Consumption Study Framework

Example output

Occasion-based consumption study framework — Meadowlight RTD iced tea

Occasion hypotheses worth testing
Your three suspected occasions are a reasonable starting frame, but a diary study should test them as hypotheses, not assume they're the full picture. Recommend going in with:
1. Desk hydration — routine, low-involvement, likely driven by habit and availability rather than flavor choice in the moment.
2. Lunch pairing — meal-adjacent, likely higher flavor and portion consideration, possibly paired with food choice.
3. Afternoon treat — reward-framed, likely higher emotional engagement and more willing to trade up in flavor or format.
4. A fourth, unnamed hypothesis to leave room for: an occasion you haven't named yet. Diary studies routinely surface an occasion the brief didn't anticipate — build in an open "what were you doing" field rather than only pre-coded options, or you'll never see it.

Recommended methodology
A two-week mobile diary study fits your goal better than a single retrospective survey. Occasion research is notoriously unreliable when reconstructed from memory days later — people misremember context, and "why did you drink this" answered after the fact tends to rationalize rather than report. In-the-moment or same-day logging captures the real trigger. At 150 participants over two weeks, expect roughly 4-8 logged occasions per participant if RTD tea is a regular habit, giving you 600-1,200 total occasion entries — treat that estimate as a planning assumption to validate against your actual incidence rate.

Occasion-differentiating questions for the instrument
For each logged occasion, capture: what were you doing right before this (activity context), who else was present, where you were, what (if anything) you were also consuming, how the drink was acquired (had on hand vs. just bought), and the one-word feeling driving the choice (routine, refresh, treat, thirst). These six fields are what let you cluster logged moments into occasions after the fact, rather than forcing participants into your three pre-named buckets in the moment, which biases the read toward confirming what you already suspected.

Sample and fielding approach
150 participants is workable for a directional occasion map but thin for statistically separating six or more distinct occasion clusters if they emerge. Recruit for known RTD tea buyers, not general beverage buyers, and stratify recruitment loosely across your three hypothesized occasion types if you have any purchase-channel signal (e.g. workplace vending vs. grocery) to avoid the sample skewing toward one occasion by accident.

Analysis plan
1. Code every logged entry against the six context fields before looking at brand or flavor choice — occasion structure should emerge from behavior, not be assumed from your hypothesis list.
2. Cluster entries into occasion groups based on shared context patterns, then check how well your original three hypotheses map onto the clusters that actually emerge.
3. For each confirmed occasion, profile the flavor, format, and trigger differences — this is the output that should feed back into innovation and marketing.
4. Flag any cluster with fewer than 10% of total entries as directional only, not a stable occasion to build a strategy around.

Assumptions to validate
- That 150 participants yields enough logged occasions per hypothesized moment to analyze separately — a lower incidence rate than expected may require extending the field period.
- That desk hydration, lunch pairing, and afternoon treat are genuinely distinct in behavior and not just distinct in how your team talks about the product internally.

Want me to draft the actual diary-entry question set, formatted for a mobile survey tool?