Methodology and reading notes

• Method: online interviewing among category users in UAE. Wave samples are independent.
• Wave-on-wave movements are tested with a two-proportion z-test at 95% confidence, on the smaller of the two wave bases.
• Bases below n=100 are flagged indicative; below n=50 values are withheld and read "Not available" rather than being charted.
• KPI direction is applied throughout: a fall in churn intention reads as an improvement, and being below the competitive-set average on it is favourable.
• Differences between brands are numerical positions within the selected set. No paired respondent-level test was run, so no brand difference is described as significant.
• Competitive-set average: simple average across the tracked brands — Kalha Coffee, Bean Souq, Roastery 21; client brand included; long tail excluded: Other.
• Funnel and conversion figures are supplied on the Total base only and are not re-based when another segment is selected.
• Where only one tracked brand carries a reportable value for a KPI, the competitive-set average would restate that brand, so the benchmark comparison is withheld rather than reported as a zero gap.
• Comparability break — Prompted awareness: Three new chains were added to the prompted list at W3. No change, interval, p-value, effect size or significance label is reported across it.
• Limitation: Prompted awareness is not comparable across W2–W3: Three new chains were added to the prompted list at W3