Promotion & Pricing Experiment Tracker

Give the experiment a short, descriptive name (e.g., 'Margherita + Side Bundle — Wed Lunch').

Who is responsible for running and reporting on this experiment (name or role).

Planned start date (YYYY-MM-DD). Keep experiments long enough to hit sample targets and cover typical weekday/weekend patterns.

Planned end date (YYYY-MM-DD). You can update this if you extend the test.

Where the test will run (e.g., single store, store IDs, online ordering only, delivery platforms). Include control locations if using A/B by venue.

State the expected outcome and why you expect it (e.g., 'A $0.50 price increase on side salad will not reduce unit sales but will increase margin per guest').

Choose the main metric you'll use to judge success.

Other metrics to watch (check all that apply).

Describe control (current price/offers) and test (new price/offer) clearly. Example: 'Control: Side salad $3.50. Test: Side salad $3.99'

How customers or locations are assigned to groups.

Enter the target sample size per group (units or guests). If you don't know the exact calculation, aim for a sample that will detect a practical effect (see guidance below).

Choose a significance level or 'practical' when exact stats aren't feasible.

Quantify expected changes for primary and secondary KPIs (e.g., '+2% units, +$0.45 revenue per guest, margin +1.8 percentage points'). Use absolute or percent changes.

Estimate how margins will change if hypothesis is true (positive or negative). Use percentage points or $ per unit.

How often you'll check results (daily/weekly) and key reporting milestones (midpoint review, final analysis date).

Which systems will supply the data (POS, delivery platform, manual counts, surveys). Include any expected manual steps.

Percent change in units sold for the test group vs control (positive or negative). Update after the experiment.

Describe how the menu mix changed (e.g., test increased sales of item A but reduced item B).

Observed change in gross margin (percentage points or $ per guest).

Quality, speed, guest feedback, unintended consequences, labor impacts, or operational issues observed during the test.

Team's confidence that observed result is real and actionable (1 = low, 5 = very high).

Decision based on results and confidence.

Explain the reasons behind the recommended action and any conditions for rollout (e.g., margin thresholds, timing, training).

Who should run the rollout if approved (name or role).

Date the rollout decision was made (YYYY-MM-DD).

Examples: • Small price increase on low‑U item: test in 1 location for 4 weeks; watch units and margin. • Bundling experiment: offer entree + side for $X; watch mix shift and throughput. Sample size guidance: if you lack formal power calculations, aim for several hundred transactions per group for moderate effects; use 'practical' significance when data is limited. Consider extending duration rather than adding new locations to maintain operational consistency.