# Phase 2 Pilot Plan

This is the part of Phase 2 that genuinely can't be done in a chat window —
it requires real households using the live deployment over weeks or months.
What's in this folder is the *instrumentation* for that pilot: the surveys,
the tracking sheet, and a description of what the app already measures
automatically. Run the pilot, then plug the real numbers into the template
in `metrics-tracker-template.csv`.

---

## Groups

| Group | Description | Target size |
|---|---|---|
| A | University students (self-catering) | 100 |
| B | Families | 50 |

## Duration

3–6 months. A minimum viable pilot is 6 weeks (2 weeks baseline / "before",
4 weeks active use / "after") if 3–6 months isn't feasible before a funding
deadline — note this explicitly to evaluators as an abbreviated pilot.

## What to measure

The brief's three headline metrics, each captured **before** participants
start using Meal Mate and again **after** a few weeks of use:

1. **Meal planning time** — minutes/week spent deciding what to cook
2. **Food budget overruns** — % of weeks the household exceeded its food budget
3. **Food wastage frequency** — self-reported, times/week food was thrown out

Use `surveys.md` to collect these (self-report, via Google Forms or
similar) and `metrics-tracker-template.csv` to log them per participant per
week.

## What the app already measures automatically — no survey needed

Two of these don't require asking anyone anything, because the software
already logs the timestamps:

- **Time from household setup to first generated plan** — a live proxy for
  "meal planning time," computed from existing `households.created_at` and
  `mealplans.created_at` timestamps. Surfaced today at
  **Analytics → Population → "Avg. time to first plan."**
- **Budget adherence** — every generated plan's estimated cost vs. the
  household's stated weekly budget is already computed and stored. Pull this
  from `mealplans` + `households` for a real "% of plans over budget" metric
  without surveying anyone.

This means the *planning time* and *budget adherence* metrics can be
reported with real numbers as soon as people start using the deployed app —
only "food wastage frequency" genuinely requires the self-report survey.

## Recruitment notes

- Group A (students): partner with the university's student affairs office
  or a hostel/cafeteria association; offer a small incentive (e.g. airtime)
  for completing both the pre- and post-survey.
- Group B (families): recruit through a community group, church, or estate
  WhatsApp group; in-person onboarding (15 min) significantly improves
  completion of the household setup step.

## Reporting the result

Once both surveys are in and the app has live usage data, the result you're
building toward is a single sentence like the brief's example:

> "Meal Mate was tested on N users, reduced meal planning time by X%,
> improved budget adherence by Y%, and Z% of users reported less food
> wastage."

Fill `metrics-tracker-template.csv`, compute the before/after deltas, and
that sentence — backed by both the survey and the live `analytics/population`
endpoint — is your Phase 2 deliverable for a funding conversation.