# Improvement Roadmap

Improve [PROJECT NAME] from observed behavior and support evidence. Do not turn every request into a feature; first determine which outcome is blocked and whether a smaller change solves it.

[Kit README](README.md) · [Project Overview](Project%20Overview.md) · [PRD](Full%20PRD%20Template.md) · [Master Build Prompt](Master%20Build%20Prompt.md) · [Completion Checklist](Project%20Completion%20Checklist.md)

## Evidence Sources

| Source | Question | Cadence | Owner | Privacy Rule |
| --- | --- | --- | --- | --- |
| Product metrics | Where does the core flow fail? | [WEEKLY] | [OWNER] | Collect minimum approved events |
| Interviews/usability | Why does behavior occur? | [CADENCE] | [OWNER] | Consent and note retention |
| Support/feedback | Which problems repeat or block trust? | [CADENCE] | [OWNER] | Remove unnecessary personal data |
| Reliability/cost | Which failures threaten operation? | [DAILY/WEEKLY] | [OWNER] | Redacted telemetry |

Track these domain measures: onboarding completion, core-action completion, crash-free sessions, sync failure rate.

## Improvement Backlog

| Opportunity | Evidence | User/Outcome | Proposed Change | Expected Signal | Effort | Risk | Decision |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [PROBLEM, NOT FEATURE] | [SOURCE] | [USER / OUTCOME] | [SMALLEST CHANGE] | [TARGET] | [S/M/L] | [L/M/H] | [TEST/BUILD/DECLINE] |

## Prioritization Rules

1. Fix security, privacy, data integrity, misleading behavior, and P0 reliability defects first.
2. Remove unnecessary steps or confusion before adding automation.
3. Improve the weakest core-flow metric with the smallest testable change.
4. Invest in performance, supportability, and cost when they constrain growth.
5. Reconsider deferred items—widgets, wearable support, advanced background sync, deep links, tablet-specific layouts—only when evidence shows a target-user outcome cannot be reached without them.

## Experiment Card

~~~text
Problem/evidence: [WHAT WE OBSERVED]
Hypothesis: If we [CHANGE], then [USER] will [BEHAVIOR], measured by [METRIC], because [REASON].
Guardrails: [QUALITY/SAFETY/COST METRICS]
Audience/duration: [SCOPE]
Pass/fail rule: [THRESHOLD]
Rollback: [HOW]
Result: [DATA]
Decision: [ADOPT / ITERATE / REJECT]
~~~

## Suggested Horizons

- **Now — Reliability and clarity:** address asking for permissions too early and data conflicts after offline use, then improve the current P0 flow.
- **Next — Proven depth:** add the highest-evidence workflow improvement for existing target users.
- **Later — Expansion:** consider widgets, wearable support, advanced background sync only with capacity and a new validation decision.

## Review Gate

- [ ] The backlog is phrased as problems and evidence.
- [ ] Every selected change has a metric and guardrail.
- [ ] The roadmap includes maintenance, accessibility, security, and cost—not only features.
- [ ] Removed or rejected ideas retain decision reasons.
- [ ] The next version remains small enough to test end to end.