Business & Product

Planning documentation for the Your Solar Guy MVP.

Success Metrics and MVP Acceptance

Purpose

This document defines how the team should measure customer value, product health, and readiness during an MVP pilot. It does not claim achieved results.

Product value funnel

Visit to Manual Submission to Useful Results to Value Action to Return

Traffic alone does not establish success. The useful signal is whether people try the product, reach trustworthy results, take a meaningful action, and return when they need more help.

The funnel connects directly to the KPIs below. Activation measures the first meaningful step, Time to Value covers the path to useful results, and Calendar Conversion measures one important action after a result set contains eligible dated events.

Product KPIs

KPI 1: Activation Rate

Definition: Percentage of eligible workspace visitors who submit at least one supported manual.

Formula: activated workspaces divided by eligible visited workspaces.

Use unique workspace events and exclude obvious internal or invalid activity where possible.

KPI 2: Time to Value

Definition: Time from the first accepted manual submission to the first usable result set.

Report a distribution such as median and upper percentile when the sample is large enough.

KPI 3: Calendar Conversion Rate

Definition: Percentage of result sets with eligible dated events that complete an authorized ICS download or accepted calendar email delivery.

Provider acceptance confirms the send request. It does not confirm inbox receipt or calendar import.

KPI 4: Task Success Rate

Definition: Percentage of observed user tasks completed without assistance in a structured usability study.

Tasks may include submitting a manual, finding an important action, verifying a source, downloading a calendar, or asking a grounded question.

Product health guardrails

Processing Failure Rate

Measure submitted processing attempts that end without usable results. Separate invalid user input, application errors, quota pauses, and provider failures.

Validation Pass Rate

Measure reviewed artifacts that pass QA/QC without correction. Report the artifact type so extraction, results, email, and calendar quality are not mixed.

Measurement integrity

Define each event once.
Deduplicate by workspace, result set, or artifact as appropriate.
Keep operational retries out of user funnel counts.
Mark unavailable denominators clearly.
Avoid trends until both periods contain enough observations.
Do not reconstruct claims from incomplete records.

Internal metrics view

An operator-only view may show four product KPIs, two health guardrails, and the result-set funnel. It must protect workspace identity and document content.

MVP acceptance criteria

Gate A: Safety and integrity

Private documents remain isolated by workspace.
Secrets and private links stay out of logs and commits.
Dates, access, approval, and ICS output use deterministic validation.
Untrusted content cannot grant authority.

Gate B: Functional completion

Web and email intake complete their expected journeys.
Saved results, private links, calendar actions, and Q&A work together.
Temporary failures preserve recoverable state.

Gate C: Extraction and evidence quality

Every page receives coverage or a visible failure state.
Critical actions retain source evidence.
Missing timing stays unresolved.
A reviewed public or synthetic corpus supports accuracy assessment.

Gate D: Calendar and validation quality

Eligible dates match deterministic calculations.
Invalid or unsupported dates do not enter calendar output.
QA/QC correction and manual-review rules behave as specified.

Gate E: User value

A pilot should show that users can identify an important action, understand its timing or uncertainty, verify the source, and complete a useful next step.

Gate F: Operational readiness

Provider authorization is valid.
Webhooks and scheduled workers run in the target environment.
Retention cleanup is observed.
Error diagnostics support stage-level investigation.

Data limitations

The 15-day retention period limits historical analysis. Small samples may not support trends. Provider acceptance does not prove delivery or calendar import. User-reported usefulness does not replace accuracy review.

Claims policy

Do not claim validated demand, reduced missed maintenance, broad manufacturer accuracy, long-term retention, willingness to pay, or large-scale throughput without direct evidence.