# Product Requirements Document

## 1. Document Control
- **PRD Title:** Invite-Only Referral Experiment for Self-Serve Teams
- **Owner:** Marco (Growth PM)
- **Date:** 2026-04-14
- **Status:** Draft
- **Review Date:** 2026-04-15

## 2. Context and Problem Statement
### Background
Referral-attributed signups have declined from 21.4% (week of 2026-03-02) to 17.1% (week of 2026-04-06), representing a 20% drop in referral conversion efficiency. Simultaneously, paid acquisition CAC is rising, creating urgency for a lower-cost growth lever. Current referral mechanics allow any user to generate referral links at any point in their journey, resulting in low-quality invites sent before users understand product value or complete team setup.

### Problem
The existing referral flow suffers from three core issues:
1. **Broad, untargeted access**: Users can generate referral links before reaching team setup completion, leading to low-intent invites.
2. **Banner blindness**: Referral CTAs appear too broadly across the app and are being ignored.
3. **Poor conversion**: 7-day invite acceptance rates dropped after recent onboarding email copy changes, and referral attribution breaks when invitees switch devices.

Launch review data shows teams with 3+ active users within the first 7 days are most likely to successfully invite others, but the current flow does not target this high-intent segment.

## 3. Goals, Non-Goals, and Success Metrics
### Primary Goal
Increase referral-attributed signups by 25% within 8 weeks of full rollout (from baseline of ~33 weekly signups to ~41 weekly signups).

### KPI Targets
- **Invite acceptance rate (7-day)**: Increase from 17.1% to 22% by targeting engaged team admins only.
- **Invite quality**: 60%+ of invites sent by eligible admins result in signup completion (vs. current ~10% overall).
- **Eligible admin activation**: 50%+ of teams reaching 3+ users in first 7 days send at least one invite within 14 days.

### Non-Goals
- Pricing or packaging changes tied to referral incentives.
- Sales-assisted or enterprise referral motions.
- Monetary incentives (requires legal review; out of scope for this experiment).
- Redesigning the entire onboarding flow beyond referral entry points.

## 4. Users and Hypothesis
### Primary User Segment
Self-serve team admins who have reached 3+ active users within their first 7 days. This segment represents ~43% of self-serve team signups (84 of 193 weekly teams) and demonstrates early product-market fit.

### Jobs To Be Done
- As a team admin with an engaged team, I want to invite trusted colleagues so we can collaborate more effectively.
- As a team admin, I want to share a product I've validated so my referrals reflect well on me.

### Experiment Hypothesis
If we gate referral access to team admins who have reached 3+ active users within 7 days, then we will increase invite quality and conversion rates because these users have experienced product value and will send invites to higher-intent contacts.

## 5. Scope
### In Scope (MVP)
- Eligibility gating: Only show referral CTA to team admins with 3+ active users in first 7 days.
- Contextual CTA placement: Surface referral invite at team milestone moment (when 3rd user joins).
- Improved invite messaging: Update email copy to emphasize value proposition from referrer's perspective.
- Event instrumentation: Track "eligible admin identified," "invite sent by eligible admin," "invite accepted within 7 days."

### Out of Scope
- Referral incentives (credits, discounts, monetary rewards).
- Multi-channel invite distribution (SMS, Slack, etc.; email-only for MVP).
- Referral dashboard or leaderboard UI.
- Attribution fixes for cross-device signup flows (separate engineering effort).
- Changes to pricing, packaging, or sales-assisted onboarding.

## 6. Requirements
### Functional Requirements
1. **Eligibility Detection**: System identifies team admins who reach 3+ active users within 7 days of team creation. Active user = logged in and performed ≥1 core action (e.g., created workspace, invited teammate, completed onboarding task).
2. **Gated CTA**: Referral invite CTA only appears for eligible admins. Non-eligible users do not see referral options in navigation, settings, or onboarding.
3. **Milestone Trigger**: When a team reaches 3 active users, show in-app modal to admin: "Your team is growing! Invite colleagues who'd benefit from [Product]." Include one-click invite flow.
4. **Invite Email Update**: Revise invite email copy to lead with referrer's name and team context: "[Referrer Name] invited you to join their team on [Product]. They're using it to [value prop]."
5. **Invite Cap (Optional Safety)**: Limit eligible admins to 10 invites per 7-day period to prevent abuse during experiment.

### Analytics and Instrumentation Requirements
1. **Event: `referral_eligibility_reached`**: Fired when team admin reaches 3+ active users within 7 days. Properties: `team_id`, `admin_user_id`, `days_to_eligibility`, `active_user_count`.
2. **Event: `referral_invite_sent_eligible_admin`**: Fired when eligible admin sends invite. Properties: `team_id`, `admin_user_id`, `invite_count`, `invite_channel` (email).
3. **Event: `referral_invite_accepted_d7`**: Fired when invitee completes signup within 7 days. Properties: `referrer_user_id`, `invitee_user_id`, `days_to_acceptance`, `attribution_source`.

### Reporting Cadence
Weekly dashboard review with Growth and Data teams. Key metrics: eligible admins identified, invites sent, 7-day acceptance rate, referral-attributed signups. First review 7 days after limited beta launch.

## 7. Launch Plan
### Rollout Phases
1. **Internal Dogfood (Week 1)**: Enable for internal teams only. Validate eligibility logic, CTA placement, and event tracking. Target: 5-10 internal teams reach eligibility.
2. **Limited Beta (Weeks 2-4)**: Roll out to 20% of newly eligible admins. Monitor invite quality, acceptance rates, and user feedback. A/B test: 20% treatment (invite-only) vs. 80% control (existing flow).
3. **Scale Decision (Week 5)**: If limited beta hits success criteria, scale to 100% of eligible admins. If stop/fix criteria triggered, pause and iterate.

### Launch Success Criteria
- 7-day invite acceptance rate ≥20% in limited beta (vs. 17.1% baseline).
- 50%+ of eligible admins send at least one invite within 14 days of reaching eligibility.
- No increase in support tickets related to referral confusion or access issues.

### Stop/Fix Criteria
- 7-day invite acceptance rate drops below 15% (worse than baseline).
- Eligible admin invite send rate <25% (low engagement with new flow).
- Support ticket volume increases >20% related to "can't find referral link" or similar confusion.

## 8. Risks and Mitigations
| Risk | Impact | Mitigation | Owner |
|---|---|---|---|
| Abuse/fraud: Eligible admins create fake teams to gain referral access | High - could inflate metrics and waste resources | Implement invite cap (10 per 7 days) and monitor for anomalous patterns (e.g., same IP, rapid team creation) | Data + Eng |
| User confusion: Non-eligible users expect referral access and complain | Medium - support burden, negative sentiment | Add help center article explaining eligibility; monitor support tickets for trends | PMM + Support |
| Funnel cannibalization: Gating reduces total invite volume more than it improves quality | High - could decrease overall referral signups | A/B test in limited beta to measure net impact on referral-attributed signups | Growth PM |
| Attribution breakage: Cross-device signup flows lose referral source | Medium - undercounts success | Document known limitation; prioritize attribution fix in Q3 roadmap | Eng + Data |

## 9. Dependencies
- **Engineering**: 2-week sprint to build eligibility detection, gated CTA logic, and event instrumentation.
- **Data**: Event schema updates and dashboard setup (1 week).
- **Lifecycle/PMM**: Revised invite email copy and help center article (3 days).
- **Legal**: Confirm no legal review needed for non-monetary referral experiment (1 day).

## 10. Open Questions
1. Should we allow team members (non-admins) to send invites if they're part of an eligible team, or restrict to admins only?
2. What happens to users who were previously shown referral CTAs but are now ineligible? Do we grandfather them in or remove access?
3. If a team drops below 3 active users after reaching eligibility, do they lose referral access?
4. Should we test alternative eligibility criteria (e.g., 5+ users, or 7-day retention threshold) in a follow-up experiment?

## 11. Decision Log
- **2026-04-14**: Decided to gate referral access to 3+ active user milestone based on launch review findings showing this segment has highest invite success rates.
- **2026-04-14**: Decided to exclude monetary incentives from MVP to avoid legal review delays and keep experiment launch timeline within one quarter.
- **2026-04-14**: Decided to use A/B test in limited beta (20% treatment) to measure net impact on referral signups and avoid full-scale risk.