| name | next |
| description | Recommend the next highest-impact marketing action based on current phase, funnel metrics, and recent activity. Read-only analysis — identifies the single most impactful thing to work on right now. For cross-domain briefings covering all domains simultaneously, use the COO skill instead.
|
| disable-model-invocation | true |
GrowthOS Next Action
This skill is read-only — it analyzes current state and recommends one action but never writes files.
Cross-domain prioritization: For a briefing that covers marketing, experiment, development, and publishing simultaneously, use the COO skill instead. This skill is marketing-only.
Read Current State
- Read
marketing/MEMORY.md and marketing/STATUS.md
- If
marketing/ does NOT exist: tell the user to run the init skill first and stop
- Resolve app context:
a. Read
marketing/apps/ to list app directories
b. If one app: analyze it
c. If multiple apps: analyze all, recommend for the most critical
Per-App Data
For each app, read:
marketing/apps/{slug}/funnel.md — funnel metrics and conversion rates
marketing/apps/{slug}/reviews.md — review metrics
Also read:
marketing/experiments/backlog.md — experiment status
marketing/positioning/icp.md — check if positioning exists
marketing/outreach/targets.md — check if outreach is set up
marketing/content/strategy.md — check if content strategy exists
marketing/partnerships/pipeline.md — check if partnerships exist
marketing/reviews/strategy.md — check if review strategy exists
Funnel Bottleneck Detection
Read marketing/apps/{slug}/funnel.md and compare each conversion stage against Average benchmarks from references/funnel-benchmarks.md:
| Stage | User Rate | Average Benchmark | Relative Performance |
|---|
| View → Add App | {rate}% | 20% | {rate/benchmark} |
| Add → Install | {rate}% | 65% | {rate/benchmark} |
| Install → Onboard | {rate}% | 50% | {rate/benchmark} |
| Onboard → Activate | {rate}% | 40% | {rate/benchmark} |
| Activate → Paid | {rate}% | 15% | {rate/benchmark} |
| 30-Day Retention | {rate}% | 75% | {rate/benchmark} |
The stage with the lowest relative performance (rate / benchmark average) is the bottleneck.
Staleness Detection
Check each marketing area for last update date. Flag any area not updated in 2+ weeks:
| Area | File Checked | Last Updated | Stale? |
|---|
| Funnel metrics | apps/{slug}/funnel.md | {date} | {yes/no} |
| Experiments | experiments/backlog.md | {date} | {yes/no} |
| Content | content/strategy.md | {date} | {yes/no} |
| Outreach | outreach/targets.md | {date} | {yes/no} |
| Reviews | reviews/strategy.md | {date} | {yes/no} |
| Partnerships | partnerships/pipeline.md | {date} | {yes/no} |
Use last_updated from YAML frontmatter where available, or file modification context.
Phase-Based Priority Weighting
Read current_phase from marketing/STATUS.md and weight priorities accordingly:
- pre-launch: positioning (5x), ASO/listing (4x), outreach (2x), experiments (1x)
- beta: outreach (5x), reviews (4x), ASO (3x), experiments (2x)
- growth: experiments (5x), content (4x), channels (3x), partnerships (2x)
- scale: partnerships (5x), content (4x), experiments (3x), channels (2x)
Single Recommendation
Present exactly ONE action in this format:
Next Action: {action_title}
Why: {1-2 sentence explanation of why this is the highest priority}
Bottleneck/Area: {the funnel stage or marketing area this addresses}
Run: {skill name to execute this action}
Priority Order
Evaluate in this order and recommend the first match:
- No marketing/ directory → "Run the init skill to set up your marketing workspace."
- No positioning → "Start with the positioning skill to define your ICP and value proposition."
- Poor-rated bottleneck → "Your {stage} conversion is at {rate}% (Poor). Run {skill} to address this."
- Stale area (2+ weeks) → "Your {area} hasn't been updated in {days} days. Run {skill} to refresh it."
- Phase-weighted bottleneck → Apply phase weights to the relative performance scores. The stage with the lowest weighted score is the recommendation.
- Phase default → If no clear bottleneck, recommend the default action for the current phase:
- pre-launch: "Refine your positioning and App Store listing."
- beta: "Focus on outreach to get your first installs."
- growth: "Start a new experiment to improve your weakest metric."
- scale: "Explore new partnership opportunities."
Multi-App Context
If multiple apps, analyze all and recommend for the app with the most critical issue. Include the app name in the recommendation:
Next Action for {app_name}: {action_title}