SKILL.md
Market Funnel — Full Journey Diagnosis
Model the full funnel, identify where prospects drop off, and hand back an analytics spec the client's engineer can implement directly.
The 6 Funnel Stages
| Stage |
Goal |
Key metric |
| 1. Awareness |
Get in front of the right audience |
Impressions, reach, branded search |
| 2. Acquisition |
Earn the click |
CTR, referral traffic, CAC |
| 3. Activation |
First value moment |
Signup rate, first-action rate |
| 4. Revenue |
First purchase or trial conversion |
Trial→paid, ARPU |
| 5. Retention |
Continued use |
DAU/MAU, churn, repeat rate |
| 6. Referral |
Organic growth |
NPS, referral rate, virality coefficient |
Pirate metrics (AARRR) plus Awareness, adapted for 2025 realities.
Workflow
- Map the existing funnel based on what's visible: the home page, the pricing page, the signup flow, the app/experience.
- Score each stage 0–10 using the rubrics below.
- Name the leak — the stage with the steepest drop-off relative to benchmark.
- Propose fixes per stage with expected impact.
- Write the GA4 event spec — every event name, params, when it fires.
Stage Rubrics
1. Awareness (0–10)
- Channels in use (organic, paid, social, referral, partnerships)
- Branded vs non-branded traffic mix
- Share of voice vs competitors (ballpark from search data)
- Content cadence
2. Acquisition (0–10)
- Landing page → form view rate
- Page load speed
- Hero-to-CTA distance
- Ad / organic message match
3. Activation (0–10)
- Signup friction (fields, steps, time)
- First value moment — is there a "wow" before the user has to commit?
- Empty state quality
- Onboarding sequence present?
4. Revenue (0–10)
- Trial length and gating
- Pricing page clarity
- Checkout friction
- Payment options
- Refund/cancel policy transparency
5. Retention (0–10)
- Product usage signal (if visible)
- Lifecycle emails
- Re-engagement paths
- Churn recovery
6. Referral (0–10)
- Explicit referral program
- Shareable moments built in
- Testimonial capture
- UGC encouragement
Benchmarks (use these, state they're rough)
| Stage |
SaaS |
E-com |
Agency |
| Visitor → signup |
2–5% |
— |
— |
| Visitor → purchase |
— |
1.5–3.5% |
— |
| Visitor → booked call |
— |
— |
2–5% |
| Trial → paid |
15–25% |
— |
— |
| Cart → checkout |
— |
40–70% |
— |
| Checkout → purchase |
— |
65–85% |
— |
| Month 1 retention |
70%+ |
15–25% (repeat) |
— |
GA4 Event Spec (deliverable)
Produce a table the client's engineer can implement in one sitting:
| Event name |
When it fires |
Params |
Stage |
page_view |
Every page load |
pagelocation, pagetitle, page_referrer |
Awareness |
scroll_50 |
Scroll depth 50% |
page_location |
Awareness |
cta_click |
Any primary CTA click |
ctatext, ctalocation, destination |
Acquisition |
sign_up |
Account creation success |
method, plan |
Activation |
onboarding_step |
Each onboarding step |
stepnumber, stepname |
Activation |
first_value |
Key activation moment (client-specific) |
feature |
Activation |
begin_checkout |
Checkout loaded |
value, currency, items |
Revenue |
purchase |
Payment success |
transaction_id, value, currency, items |
Revenue |
subscription_started |
First recurring charge |
plan, value, billing_period |
Revenue |
feature_used |
Core feature event (client-specific) |
feature, count |
Retention |
churn_risk |
Inactivity 14d+ |
last_active |
Retention |
referral_sent |
Share action |
channel |
Referral |
Adjust events to the specific product. Always include the stage mapping so the client can build funnel reports in GA4 directly.
Output: FUNNEL-ANALYSIS.md
# Funnel Analysis — [Brand]
**URL:** [url] | **Date:** [YYYY-MM-DD]
## Funnel Score: [X]/100
[1-paragraph diagnosis: where the funnel leaks worst and why.]
## Stage Scores
| Stage | Score | Status | Top Issue |
|---|---|---|---|
## Stage-by-Stage Analysis
### 1. Awareness — X/10
[Findings, evidence, fixes]
... repeat for all 6 stages ...
## The Leak
[Name the #1 drop-off point. Estimate its cost. Explain why.]
## Fixes by Stage
### Quick wins (this week)
### High-leverage (this month)
### Strategic (this quarter)
## GA4 Event Spec
[Full table as above]
## Reporting
[3 GA4 exploration reports the client should build: funnel exploration, path exploration, retention.]
Quality Bar
- Be honest about what you can and can't see. The homepage shows you stages 1–3 clearly; 4–6 require the client's analytics data. If you can't see it, flag it.
- The GA4 spec is the most actionable part — it should be copy-paste-ready for a developer.
- Don't invent conversion rates. If you don't know actual CR, say "assume 3% baseline" and compute impact from there.