AgentSkillsCN

funnel-analytics

全面梳理并衡量 Hedge Edge 用户从首次接触直至留存与推荐的完整旅程。在漏斗的每个阶段追踪转化率、流失节点以及收入影响:内容获取、注意力吸引、客户捕获、销售成交、交付使用、用户留存、数据分析。整合 GA4、Vercel Analytics、Supabase、Creem.io、Discord 以及 Google Sheets 的数据,精准定位最具杠杆效应的优化点。

SKILL.md
--- frontmatter
name: funnel-analytics
description: |
  Map and measure the complete Hedge Edge user journey from first touch through retention and referral. Track conversion rates, drop-off points, and revenue impact at every funnel stage: Content  Attention  Capture  Sales  Delivery  Retention  Analytics. Integrate data from GA4, Vercel Analytics, Supabase, Creem.io, Discord, and Google Sheets to identify the highest-leverage optimization points.

Funnel Analytics

Objective

Identify exactly where Hedge Edge is losing potential revenue in the user journey and quantify the dollar impact of fixing each bottleneck. The funnel is not a static diagram it is a living, measured system where every stage has a conversion rate, a benchmark, a trend, and a recommended action. The goal is to make funnel optimization the primary growth lever by turning qualitative hunches into quantitative priorities.

When to Use This Skill

  • Weekly funnel review: Standard weekly analysis of full-funnel conversion rates and WoW changes
  • Campaign launch: When Marketing Agent launches a new campaign and needs baseline + tracking setup
  • Conversion drop detected: When KPI Dashboards flag a conversion rate anomaly at any funnel stage
  • Channel comparison: When evaluating which acquisition channel (YouTube, Discord, paid, IB) has the best funnel performance
  • Landing page changes: When Developer Agent or Content Creator Agent pushes landing page updates and needs impact measurement
  • Growth planning: When Business Strategist Agent needs funnel data to model growth scenarios
  • New funnel stage added: When a new touchpoint is introduced (e.g., webinar funnel, MT4 launch)

Input Specification

Required Inputs

FieldSourceDescription
date_rangeRequest parameterAnalysis period with comparison period
ga4_traffic_dataGA4 APISessions, users, source/medium, landing pages, UTM parameters, conversion events
ercel_page_dataVercel AnalyticsPage views, unique visitors, bounce rate by page, referrer data
supabase_signupsSupabase uth.usersSignup events with created_at, source, utm_source, utm_medium, utm_campaign
supabase_trialsSupabase subscriptionsTrial start events with user_id, plan, started_at
creem_conversionsCreem.io APITrial-to-paid conversions, plan selections, payment timestamps
creem_retentionCreem.io APIRenewal events, cancellation events, upgrade/downgrade events

Optional Inputs

FieldSourceDescription
discord_signupsDiscord Bot + Supabase joinUsers who joined Discord before/after signup (community influence measurement)
email_eventsn8n/email platformEmail sends, opens, clicks by campaign (for email funnel stages)
ib_activationsGoogle Sheets CRMIB broker activation events mapped to user journey
support_ticketsSupabase/Google SheetsSupport interactions mapped to funnel stage (pre-sale vs. post-sale)
segment_filterRequest parameterFilter by channel, plan, prop firm, geography, device

Step-by-Step Process

Step 1: Define the Funnel Stages

Map the Hedge Edge funnel with measurable events at each transition:

` CONTENT (Awareness) Event: Video view, blog visit, social impression Metric: Impressions, watch time, profile visits Source: YouTube Analytics, social platforms

ATTENTION (Interest) Event: Click to landing page, ad click, email open Metric: CTR, CPC, engagement rate, open rate Source: GA4, Vercel, email platform

CAPTURE (Lead) Event: Landing page signup, lead magnet download, Discord join Metric: Lead capture rate, leads per channel Source: Vercel + Supabase (signup event)

SALES (Conversion) Event: Trial start, trial-to-paid, plan selection Metric: Trial start rate, close rate, avg plan value Source: Supabase + Creem.io

DELIVERY (Activation) Event: First hedge executed, MT5 EA connected, 3+ accounts added Metric: Time-to-first-value, activation rate, support tickets Source: Supabase usage logs

RETENTION (Loyalty) Event: 30-day renewal, 60-day renewal, upgrade, referral sent Metric: 30/60/90-day retention, NRR, referral rate Source: Creem.io + Supabase

EXPANSION (Advocacy) Event: Referral conversion, IB activation, plan upgrade, review posted Metric: Viral coefficient, IB activation rate, expansion revenue Source: Supabase + Google Sheets CRM `

Step 2: Data Collection & Event Mapping

  1. Pull GA4 session data with UTM parameters to attribute traffic to channels
  2. Pull Vercel Analytics for landing page behavior (pageviews, scroll depth, CTA clicks)
  3. Query Supabase for signup events, matching utm_source to GA4 sessions where possible
  4. Pull Creem.io for trial starts, conversions, and renewals
  5. Query Supabase usage logs for activation events (first hedge, EA connection, account additions)
  6. Map each user to their funnel journey: assign timestamps for each stage transition
  7. Handle multi-session attribution: if a user visits 3 times before signing up, capture all touchpoints

Step 3: Conversion Rate Calculation

For each stage transition, calculate:

  1. Absolute conversion rate: Users who completed stage N+1 / Users who entered stage N
  2. Cumulative conversion rate: Users at stage N / Total users at top of funnel
  3. Time-to-convert: Median and P90 time between stages (e.g., signup to trial start: median 0.5 days, P90 3 days)
  4. Drop-off count: Absolute number of users lost at each stage
  5. Revenue impact of drop-off: Lost users average LTV of converted users = revenue left on table

Example calculation for Hedge Edge: ` Landing Page Visitors: 10,000/mo Signups: 800 (8.0% capture rate) Trial Starts: 560 (70% of signups) Trial-to-Paid: 168 (30% close rate) 30-Day Retained: 135 (80% retention) 90-Day Retained: 101 (60% of original paid) IB Activated: 42 (25% of paid users)

Revenue per converted user: ARPU 8 mo avg tenure = LTV

  • IB value: 25% /mo IB commission 8 mo = IB LTV contribution Blended LTV: ~

Revenue impact of +1% signup rate improvement: 100 additional signups 70% trial 30% convert LTV = ,862/year `

Step 4: Channel-Specific Funnel Analysis

Build separate funnels for each acquisition channel:

  1. YouTube Organic: Video view Profile visit Landing page click Signup Trial Paid
    • Key metric: Content-to-signup rate, video-assisted conversion rate
  2. Discord Community: Discord join Engagement (5+ messages) Landing page visit Signup Trial Paid
    • Key metric: Community-influenced conversion rate, time-in-community before conversion
  3. Paid Ads: Ad impression Click Landing page Signup Trial Paid
    • Key metric: CAC (fully loaded), ROAS, payback period
  4. IB Partner Referral: Partner link click Broker signup Hedge Edge signup Trial Paid
    • Key metric: Partner-attributed revenue, co-registration rate
  5. Organic Search: Search impression Click Landing page Signup Trial Paid
    • Key metric: Keyword conversion rates, SEO ROI
  6. Email: Email received Opened Clicked Landing page Signup/Trial/Upgrade
    • Key metric: Email-attributed conversions, nurture sequence effectiveness
  7. Word of Mouth/Referral: Referral link shared Click Signup Trial Paid
    • Key metric: Viral coefficient, referral conversion premium vs. organic

Step 5: Drop-Off Analysis & Bottleneck Identification

  1. Rank all stage transitions by drop-off rate (worst conversion % first)
  2. For the top 3 bottlenecks, investigate:
    • Who drops off: Segment by channel, device, plan interest, geography
    • When they drop off: Time-based patterns (day of week, time of day, days since previous stage)
    • Where they drop off: Specific pages, form fields, onboarding steps
    • Behavioral signals: What did drop-offs do differently than converters? (e.g., watched demo video vs. didn't, joined Discord vs. didn't)
  3. Calculate the revenue impact of improving each bottleneck by 10%, 25%, 50%
  4. Rank bottlenecks by: Revenue Impact Confidence of Improvement Implementation Ease

Step 6: Funnel Velocity Analysis

  1. Calculate time-to-convert at each stage (median, P25, P75, P90)
  2. Identify slow stages where users stall (e.g., signup-to-trial >3 days = risk of drop-off)
  3. Correlate speed with conversion: do faster-moving users have higher LTV?
  4. Identify "fast path" users (signup to paid in <24 hours) and analyze what they have in common
  5. Set velocity benchmarks and flag users who are falling behind (trigger nurture automation)

Step 7: Output Generation

  1. Build funnel visualization with conversion rates at each stage (Sankey-style data for rendering)
  2. Write stage-by-stage data to Google Sheets with conditional formatting
  3. Generate Notion report with:
    • Funnel diagram with current rates and MoM changes
    • Top 3 bottlenecks with revenue impact quantification
    • Channel comparison table
    • Velocity analysis with benchmark compliance
    • Recommended actions with ICE scores
  4. If a conversion rate drops >15% WoW, trigger n8n alert

Output Specification

Funnel Report Structure

`

Funnel Analytics Report [Date Range]

Executive Summary

  • Total funnel throughput: [X] visitors [Y] paid users ([Z]% end-to-end)
  • MoM change in end-to-end conversion: [+/- %]
  • Biggest bottleneck: [Stage] at [X]% conversion (target: [Y]%)
  • Estimated revenue impact of fixing top bottleneck: $[amount]/month

Stage-by-Stage Breakdown

StageVolumeConversion RateWoW ΔMoM ΔTargetStatus
Landing Page Visitors10,000+5%+12%
Signups8008.0%-0.5%+1.2%8%
Trial Starts56070.0%+2%+5%65%
Trial-to-Paid16830.0%-3%-1%35%
30-Day Retained13580.4%+1%0%80%
IB Activated4225.0%+4%+8%30%

Channel Comparison

ChannelVisitorsEnd-to-End CVRCACLTVCAC:LTVPayback
YouTube Organic4,2002.1%1:550.4 mo
Discord1,8003.5%1:1700.2 mo
Paid (Google)2,5001.2%1:93.2 mo
Organic Search1,2001.8%1:800.5 mo
Referral3005.2%0 mo

Bottleneck Analysis

Bottleneck 1: Trial-to-Paid (30% vs. 35% target)

  • Revenue at stake: $[X]/month
  • Root cause hypothesis: [analysis]
  • Segment most affected: [detail]
  • Recommended action: [specific action]
  • ICE Score: [Impact Confidence Ease]

Funnel Velocity

TransitionMedian TimeP90 TimeBenchmarkStatus
Visit Signup2.3 sessions7 sessions3 sessions
Signup Trial0.5 days3.2 days1 day
Trial Paid6.8 days13 days7 days

Recommended Actions (ICE-ranked)

  1. [Action] Impact: [H/M/L], Confidence: [H/M/L], Ease: [H/M/L] Score: [X/30]
  2. ... `

API & Platform Requirements

PlatformEndpoint/MethodAuthPurpose
GA4Data API v1
unReportGA4_MEASUREMENT_ID + GA4_API_SECRETTraffic source, UTM attribution, session data
Vercel Analytics/v1/analyticsVERCEL_ANALYTICS_TOKENLanding page behavior, CTA click tracking
Supabase/rest/v1/users, /rest/v1/eventsSUPABASE_URL + SUPABASE_KEYSignup events, activation events, usage logs
Creem.io/v1/subscriptions, /v1/eventsCREEM_API_KEYTrial, conversion, renewal, churn events
Google SheetsSheets API v4GOOGLE_SHEETS_API_KEY + service accountDashboard output, CRM data for IB funnel
Notion/v1/pagesNOTION_API_KEYReport storage and distribution
Discord Bot/guilds/{id}/members, /channels/{id}/messagesDISCORD_BOT_TOKENCommunity engagement funnel data
n8nPOST to N8N_WEBHOOK_URLWebhook URLAlert triggers for conversion anomalies

Quality Checks

  • Full funnel coverage: Every defined stage has a measured conversion rate with sample size shown
  • Attribution integrity: >90% of signups have a tracked source (UTM or referrer). Unattributed <10%.
  • No double-counting: Each user counted once per stage, even with multiple sessions
  • Conversion windows: Time windows are defined and consistent (e.g., 14-day trial window, 30-day attribution window)
  • Statistical validity: Conversion rate changes flagged only when sample size >100 and change is >2 percentage points or statistically significant at p<0.05
  • Channel isolation: Channel-specific funnels do not overlap (multi-touch handled in Attribution Modeling skill, not here)
  • Revenue impact calculated: Every bottleneck has a dollar-value impact estimate, not just a percentage
  • Velocity benchmarks set: Each stage transition has a target time-to-convert with actual vs. target comparison
  • Comparison context: Every metric shows WoW and MoM change plus target benchmark
  • Actionable output: Report includes at least 3 ICE-scored recommended actions tied to specific funnel stages