Growreach Posthog Audit and Implementation Guide

GrowReach PostHog Audit, Analytics Report & Implementation Guide

Summary: Audit of GrowReach’s PostHog implementation — tracking gaps, event taxonomy, and the recommended setup to make CRO analysis reliable.

Date: July 28, 2026 Author: CRO Agent (via PostHog MCP) Data Window: June 29 – July 28, 2026 (30 days) Project: GrowReach (growreach.app) — AI Engagement Automation SaaS

---

Table of Contents

1. Executive Summary
2. For the Product Manager — What’s Happening with Our Visitors & Users
3. For the Development Team — Technical Issues to Fix
4. For the Designer — UI/UX Improvements Needed
5. PostHog Implementation Guide — Complete Setup for Dev Team
6. Appendix: Raw Data Tables

---

1. Executive Summary

The Big Picture

Over the last 30 days, 239 unique visitors came to GrowReach. Of those:

• 16 signed up (6.7% conversion)
• 4 paid (1.7% conversion)
• 101 hit a JavaScript error (42% of all visitors)
• 44.3% bounced without viewing a second page

The #1 Problem: A TypeError Is Breaking the Site for Half Your Visitors

1,864 JavaScript exceptions hit 101 unique users in 30 days. 99% are TypeErrors. The homepage alone has 938 TypeErrors affecting 76 users — that’s ~49% of homepage visitors hitting a broken page.

This is almost certainly a null reference error in a JavaScript file (likely a WordPress theme or plugin script). The error message property ($exception_message) isn’t captured in the PostHog taxonomy, so we can’t see the exact error text from analytics — you need to open the browser console on growreach.app to find it.

Conversion Funnel at a Glance

239 visitors → 154 homepage views → 11 clicked "Start Free Trial" → 16 signed up → 4 paid

Key drop-offs:

• Homepage → CTA click: 92.9% drop-off (only 7.1% click the primary CTA)
• Pricing page: 57.1% bounce rate
• US visitors: 68.4% bounce rate (vs 41.9% for India)
• Mobile: 53.2% bounce rate (vs 38.2% desktop)

What’s Working

• Trial-to-paid conversion is healthy: 4 paid from 16 signups (25%)
• Reddit is driving real traffic: 31 users from Reddit (web + app)
• AI observability is set up: 8 insights tracking costs, latency, errors
• Autocapture is working: Pageviews, clicks, exceptions all captured

What’s Missing

• No custom user_signed_up event — can’t build accurate retention cohorts
• No Actions defined — can’t group related events or rename anything
• No UTM tracking on 99% of traffic — can’t attribute any campaigns
• No session recordings visible — can’t see what users actually experience
• No dashboards for marketing or conversion — only AI observability dashboard exists
• No feature flags or experiments — can’t A/B test anything

---

2. For the Product Manager

What’s Happening with Our Website Visitors

Traffic Overview (30 Days)

Metric	Value
Unique Visitors	239
Total Pageviews	1,002
Sessions	406
Avg Session Duration	5m 38s
Bounce Rate	44.3%
Signups	16 (6.7%)
Paid Conversions	4 (1.7%)

Where Are Our Visitors Coming From?

Channel	Visitors	% of Total	Bounce Rate
Direct	159	64.6%	41.7%
Organic Social	40	16.3%	61.0%
Organic Search (Google)	23	9.3%	40.7%
Referral	22	8.9%	45.7%
AI (chatgpt.com)	2	0.8%	100%

External referrers by source:

• Google organic search: 21 users
• Reddit (web): 16 users
• Reddit (mobile app): 15 users
• LinkedIn (web): 8 users
• LinkedIn (mobile app): 2 users

Note: Reddit is our #2 external traffic source (31 users combined). This channel is working — keep investing in it.

Where Are Our Visitors From?

Country	Visitors	%	Bounce Rate
India	70	29.8%	41.9%
United States	54	23.0%	68.4%
Indonesia	21	8.9%	25.0%
UAE	15	6.4%	33.3%
China	15	6.4%	100%
Germany	9	3.8%	22.2%
UK	6	2.6%	19.2%

⚠️ US visitors have 68.4% bounce rate — nearly 1.6x higher than India. The site may not resonate with US audience. This needs investigation.

Device Breakdown

Device	Visitors	%	Bounce Rate
Desktop	135	58.7%	38.2%
Mobile	95	41.3%	53.2%

⚠️ Mobile bounce rate is 15 percentage points higher than desktop. Mobile also gets heavy TypeErrors (279 from 24 users on Mobile Safari).

Page Performance

Page	Pageviews	Users	Bounce Rate	Avg Time
Homepage (/)	317	154	55.7%	~4s
Pricing	62	25	57.1%	~4.2s
Build My Personal Brand	35	10	0%	~1s
Generate Leads	11	4	0%	~0.02s
Comparisons (TryKondo)	6	5	—	—
Features	4	4	0%	~0.02s

Homepage has 55.7% bounce rate and average time of only ~4 seconds. The pricing page is even worse at 57.1% bounce.

Homepage CTA Performance

CTA Element	Clicks	Unique Users	CTR (of 154 visitors)
Pricing	21	16	10.4%
Sign In	13	11	7.1%
I Want To Build My Brand	13	7	4.5%
Start 14-Day Free Trial	12	11	7.1%
I want to Generate Leads	9	4	2.6%
Watch Demo	5	5	3.2%
Features	3	3	1.9%

The primary CTA “Start 14-Day Free Trial” gets only 7.1% CTR. “Pricing” actually gets more clicks (21 vs 12). The persona buttons (“Build My Brand” + “Generate Leads”) combined get 22 clicks — nearly 2x the primary CTA. This suggests visitors want to understand the product before committing to a trial.

Signup & Conversion Funnel

Step	Users	% of Visitors
Any pageview	239	100%
Homepage visit	154	64.4%
Pricing page visit	25	10.5%
Clicked “Start Free Trial”	11	4.6%
Email registered	16	6.7%
Persona configured	14	5.9%
LinkedIn connected	6	2.5%
Paid plan purchased	4	1.7%

Note: More people signed up (16) than clicked “Start Free Trial” (11). This means some users are signing up through other paths (direct /start page access, sign-in page). The funnel isn’t linear.

Error Impact on Business

1,864 TypeErrors in 30 days. If even 10% of those 99 affected users would have converted without the error, that’s ~10 lost signups. At $19–$69/mo, that’s $190–$690/mo in directly lost revenue from a single JavaScript bug.

---

3. For the Development Team

Technical Issues to Fix — Prioritized

---

🔴 P0: Fix Homepage TypeError (Do Today)

The Problem: 938 TypeErrors on the homepage affecting 76 users. 441 more on the pricing page. 1,845 total TypeErrors from 99 users.

What We Know:

• Error type: TypeError (confirmed via $exception_types property)
• Cross-browser: Chrome (873), Brave (306), Mobile Safari (279), Edge (270), Firefox (60)
• Cross-OS: Mac (833), Android (360), iOS (309), Windows (278), Linux (74)
• Concentrated on: Homepage (938), Pricing page (441), Build My Personal Brand page (257)
• The $exception_message property is NOT in the PostHog taxonomy — we can’t see the exact error text from analytics

How to Debug:

1. Open https://growreach.app/ in Chrome DevTools → Console tab
2. Look for red TypeError messages — they’ll show the exact file, line number, and error text
3. Common causes: Cannot read properties of null (reading ‘classList’ or ‘addEventListener’), a missing DOM element that a script expects
4. Check WordPress theme’s homepage.js, main.js, or any third-party plugin scripts
5. Test in incognito mode to rule out extension conflicts

Fix: Add a null check before the failing operation. This is likely a 1-line fix.

// Before (likely what's breaking):
header.classList.add('sticky');
 
// After (fix):
if (header) {
    header.classList.add('sticky');
}

Expected Impact: Fixing this alone could recover 5-10% of lost conversions.

---

🔴 P0: Add user_signed_up Custom Event

The Problem: We’re using email_registered_on_signup as our signup event, but PostHog’s autocapture can’t give us a reliable signup event. We need an explicit custom event from both frontend and backend.

What to Do:

Add this to your signup success handler (frontend):

posthog.capture('user_signed_up', {
    signup_method: 'email',        // or 'google', 'linkedin'
    plan_type: 'free_trial',       // or 'free', 'paid'
    referral_source: 'direct',     // or 'reddit', 'google', etc.
    is_first_time: true
});
 
// Also identify the user
posthog.identify('user_distinct_id', {
    email: 'user@example.com',
    name: 'John Doe',
    plan: 'free_trial',
    signup_date: '2026-07-28'
});

Also add backend confirmation (Node.js):

client.capture({
    distinctId: 'user_distinct_id',
    event: 'user_signed_up',
    properties: {
        signup_method: 'email',
        plan_type: 'free_trial',
        is_first_time: true
    }
});

---

🔴 P0: Create Actions in PostHog

The Problem: Zero Actions defined. Actions are how you group related events, rename events, and create conversion goals.

What to Do: Go to Data Management → Actions in PostHog and create:

Action Name	Events to Group	Purpose
Signup Completed	email_registered_on_signup + user_signed_up	Single signup metric
Homepage CTA Clicked	All autocapture clicks on “Start 14-Day Free Trial”, “Pricing”, “Sign In”	Track CTA effectiveness
Onboarding Started	persona_settings_configured	Funnel step
LinkedIn Connected	linkedin_account_connected	Funnel step
Paid Conversion	free_plan_purchased + paid_plan_purchased	Revenue tracking

---

🟡 P1: Add UTM Parameters to All Outbound Links

The Problem: 99% of traffic has no UTM source. We can’t attribute any campaigns.

What to Do: Add UTMs to every link you share:

https://growreach.app/?utm_source=reddit&utm_medium=social&utm_campaign=launch_post
https://growreach.app/?utm_source=linkedin&utm_medium=social&utm_campaign=profile_bio
https://growreach.app/?utm_source=google&utm_medium=organic&utm_campaign=seo
https://growreach.app/?utm_source=twitter&utm_medium=social&utm_campaign=tweet

PostHog automatically captures: utm_source, utm_medium, utm_campaign, utm_content, utm_term, gclid, fbclid, msclkid.

---

🟡 P1: Enable Session Recordings

The Problem: We can’t see what users actually experience on the site. Session recordings would show us exactly where they get stuck.

What to Do: Check your PostHog JS config. Ensure disable_session_recording is NOT set to true:

posthog.init('<token>', {
    api_host: 'https://eu.i.posthog.com',
    // Make sure this is NOT set:
    // disable_session_recording: false,  // default is false — good
    session_recording: {
        maskAllInputs: true,
        maskTextSelector: '.ph-no-capture'
    }
});

Also enable in Project Settings → Session Recording → Record user sessions.

---

🟡 P1: Set Up a Reverse Proxy

The Problem: Ad blockers and tracking protection (especially on Safari and Firefox) can block PostHog events. A reverse proxy sends events through your own domain.

What to Do:

# Nginx config — add to your server block
location /ph/ {
    proxy_pass https://eu.i.posthog.com/;
    proxy_set_header Host eu.i.posthog.com;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
}

Then update PostHog init:

posthog.init('<token>', {
    api_host: 'https://growreach.app/ph',  // Your domain
});

---

🟡 P1: Fix Exception Message Capture

The Problem: The $exception_message property isn’t in the PostHog taxonomy. We can only see error types (e.g., “TypeError”) but not the actual error text.

What to Do: Update to the latest posthog-js SDK version. The $exception_list property (available in newer versions) contains the actual error message in the value field. You can then query:

SELECT 
    properties.$exception_list[1].type AS error_type,
    properties.$exception_list[1].value AS error_message,
    count() AS count
FROM events 
WHERE event = '$exception' 
  AND timestamp >= now() - INTERVAL 7 DAY
GROUP BY error_type, error_message
ORDER BY count DESC

---

🟢 P2: Standardize Event Naming Convention

Adopt PostHog’s recommended [object] [verb] format:

Current Name	Recommended Name	Properties
email_registered_on_signup	user signed_up	signup_method, plan_type, referral_source
persona_settings_configured	persona configured	persona_type, industry
linkedin_account_connected	linkedin connected	account_type
paid_plan_purchased	plan purchased	plan_tier, price, billing_cycle
commenting_voice_configured	voice configured	voice_type, tone

Use Actions to group old + new names so you don’t lose historical data.

---

🟢 P2: Add More Custom Events for Key Actions

Track these additional events to get full visibility:

// Pricing page interactions
posthog.capture('pricing viewed', { plan_tier: 'pro', page: 'pricing' });
posthog.capture('pricing plan_selected', { plan_tier: 'pro', billing: 'monthly' });
 
// Feature page engagement
posthog.capture('feature viewed', { feature: 'linkedin-comments', page: '/features/' });
 
// Comparison page engagement
posthog.capture('comparison viewed', { competitor: 'trykondo', page: '/comparisons/' });
 
// Free tool usage
posthog.capture('free_tool used', { tool: 'linkedin-carousel-generator' });
 
// Demo request
posthog.capture('demo requested', { source: 'homepage_cta' });

---

4. For the Designer

UI/UX Improvements Needed on Web Pages

Based on the analytics data, here are specific design improvements that could increase conversion:

---

4.1 Homepage — Above the Fold

Problem: 55.7% bounce rate, average time ~4 seconds. Visitors aren’t staying long enough to understand the value proposition.

Data:

• “Start 14-Day Free Trial” CTA gets only 7.1% CTR (12 clicks from 154 visitors)
• “Pricing” gets more clicks (21) than the primary CTA
• Persona buttons (“Build My Brand” + “Generate Leads”) combined get 22 clicks — nearly 2x the primary CTA

Recommendations:

1. Make the value proposition instantly scannable — 4 seconds is not enough time to read paragraphs. Use a bold headline + 3 bullet benefits max.

2. Test moving the persona buttons above the fold — They’re getting more engagement than the primary CTA. Consider making them the primary interaction point.

3. Add social proof near the CTA — Number of users, testimonials, or logos build trust before commitment.

4. Consider an inline email capture — A simple “Get early access” email field above the fold could capture visitors who aren’t ready to commit to a 14-day trial.

5. “Watch Demo” only gets 5 clicks — Either the button isn’t visible enough, or it doesn’t lead to an actual demo. Check where it goes.

---

4.2 Pricing Page

Problem: 57.1% bounce rate, 73.9% exit rate. Only 4 of 25 visitors click through to trial.

Data:

• 441 TypeErrors on this page (23 users affected)
• Quarterly (-20%) and Yearly (-40%) badges get clicks but don’t convert
• “Continue With Free” gets 7 clicks

Recommendations:

1. Fix the TypeError first — 92% of pricing page visitors hit a JS error. No design fix matters until this is resolved.

2. Add a comparison table or feature highlights — Visitors clicking “Pricing” from the homepage want to understand what they get. Make feature differentiation between tiers obvious.

3. Add a “Most Popular” badge — Standard SaaS pattern that guides decision-making.

4. Consider a monthly vs annual toggle — Currently showing both side by side with discount badges. A toggle is cleaner and reduces cognitive load.

5. Add FAQ below pricing — Common questions about cancellation, LinkedIn safety, and results timeline are already on the homepage. Move them to the pricing page where they’re more relevant.

---

4.3 Mobile Experience

Problem: 53.2% bounce rate (vs 38.2% desktop). Mobile Safari has 279 TypeErrors from 24 users.

Recommendations:

1. Test the full mobile flow — Open growreach.app on an iPhone Safari. If the TypeError breaks the page, mobile users see a broken site.

2. Ensure CTA buttons are thumb-friendly — Minimum 48px touch target, adequate spacing between buttons.

3. Reduce above-the-fold content — Mobile screens have limited real estate. Prioritize the value prop and one primary CTA.

4. Test page load speed on 4G — Mobile users on slower connections may leave before the page finishes loading.

---

4.4 General Design Improvements

1. “Start 14-Day Free Trial” vs “Pricing” — The fact that “Pricing” gets more clicks suggests visitors want to evaluate before committing. Consider:

   • Adding pricing information to the homepage hero
   • Testing “See Plans & Pricing” as the primary CTA
   • Adding a “Free forever” or “No credit card required” trust signal

2. Persona-based onboarding is working — The “I Want To Build My Brand” and “I want to Generate Leads” buttons get strong engagement. Consider making this the primary navigation pattern.

3. Trust signals are needed — Questions about LinkedIn safety and cancellation appear in FAQ clicks. Surface these concerns earlier in the experience.

---

5. PostHog Implementation Guide

Complete Setup Checklist for the Dev Team

Phase 1: Fix Critical Issues (This Week)

• Fix homepage TypeError — Debug and fix the null reference JS error
• Add user_signed_up custom event — Frontend + backend
• Create Actions — Signup Completed, Homepage CTA Clicked, Onboarding Started, LinkedIn Connected, Paid Conversion
• Add UTM parameters to all outbound links

Phase 2: Build Analytics Foundation (This Month)

• Enable session recordings — Config + project settings
• Set up reverse proxy — Nginx config for /ph/ endpoint
• Fix exception message capture — Update posthog-js SDK
• Create Growth Funnel dashboard — Pageview → Signup → Onboarding → LinkedIn → Paid
• Create Marketing dashboard — Traffic by channel, country, device, bounce rate
• Create Error Monitoring dashboard — TypeError count, error rate by page, rageclicks
• Create Revenue dashboard — Signup → Paid conversion, plan distribution

Phase 3: Advanced Features (Next Quarter)

• Standardize event naming — Adopt [object] [verb] convention
• Add custom events for pricing, features, comparisons, free tools, demo requests
• Enable Marketing Analytics beta — Connect ad platforms
• Set up A/B testing with feature flags
• Add user surveys — Exit intent, post-signup, NPS
• Create cohorts — Trial users, power users, US visitors, mobile visitors
• Save HogQL views — Daily error rate, signup funnel, mobile vs desktop conversion

PostHog JS Config (Reference)

posthog.init('<ph_project_token>', {
    api_host: 'https://eu.i.posthog.com',          // EU region
    // After reverse proxy:
    // api_host: 'https://growreach.app/ph',
    
    // Recommended defaults for new projects
    defaults: '2026-05-30',
    
    // Session recording
    session_recording: {
        maskAllInputs: true,
        maskTextSelector: '.ph-no-capture'
    },
    
    // Exception autocapture (enabled by default in recent versions)
    // enableExceptionAutocapture: true,
    
    // Person profiles
    person_profiles: 'identified_only',  // or 'always'
});

Key HogQL Queries to Save as Views

Daily TypeError Rate:

SELECT 
    toStartOfDay(timestamp) AS day,
    count() AS errors,
    count(DISTINCT person_id) AS affected_users
FROM events 
WHERE event = '$exception' 
  AND timestamp >= now() - INTERVAL 30 DAY
GROUP BY day
ORDER BY day

Signup Funnel with Error Impact:

SELECT 
    countIf(event = '$pageview') AS pageviews,
    countIf(event = 'email_registered_on_signup') AS signups,
    countIf(event = 'paid_plan_purchased') AS paid,
    countIf(event = '$exception') AS errors
FROM events 
WHERE timestamp >= now() - INTERVAL 30 DAY

Mobile vs Desktop Conversion:

SELECT 
    multiIf(
        properties.$device_type = 'Desktop', 'Desktop',
        properties.$device_type = 'Mobile', 'Mobile',
        'Other'
    ) AS device_type,
    count(DISTINCT person_id) AS visitors,
    countIf(event = 'email_registered_on_signup') AS signups
FROM events 
WHERE timestamp >= now() - INTERVAL 30 DAY
  AND event IN ('$pageview', 'email_registered_on_signup')
GROUP BY device_type

---

6. Appendix

Raw Data Tables

A. Exception Types (30 days)

Exception Type	Count	Affected Users
TypeError	1,845	99
Error	12	4
ReferenceError	4	2
SyntaxError	2	2
UnhandledRejection	1	1
Total	1,864	101

B. TypeErrors by Page

Page	TypeErrors	Affected Users
Homepage (/)	938	76
Pricing page	441	23
Build My Personal Brand	257	10
Generate Leads from Comments	65	3
Features	6	3

C. TypeErrors by Browser

Browser	Exceptions	Users
Chrome	873	49
Brave	306	11
Mobile Safari	279	24
Microsoft Edge	270	5
Firefox	60	6
Chrome iOS	39	3
Safari	27	2

D. TypeErrors by OS

OS	Exceptions	Users
Mac OS X	833	28
Android	360	24
iOS	309	25
Windows	278	14
Linux	74	10

E. Signup & Conversion Events (30 days)

Event	Count	Unique Users
email_registered_on_signup	16	16
persona_settings_configured	14	14
linkedin_profile_url_submitted_on_signup	13	13
is_manual_commenting_configured	8	8
linkedin_account_connected	6	6
paid_plan_purchased	4	4
commenting_voice_configured	2	1
free_plan_purchased	1	1

F. $identify Events (30 days)

Metric	Value
Total $identify events	37
Unique identified users	22

G. Rageclicks (30 days)

Metric	Value
Total rageclick events	48
Affected users	16

H. Existing PostHog Insights

Insight Name	Created
Generations by HTTP status	Jul 23
Generation latency by model (median)	Jul 23
AI Errors	Jul 23
Generation calls	Jul 23
Cost by model (USD)	Jul 23
Cost per user (USD)	Jul 23
Total cost (USD)	Jul 23
Generative AI users	Jul 23
Traces	Jul 23
Pageview funnel, by browser	Mar 6
Referring domain (last 14 days)	Mar 6
Growth accounting	Mar 6
Retention	Mar 6
Weekly active users (WAUs)	Mar 6
Daily active users (DAUs)	Mar 6

I. Existing Dashboards

Dashboard Name	Created
AI observability default	Jul 23
PROD Dashboard	Mar 6

---

Related

• growreach-cro-deep-dive-report

---

Report generated by CRO Agent via PostHog MCP. Data source: PostHog EU Cloud, Project ID 136693 (GrowReach PROD).