PostHog shows you what GA4 can't — sessions, replays, and the 'why'

GA4 tells you what happened. PostHog tells you how it happened.

That distinction sounds small. It isn't. GA4 can tell you that 60% of visitors to your pricing page bounced. PostHog can show you the session replay where they scrolled down, hovered over the enterprise tier for 4 seconds, looked for a "talk to sales" button that didn't exist, and left. One is a metric. The other is an explanation.

We run both on warehows.ai. Most of our clients run both. They answer different questions.

What PostHog actually is

PostHog is an open-source product analytics platform. It does:

• Session replays — watch exactly what a user did on your site, click by click
• Event tracking — autocapture clicks, pageviews, form submissions without manual instrumentation
• Funnels — define a multi-step path and see where users drop off
• Feature flags — roll out changes to a percentage of users
• Heatmaps — see where users click and how far they scroll
• Surveys — collect in-app feedback tied to specific pages or events

The self-hosted version is free. The cloud version has a generous free tier (1M events/month). For most SMBs, the cost is effectively zero.

GA4 vs PostHog — what each does better

The pattern is clear: GA4 is a traffic tool. PostHog is a product tool. GA4 tells you where visitors came from and how many there were. PostHog tells you what they did and why they left.

The "why" gap

This is the gap that burns most teams.

Your GA4 funnel shows a 40% drop-off between the pricing page and the contact form. You know that people are dropping off. You don't know why. Was the form too long? Was the pricing confusing? Did they get distracted by a broken layout on mobile?

GA4 can't answer these questions. PostHog can.

Session replay is the feature that changes the conversation. You pull up the 20 sessions that dropped off at the pricing page and watch them. Within 10 minutes, you notice a pattern: mobile users can't see the "Book a call" button because it's below the fold. Or desktop users spend 15 seconds looking at the enterprise tier and leave because there's no way to talk to a human without committing to a form.

That's actionable in a way that "40% drop-off" never is.

Setting it up

PostHog setup is a single script tag or SDK install:

// Next.js / React — PostHog provider
import posthog from 'posthog-js'
 
posthog.init('phc_your_project_key', {
  api_host: 'https://us.i.posthog.com',
  capture_pageview: false,       // manual pageview for SPA routing
  capture_pageleave: true,
  session_recording: {
    recordCrossOriginIframes: true,
  },
})

Autocapture handles the rest — clicks, pageviews, form interactions are all tracked without writing custom event code. Session recording starts automatically.

For specific conversion events, add manual captures:

// Track form submission with metadata
posthog.capture('contact_form_submitted', {
  form_page: window.location.pathname,
  has_company: Boolean(formData.company),
  message_length: formData.message.length,
})

Combining PostHog with GA4 in your warehouse

PostHog alone is powerful. PostHog + GA4 + your CRM in a warehouse is a different level.

The architecture:

GA4 ──→ BigQuery (native export)
PostHog ──→ BigQuery (via Fivetran, API export, or PostHog warehouse connector)
CRM ──→ BigQuery (via Fivetran)
         │
         ▼
      dbt models
         │
         ▼
   Sigma / Looker

Staging the PostHog data

-- models/staging/stg_posthog__events.sql
SELECT
    uuid                    AS event_id,
    event                   AS event_name,
    distinct_id             AS posthog_user_id,
    TIMESTAMP(timestamp)    AS event_at,
    properties.$current_url AS page_url,
    properties.$referrer    AS referrer_url,
    properties.$device_type AS device_type,
    properties.$session_id  AS session_id
FROM {{ source('posthog', 'events') }}
WHERE event NOT IN ('$feature_flag_called')  -- filter noise

Joining PostHog sessions with GA4 traffic

-- models/intermediate/int_unified_sessions.sql
SELECT
    COALESCE(ph.session_id, ga.ga_session_id) AS session_id,
    ga.utm_source,
    ga.utm_medium,
    ph.total_events,
    ph.session_duration_seconds,
    ph.pages_viewed,
    ph.had_rage_click,
    ph.reached_contact_form
FROM {{ ref('int_posthog__sessions') }} ph
LEFT JOIN {{ ref('int_ga4__sessions') }} ga
    ON ph.landing_page_url = ga.landing_page_url
    AND ph.session_start BETWEEN ga.session_start AND ga.session_end

Now you can answer: "Users from LinkedIn organic who rage-clicked on the pricing page — which ones ended up submitting the contact form?"

What changes when you add PostHog

PostHog for visitor identification

One pattern we've explored: using PostHog as a lightweight visitor deanonymization layer.

PostHog captures IP addresses (configurable). With an IP-to-company enrichment service (IPinfo, Clearbit), you can identify which companies are visiting your site — without buying a dedicated tool like RB2B or Leadfeeder.

-- models/intermediate/int_posthog__company_visits.sql
SELECT
    ph.ip_address,
    ip.company_name,
    ip.company_domain,
    ip.employee_count,
    COUNT(DISTINCT ph.session_id) AS sessions,
    MIN(ph.session_start)         AS first_visit,
    MAX(ph.session_start)         AS last_visit
FROM {{ ref('stg_posthog__events') }} ph
LEFT JOIN {{ ref('stg_ipinfo__companies') }} ip
    ON ph.ip_address = ip.ip_address
WHERE ip.company_name IS NOT NULL
GROUP BY 1, 2, 3, 4

It's not as granular as RB2B (you get company, not individual), but it's effectively free and doesn't require a third-party pixel.

The honest take

PostHog doesn't replace GA4. GA4 is better at traffic attribution, search console integration, and the BigQuery export ecosystem. PostHog is better at everything that happens after the click — understanding behavior, diagnosing friction, watching users interact with your product.

If you're a content site optimizing for SEO traffic, GA4 alone might be enough. If you're a SaaS product, an e-commerce store, or any business where the on-site experience drives conversion, PostHog fills a gap that GA4 leaves wide open.

Your GA4 says bounce rate is 60%. PostHog shows you the exact 8 seconds where every user gets confused and leaves. One tells you there's a problem. The other shows you the fix.

We wire PostHog into warehouse stacks alongside GA4, CRM, and ad platforms — so the "why" joins the "what" and the "who." If your analytics setup gives you numbers but not explanations, book a discovery call.