Migrate from Google Analytics 4 to Plausible: Step-by-Step Guide

Why migrate

You’re considering Plausible because GA4 keeps a lot of data Plausible never collects: device fingerprints, persistent client IDs, cross-property cookies, and the inferential machinery that turns those into “audiences”. Plausible drops all of that. The benefit is GDPR-clean by default and a UI you can actually read in 30 seconds.

The cost: you lose audiences for ad platforms (CAPI is your fallback), you lose the GA4 demo-graphics layer, and your numbers will not match GA4. They are not supposed to.

Pre-migration checklist

1. Inventory custom events. Pull a 90-day event report from GA4. Group by event name + parameter shape.
2. Inventory custom dimensions. Plausible calls these “custom properties”. Each goal can carry up to 30. If you need more, you have a structural problem you should fix anyway.
3. Inventory audiences. Plausible has none. Audiences live in your ad platform now. Plan a CAPI or Conversions API path per ad surface.
4. Decide hosting. Plausible Cloud (~$9–$190/mo) or self-host (Docker, single binary). Self-host costs more in dev hours, less in monthly fees.
5. Decide proxying. Plausible’s plausible.io/js is blocked by Brave, uBlock, and many corporate firewalls. A proxy on your domain (/stats/script.js) is mandatory for accuracy.

Step-by-step

1. Audit (Days 1–3)

Export GA4 events and parameters. Document the ones with non-zero traffic. Anything below 1% can be dropped — Plausible isn’t your archive.

2. Map (Days 3–5)

Translate each GA4 event to a Plausible goal + custom properties. The Event Mapping Wizard handles the common ones. For e-commerce, model purchase with props:

<br />
gtag(‘event’, ‘purchase’, {<br />
  transaction_id: ‘T_12345’,<br />
  value: 99.00,<br />
  currency: ‘EUR’,<br />
  items: [{ id: ‘sku_1’, price: 99 }]<br />
});</p>
<p>// Plausible equivalent<br />
plausible(‘Purchase’, {<br />
  props: {<br />
    transaction_id: ‘T_12345’,<br />
    value: 99,<br />
    currency: ‘EUR’<br />
  }<br />
});<br />

3. Install Plausible (Day 5)

Add the script tag with the data-domain set to your apex (no www). Use the proxy build if you have ad-blocker exposure:

<br />
<script defer data-domain="example.com" src="/stats/script.js"></script><br />

4. Dual-tag (Days 6–35)

Run both tracking scripts in parallel for at least 30 days. Daily diff with the Parallel-Run Validator. Expect 15–30% session count delta — that’s normal, not a bug. Investigate only deltas above your tolerance threshold.

5. Cutover (Day 36)

Remove the GA4 snippet. Keep BigQuery export running for one more billing cycle as a hedge. Document the final delta numbers in a “what changed” memo for stakeholders.

6. Verify (Days 37–60)

Compare Plausible’s first month vs GA4’s last month at the metric level — pageviews, sessions, top pages, top referrers. Note the structural deltas (referrer stripping, bot filtering) so the team stops asking.

Common gotchas

• Single-page apps: Plausible doesn’t auto-track route changes. You need plausible('pageview') on every history.pushState.
• Cross-domain: Plausible has no built-in session stitching across domains. If you need that, model it yourself with a shared visitor_id property.
• Ad-blocker exposure: Without a proxy, expect 30–50% under-counting on tech-savvy audiences.
• Goal limits: 100 goals per site on Cloud. If you need more, group goals by category and use props.