Mag Analytics and your privacy

Ad platforms tell you what happened up to the click. Mag Analytics tells you what happened after it. Without that half, Mag is tuning campaigns against a proxy for success rather than success itself.

It is a small script on your site. Mag gives you the right install path for your platform — a prompt to paste into Lovable, the Custom Code panel in Webflow or Framer, an app install on Shopify — and then watches for the first event to confirm it is live.

What it sends

Page views, clicks, and the conversions you define, plus enough attribution to tie a session back to the ad that started it.

What it strips, by default

Redaction is on by default — not a setting you have to find:

• Path templating — identifiers in URLs are replaced by placeholders, so a path like an order confirmation page does not carry the order in it.
• Query allowlisting — query parameters are dropped unless explicitly allowed.
• Click-text masking — the text of what a visitor clicked is masked, so a click on their own name or email does not become an event property.

What is never stored

Your visitors' IP addresses are never persisted. An IP is resolved to a country at the moment the event arrives and then discarded. Mag knows a visitor was in Peru; it has no record of where they were more precisely than that.

Reading it

The funnel view is the one that matters: it joins ad spend to on-site behaviour to the outcome, per campaign. That join is what lets Mag say a campaign is cheap but worthless, or expensive and worth more. You can query the same funnel from your own tools through the MCP server.