← Webbership

webbership tracking

Conversion tracking that still counts a sale when the browser blocks ad scripts

Role

Founder, sole engineer

What it runs on

Go, PostgreSQL, HTMX

Why I built it

Before I spend money on ads, I want conversion data I can trust. Browser tracking loses more events every year to blockers, browser limits and cookie rules, and with bad data Google and Meta optimize the wrong way.

How it works

A small script in the browser runs only with consent, and a collector on my server passes each purchase straight to the ad platforms, with source, campaign, gclid and fbclid kept and duplicate events dropped. Through its MCP server an AI agent can manage the Google Ads and GA4 accounts and Merchant Center. It cannot write anything without a preview or a dry run first and a confirm after, and every call it makes is logged.

What it runs on

• Go, PostgreSQL and HTMX
• A first-party browser script with no outside dependencies, loaded only after consent
• Server-to-server calls to the Google and Meta APIs
• An MCP server with dry runs, confirms and an audit log

Where it stands

It runs on ovride.coffee and follows each campaign from the ad click to the purchase. For now it's an internal tool for my own projects.

What this means for you

Your team loses hours on work a system could do? I build it around how you work.

Talk about a system