Mateusz K.

I build the measurement layer that sits under an online business: events defined once before anything is coded, collected at the source in the store, sent server-side to GA4 and the ad platforms, and reported on top of that. The goal is a setup your team can plan a budget on and hand to the next person - documented definitions, not tribal knowledge about which tag does what. I'm a digital analyst at a performance marketing agency in Poland, full time, and I take on a handful of projects here alongside it. Same work, smaller teams: tracking that holds up outside a demo store, and reporting somebody actually opens on a Monday. Most of it starts inside the store. On Shopify I rebuild tracking at the source, with a custom pixel in Customer Events instead of another app that half works, checkout and order status covered, every event gated on consent. WooCommerce and custom builds get the same treatment, with the dataLayer written by hand wherever the plugin won't cooperate. And when the browser stops being a reliable place to measure from, I move the whole thing server-side with Stape: your own tracking domain, first-party cookies, Meta Conversions API with event_id deduplication so one purchase never gets counted twice, Google Ads Enhanced Conversions, plus TikTok, LinkedIn or Microsoft if that's where the budget goes. Consent is where most setups quietly break, so I test both halves of it. Nothing fires before the visitor agrees, ad_user_data and ad_personalization are set the way Google expects, and I check the denied path as carefully as the granted one, because that's the one nobody looks at. Cookiebot, OneTrust, Consentmo, or a banner somebody wrote in-house years ago. The point is numbers you can defend when legal asks about GDPR. Then there's the part you look at every week. Looker Studio for most teams, GA4 exported into BigQuery when the questions outgrow the interface, and if a client needs something Looker can't do, I build the dashboard in Next.js straight off the warehouse, so a whole marketing month fits on one screen. Getting the data there is usually n8n: pipelines pulling marketing and sales APIs into one PostgreSQL warehouse that stays current, with retries, deduplication and backfills you can re-run without breaking anything. Some of what I've built lately: - Rebuilt Shopify tracking with a custom Web Pixels pixel and theme dataLayer events, view_item through purchase, Consent Mode v2 gating, verified end to end. - Google Ads Enhanced Conversions on advanced consent: user data captured and validated on the page, but it only leaves once ad_storage and ad_user_data come back granted. - The full WooCommerce purchase funnel in GTM, custom dataLayer variables and a bit of JavaScript to get transaction and product data out of the theme. - Meta, LinkedIn and Microsoft Ads running side by side in one container, consent-aware triggers, revenue that matches across all three. - A marketing warehouse fed from five APIs with n8n, with a Next.js dashboard on top. I don't hand anything over unverified. Tracking comes with GTM Preview, GA4 DebugView, Events Manager and the actual request payloads, automations come with test runs and logs. You get a short write-up of what changed and why, you keep every account and container, and nothing in the setup depends on me still being around. Every project I've finished here is rated 5.0, with recommendations the clients wrote themselves. If you want a starting point, send me your site, your ad platforms and the one number you don't trust. I'll map how your data flows today and tell you what's broken, what's missing and what it would take to fix, before you commit to anything. I'm on CET and normally reply the same day. If you got here searching for something specific: GA4 audits and migrations, Google Tag Manager, server-side GTM on Stape, Meta CAPI and event deduplication, Google Ads conversion tracking, Consent Mode v2, Shopify Web Pixels and Customer Events, WooCommerce, cross-domain and lead tracking, BigQuery, Looker Studio, n8n, PostgreSQL, Next.js.