The operator's guide to trustworthy conversion data — what conversion tracking is, how signals break, and how to set up, fix, and harden tracking on every platform.
Kay Vink
Conversion tracking is the practice of reporting the actions that matter —
purchases, leads, signups — back to the ad platforms and analytics tools that
optimize toward them. Signal quality is what separates tracking that merely
exists from tracking you can bid on: complete, deduplicated, on-time events
that carry real values.
Smart Bidding, Advantage+, and every other automated bidder optimizes toward
the conversions you feed it. When events go missing, arrive twice, land late,
or carry no values, the algorithms don't pause — they optimize toward the
noise with the same efficiency. That makes signal quality an operations
discipline, not a one-time setup task: the signal decays as sites, consent
rules, and platform APIs change, and the spend keeps flowing either way.
This guide covers the territory job by job: set up conversion tracking per
platform, fix the mismatches between tools, harden delivery server-side, and
feed bidding the values that make value-based strategies work.
Wondering whether Buron and your product-analytics tool do the same job? They don't, and this draws the line: what each one answers, what Buron deliberately is not, and how to decide for your own stack.
Smart Bidding and Advantage+ optimize toward whatever conversion data arrives. Broken, duplicated, late, or value-less signals don't just mis-report; they mis-spend. The four signal defects, why tracking decays after setup, and what running signal quality as an operations discipline looks like.
Your Mixpanel numbers and your ad platform's numbers will never match, and mostly that's fine. Here's how to tell an expected gap from a real defect, and the one-page memo to bring to the budget review instead of a panic.
Marketing mix modeling is sold as a modeling problem. It's really a labeling problem, and the campaign names you write this week decide how much of it you'll be able to run in two years. Here's why history can't be re-labeled, and the naming system that fixes it going forward.
Sum your Meta, Google, and TikTok dashboards and you'll get more conversions than you have customers. That's not fraud and not a bug: each platform attributes within its own window, independently, view-through included. What each number is actually for, and where the real count lives.
Which Smart Bidding strategy you pick matters less than the conversion data you send it. Get the four strategies in one table, then the five inputs that decide whether any of them work: conversion completeness, value variance, freshness, consent-modeled data, and volume. Audit the inputs before you blame the algorithm.
Cost per conversion treats a $99 signup and a $50,000 deal as the same win, and Smart Bidding believes it. Here's what value-based bidding actually changes, how to build values you can trust, and the honest checklist for when switching will make things worse instead of better.
Third-party cookies didn't die the way everyone braced for. The exposure that's actually truncating your attribution is first-party cookie lifetime, and it hits every stack, consent banner or not. A per-browser status check, the lifetime table to bookmark, and the three tracking mechanisms that still work.
Server-side tracking recovers the conversions ad blockers and Safari quietly delete, but it costs money and upkeep to run. Here's what it fixes, what it doesn't, and an honest table for whether your account has crossed the line where it pays off.
What first-party data actually is, versus second-, third-, and zero-party, and the three-move operating model (collect, connect, activate) shown on a real stack: first-party pixel, CRM and order feeds, identity map, warehouse. No CDP required.
Recover the Google Ads conversions that cookies quietly lose, and know exactly what customer data you're sending Google in exchange. What enhanced conversions fix, what they leak, who should skip them, and the 2026 settings merge most guides predate.
Set up Google Ads conversion tracking so the numbers stay trustworthy: the counting, value, and attribution decisions that matter more than the tag, both install paths (gtag and GTM), where to find your conversion ID and label, and the seven ways tracking silently breaks, each with a detection step.
Events, key events, conversions, imported conversions: one table that ends the GA4 naming confusion, then the setup path that actually feeds Google Ads. Define the event, promote it to a key event, import it, and skip the double-counting trap that catches most imports.
Meta conversion tracking is one pipeline, not three products. Install the pixel, add the Conversions API on whichever of three paths fits your stack, and make sure Meta counts each sale once instead of twice. Includes the four reasons that count quietly doubles.
The Meta pixel and the Conversions API carry the same conversions by different routes, so it's almost never one or the other. Here's what changes in your numbers when you run both, the shared-ID mistake that silently double-counts them, and why Meta's free 2026 setup removed the old reason to skip it.
Check one maintained table instead of ten platform docs: the click-ID parameter, client and server tracking, value support, offline import, and consent handling for every ad platform, plus a 60-second setup pointer for each, from X and TikTok to Bing, Reddit, and phone calls.
Conversion tracking breaks silently after launch, and month-end is a bad time to find out. Here's a re-runnable checklist you run like a test suite after every site change: what to check, which tool checks it, and what a pass looks like, plus how to fire test conversions without polluting your data.
Your Amplitude or PostHog attribution will never match Google Ads, and most of the gap is by design. Here's how to audit it per tool: which differences to expect, and which ones are real capture defects to fix.
The ?fbclid= and ?gclid= codes in your URLs are how each platform matches an ad click to the conversion it caused. Here's what every click ID (fbclid, gclid, wbraid, msclkid, ttclid, li_fat_id) actually does, where they get lost in transit, and why stripping them quietly breaks your measurement.
Every UTM parameter, the GA4 dimension it populates, and the failure cases that silently break attribution: case splits, redirect stripping, fbclid pile-ups, auto-tagging precedence, and self-tagged internal links.
What a tracking pixel actually is, from 1×1 image beacon to JavaScript event tag, taken apart on a live first-party pixel: the fields one request carries, what browsers now strip or cap, and why pixels are moving server-side rather than dying.
The Meta (Facebook) pixel, explained for how it actually behaves in 2026: what it does, where your pixel ID lives, the pixel-vs-dataset-vs-CAPI vocabulary sorted out, and the part most guides skip, which is what it can no longer see after ITP, ATT, and consent.
Installing the Meta pixel on Shopify isn't a theme edit anymore: checkout is sandboxed, and the old snippet tutorials leave you with a pixel that goes blind at the one moment that pays. Here's the channel app vs custom pixel call, how to keep purchases from counting twice, and how to prove it fired.
Most Shopify stores quietly count every purchase two or three times: the Google & YouTube app, a leftover manual tag, and a GA4 import all firing at once. Here's how to pick one to trust, silence the rest, and stop Smart Bidding from chasing conversions that never happened.
Generic server-side tracking advice breaks on Shopify, because the platform decides what any pixel is allowed to see before you install a thing. Here are the three routes that actually work, when each is worth the trouble, and how to keep browser and server events from counting every sale twice.
Facebook and Shopify almost never report the same number of conversions, and usually that's fine: each counts something different. Here's how to tell a normal gap from a broken one, and how to fix the five things that cause a bad one.
GA4 and Google Ads count different things on different dates with different attribution, so they never match exactly. Here are the six fixable failure modes, each with a detection step and a fix: date attribution, conversion lag, consent gaps, cross-domain loss, import mixing, and duplicate tags. Plus the residual gap that's supposed to be there.
Denied consent isn't a lost conversion; it's a modeled one, but only if you deploy consent mode the right way. How the four signals work, why the basic-versus-advanced choice quietly decides how many EEA conversions you keep, and how to read your Ads and GA4 numbers once modeling is in them.
When your funnel crosses to an external checkout, the session dies at the border and your best-converting clicks get credited to the checkout domain instead of the ad that earned them. How to spot it, fix the GA4 side and the ads side, and stitch the data when you don't control the other domain.
Stand up a server GTM container the right way: provisioning, custom subdomain, and wiring for GA4, Google Ads, and Meta. Plus what it actually costs to run each month by hosting route, and the honest cases where you shouldn't build one at all.
Identity resolution stitches one person's scattered fragments (anonymous visitor, lead email, CRM contact, customer) into a single entity you can trust for revenue reporting. The match keys ranked, deterministic vs probabilistic settled honestly, a worked stitching pass, and where stitching fails.
Match closed CRM revenue back to your Google Ads clicks with a hashed email instead of a click ID. The setup, the CRM half nobody wires up, why match rates crater, and when to run it beside GCLID imports.
On LinkedIn, a fired Insight Tag only tells you someone filled a form, which for B2B is the cheap half of the story. Here's how to track LinkedIn conversions all the way to pipeline: install the tag, count qualified stages instead of form loads, capture li_fat_id on every lead, and check LinkedIn's claims against real CRM revenue.
Get your closed CRM deals back into Google Ads so bidding optimizes toward revenue, not form fills. The full setup, the 2026 Data Manager migration most guides miss, and the two silent failures that quietly zero out your imports.
Meta retired its standalone offline conversions tool, so CRM uploads now run through the Conversions API. Here's what your export has to send for events to match, and how to reuse the exact pipeline you already built for Google Ads instead of starting over.
Learn which LTV formula fits your business, how to compute it from data you already have, what a healthy LTV:CAC ratio really is, and the step most guides skip: feeding lifetime value back into your bidding so you buy customers, not clicks.
Every predictive-LTV guide assumes repeat purchases, which leaves B2B SaaS uncovered. Here's how to predict a deal's value before the CRM knows it: the early-life signals that work, the data you need in place first, the Google plumbing to feed it into bidding, and an honest buy-vs-build call.
Yes — run both. The browser pixel and the server-side conversions API carry
the same events over different transports: the pixel is easy to block, the
API is durable but blind to some on-page context. Send each event through
both with a shared event ID and let the platform deduplicate; you keep
coverage without double-counting.
Why don't Google Ads and GA4 report the same conversions?
Because they count different things by design. Google Ads credits a
conversion to the click date and applies its own attribution and conversion
windows; GA4 credits the conversion date with its own model. Add
consent-driven modeling and imported-versus-native events, and matching
totals would be the surprise. Compare trends, not absolutes — and alert on
divergence, not difference.
Continuously, in practice. Tracking decays without anyone touching the ad
account: a site release drops a dataLayer event, a consent banner update
suppresses a tag, a platform changes an API version. Quarterly manual audits
catch problems a quarter late; continuous automated checks on coverage,
deduplication, and value presence catch them the week they start.
