Conversion Tracking: The Complete Guide to GA4, Google Ads and Meta Setup

Without conversion tracking, marketing reports stop at clicks and sessions, and budgets get allocated on gut feeling. With it, you know which keyword, ad, article or AI referral produces revenue. The vocabulary changed recently: Google Analytics now calls important events key events, and reserves the word "conversion" for actions used by Google Ads. This guide covers the full chain, from choosing what to track to reading the numbers without being fooled by them.

What is conversion tracking?

Conversion tracking is the measurement of completed actions against a goal. A piece of code (a JavaScript tag, a pixel or a server-side call) fires when a visitor completes the action, sends an event with its details (name, value, currency, order ID) to an analytics or advertising platform, and the platform attributes the event to the traffic source that led to it.

It answers three questions that traffic alone cannot answer: which channels generate results, how much each result costs, and where visitors drop out before converting. It is the foundation of every marketing analytics practice and the starting point of conversion rate optimization: you cannot improve a rate you do not measure.

Which conversions should you track?

Track few conversions, well defined. Most sites need one or two primary conversions and a handful of secondary signals.

Give each macro conversion a value: the order amount for e-commerce, or an estimated lead value for B2B (average deal size multiplied by the lead-to-customer rate). With values, platforms can optimize for revenue, and tools like a ROAS calculator become meaningful. To check your current baseline, run the numbers through a conversion rate calculator.

How do you set up conversion tracking, step by step?

1. Write a measurement plan

List each conversion, its trigger (thank-you page, button click, server confirmation), its value, and the platforms that must receive it. Name events consistently (for example generate_lead, purchase, sign_up, which are recommended GA4 event names). This one-page document prevents most duplicates later.

2. Deploy one tag on every page

Install the Google tag directly, or deploy Google Tag Manager, which lets you add Google Ads, Meta and other tags without editing the site code each time. Fire conversion events on the confirmation of the action (an order confirmation, a successful form submission), not on the button click, which also counts failed attempts.

3. Mark key events in GA4

In Google Analytics 4, every interaction is an event. Open Admin, then Key events, and mark the events that represent business results. Check them in DebugView first: a wrong parameter at this stage propagates to every report. If you also use GA4 for organic search, our guide to Google Analytics for SEO shows how to read key events by landing page.

4. Create conversion actions in Google Ads

Two options: import GA4 key events into Google Ads (simple, one source of truth), or use the native Google Ads conversion tag (faster attribution, better for bidding). Choose one per action, never both as primary. Then turn on enhanced conversions: the tag sends first-party data such as an email address, normalized and hashed with SHA256, which Google matches with signed-in accounts to recover conversions that cookies miss. The variant "enhanced conversions for leads" matches offline sales back to the original click. Our guide to Google Ads covers campaign structure around these actions.

5. Add the Meta Pixel and the Conversions API

The Meta Pixel tracks in the browser; the Conversions API sends the same events from your server, CRM or platform, which survives ad blockers and browser restrictions. Meta recommends running both (a "redundant setup"). In that case, send the same event name and the same event ID through both channels: Meta deduplicates the pair when both match within 48 hours. Without a shared event ID, every purchase counts twice.

6. Tag every campaign link

Emails, social posts, partner links and QR codes carry no reliable referrer. Add UTM parameters (source, medium, campaign) with a UTM builder, with a fixed naming convention, otherwise the conversions land in "direct" or "unassigned".

7. Test before trusting

Complete a real conversion yourself and follow it through GA4 DebugView, Google Tag Assistant and Meta's Test Events tool. Check the value, the currency, the count (one per order) and the source. Repeat the test after every site release: theme updates and checkout changes break tags silently.

Which tool measures what?

How does consent affect conversion tracking?

In Europe and in a growing number of jurisdictions, advertising and analytics cookies require consent. Google's consent mode passes the visitor's choice to Google tags through four parameters: ad_storage, analytics_storage, ad_user_data and ad_personalization. Denying ad_user_data also disables the hashed data used by enhanced conversions.

There are two implementations. In basic consent mode, tags stay blocked until the visitor accepts, and nothing is sent if they refuse. In advanced consent mode, tags load with consent denied by default and send cookieless pings, which lets Google model the missing conversions once a data threshold is reached. Either way, you need a consent banner that sends these signals; our comparison of cookie consent software lists the main options. Expect a gap between real sales and tracked conversions: compare tracked revenue with your back-office revenue every month to know its size.

Which attribution model should you use?

Attribution decides which touchpoint gets the credit. Google Ads now supports two models: data-driven, the default for most conversion actions, which splits credit based on your account's history, and last click. According to Google Ads Help, first click, linear, time decay and position-based models are no longer supported, and actions that used them were moved to data-driven.

Keep in mind that every platform counts its own way. Google Ads and Meta each claim the conversions their ads touched, with their own click and view windows, so the sum of platform reports is always higher than your real sales. GA4 offers a neutral cross-channel view, and your CRM holds the truth on revenue. Part of the influence is invisible to all of them: a buyer who read your brand's name in an AI answer, then typed your URL, shows up as direct traffic. This is the problem of zero-click attribution.

What mistakes skew conversion data?

1. Double counting: the GA4 import and the native Google Ads tag both set as primary for the same action, or a Pixel and a Conversions API event without a shared ID.
2. Thank-you page reloads: a page-view trigger counts each refresh. Use the order ID as a transaction ID so platforms ignore repeats.
3. Payment gateway self-referrals: a visitor returning from PayPal or a bank's 3-D Secure page gets credited to that domain. Add these domains to the list of unwanted referrals in GA4.
4. Too many primary conversions: when page views and scrolls are primary, bidding optimizes for cheap, useless actions.
5. Missing values: without a value, every lead weighs the same and ROAS cannot be calculated.
6. Internal traffic: your team's test orders and form fills inflate results. Filter internal IP addresses and remove test transactions.
7. Silent breakage: a new form plugin or checkout step stops the tag. A weekly check of key event counts catches it within days.

Tracking also only measures the destination. If the page that receives the traffic converts poorly, the fix lies in the page; see our guide to building a landing page that converts.

How do you track conversions from SEO and AI?

Organic conversions use the same key events: in GA4, filter on the Organic Search channel and break results down by landing page to see which articles produce leads. Search Console shows clicks and queries, not conversions, so the two tools work together.

Visits from AI assistants have their own channel. GA4's default channel group now includes an AI Assistant channel, for visitors arriving from sources like ChatGPT. That covers AI-referred traffic that keeps its referrer; the rest arrives as direct traffic. Our step-by-step guide to measuring AI citations in Google Analytics shows how to isolate these sessions and their key events.

Because many AI answers mention a brand without sending a click, conversions only tell part of the story. Sorank's SEO and GEO dashboard puts Search Console performance next to your visibility in AI answers, so you can relate organic conversions to the queries and prompts that feed them.

Measure first, then optimize

Conversion tracking is only useful if every number can be trusted: few conversions, real values, one tag, deduplicated events and consent handled properly. Once the data is clean, you can shift budget to what works and grow the channels that cost nothing per click. To see whether ChatGPT, Gemini and Perplexity recommend your brand before visitors even reach your site, try Sorank's AI mention tracking.