Definition
Conversion tracking is the system of recording user actions (purchase, signup, demo request, etc.) and feeding them back to ad platforms — enabling those platforms to optimise bidding, audiences, and creative toward the actions that matter.
Without conversion tracking, paid advertising optimises for clicks. With conversion tracking, it optimises for outcomes. The difference is the difference between a low-CPC campaign with no revenue and a higher-CPC campaign that pays for itself. Most accounts that struggle to scale paid spend have a conversion-tracking problem at the root.
Modern conversion tracking is more involved than it used to be. iOS 14.5 (2021) and Apple's App Tracking Transparency, Chrome's third-party cookie deprecation, and EU privacy regulations have all eroded client-side tracking. Server-side conversion tracking (sending events from your backend to ad platforms via Conversion API or equivalent) is now the operational baseline for serious advertisers, not the future.
Origin
Google AdWords introduced conversion tracking in 2003. Facebook Pixel launched in 2015 (replacing the older Custom Audiences pixel). Conversion API (server-side) launched in 2020 in response to iOS 14 changes. Privacy-driven evolution continues year-over-year.
How it works
- Define what counts as a conversion — purchase, signup, qualified lead, etc.
- Implement client-side tracking (Pixel, gtag.js, or via Tag Manager).
- Implement server-side tracking (Conversion API, Enhanced Conversions, server-side GTM).
- Send conversion values when possible (revenue, not just count).
- Verify event reliability — Pixel Helper, Tag Assistant, server-side payload inspection.
- Set up de-duplication between client and server events.
- Test by triggering a real conversion and confirming it lands in the ad platform within an hour.
When to use it
Use when
- On every paid-advertising account. Without conversion tracking, optimisation is blind.
- Server-side becomes essential if you spend more than ~$10K/month or run iOS-heavy traffic.
- Before scaling spend. Bad tracking + scaled spend = scaled waste.
Skip when
- Without verifying. Half-implemented tracking is worse than none — the ad platform thinks it has signal but it's noise.
- On ad spend you don't intend to optimise. Pure brand campaigns can survive on impressions; performance campaigns cannot.
Key metrics
- Event match quality (Meta's score; aim for 8+).
- Server-to-client conversion ratio (server should match or exceed client).
- Time-to-event (server should fire within seconds of conversion).
- Conversion delta vs platform-reported (less than 30% gap is healthy).
Examples
- Conversion tracking was firing on add-to-cart, not purchase. The algorithm was optimising for the wrong action and ROAS was tanking.
- After we added Conversion API, Meta's reported CPA matched the CRM within 8% — was 35% off before.
- iOS 14 cut our client-side conversions 40%. Server-side recovered 90% of the loss.
In practice at Makreate
Makreate audits conversion tracking on day one of every paid engagement — and we routinely find it broken. A recent SaaS client had Google Ads optimising on the wrong event (page view, not signup), Meta's pixel not firing on iOS, and no server-side fallback. We rebuilt the stack: Conversion API on Meta, Enhanced Conversions on Google, server-side via GTM, with proper event deduplication. After 30 days, ad-platform-reported conversions matched the CRM within 12%, the algorithms re-optimised on the right signal, and CPA dropped 41% without changing budget or creative. The tracking fix was the highest-ROI work of the engagement.
Advertising →Common mistakes
- Tracking only proxy events (page views, clicks). The algorithm optimises on what you tell it; tell it the truth.
- Skipping server-side tracking. Post-iOS 14, client-side alone undercounts by 30-50%.
- Not de-duplicating between client and server. Inflated conversion counts mislead the algorithm.
- Setting up tracking and never auditing. Tracking breaks silently — checkout flow changes, third-party tags update, devs ship code.
Frequently asked
Client-side vs server-side conversion tracking?
Client-side is simpler but increasingly unreliable on iOS and privacy-protected browsers. Server-side is more reliable, requires backend work. Most serious accounts run both, deduplicated.
How important is event match quality?
Critical on Meta — sub-7 EMQ noticeably hurts campaign performance. Above 8 is the operational target. Achievable via Conversion API + complete user data (email, phone, IP).
How long does tracking take to implement?
Client-side: hours. Server-side: 1–3 weeks depending on stack. Verifying and tuning takes another 2 weeks. Plan accordingly.