Why tracking matters
Every CAC number, every ROAS number, every attribution model output starts with a tracking event. If 30% of your purchases never made it into the tracking log, your CAC is reported 30% too high, your ROAS 30% too low, and your channel allocation is being driven by data that doesn’t reflect reality.
Tracking is also where most marketing problems quietly originate. Teams blame attribution models or ad platform algorithms for declining performance when the actual cause is a tracking gap that gradually widened over months.
What “tracking” includes
A complete tracking layer captures four things:
- Events, discrete user actions (page view, add to cart, purchase, signup) with their context (timestamp, product IDs, value)
- Identifiers, anonymous IDs, login IDs, hashed emails, the keys that let you connect events to the same user
- Source attribution, UTM parameters, click IDs, referrer URLs, what brought the user here
- Consent state, what the user has agreed to be tracked for, in what form
Drop any of these and the downstream measurement degrades correspondingly.
What breaks tracking in 2026
Five common failure modes:
- ITP / ETP, Safari and Firefox cap or strip client-side cookies, breaking cross-session identity
- Ad blockers, typically block third-party pixels entirely (Meta, TikTok, Google Tag Manager)
- Consent enforcement, Consent Mode V2 prevents tracking for users who haven’t opted in
- CNAME deprecation, Apple and Mozilla shorten or block CNAMEd third-party cookies
- Pixel failure, async loading, network errors, race conditions before the user navigates away
The combined effect: client-side tracking alone typically loses 30-40% of events.
How to fix tracking
The modern resilient stack is three layers:
- Server-side first-party tracking, events fire from your own domain, server-to-server to ad platforms. Bypasses ad blockers and client-side restrictions.
- Strong identity resolution, stitching anonymous, cross-device, cross-session events into single user profiles via deterministic and probabilistic matching.
- Conversions API (CAPI) integration, for Meta, Google, TikTok and others, push events from your server directly into their conversion pipelines instead of relying on the browser pixel.
Each layer recovers a chunk of the lost data. Together they restore most of it.
Common mistakes
- Relying on a single client-side pixel. Single point of failure. 30-40% data loss is normal.
- Not auditing the tracking regularly. Releases break events silently. Quarterly audits catch drift.
- Treating CAPI as optional. In 2026 it isn’t, without server-side conversion data, the ad platforms’ optimization is blind.
FAQ about Tracking
Why is tracking broken in 2026?
A combination of ITP, ETP, ad blockers, third-party cookie deprecation, and consent restrictions removes 30–40% of conversion events from typical pixel-only tracking.
How do I fix broken tracking?
Move to server-side first-party tracking. Pair browser pixels with CAPI (Conversions API) and use event-ID deduplication. Add a strong identity graph so anonymous sessions stitch to known users.
What is the difference between client-side and server-side tracking?
Client-side fires from the browser, vulnerable to ad blockers and ITP. Server-side fires from your own server to ad platforms, bypassing both. The two are usually paired for resilience.