Server-Side Tracking and Measurement Limits

Client-Side Is Dying

Traditional client-side tracking (JavaScript tags firing in the browser) faces compounding threats:

• • Ad blockers: Some visitors block browser tags
• • Safari ITP: First-party cookies limited to 7 days, third-party blocked entirely
• • Firefox ETP: Tracking protection can limit some browser identifiers
• • Brave, DuckDuckGo: Growing privacy-first browsers that block all tracking by default
• • Consent: A denied choice must still be respected, whatever the transport

The effect depends on the audience, consent choices, browser behaviour and measurement design. Reported ROAS may differ from actual commercial contribution and should not be treated as a complete account of demand.

The Scale of Data Loss

Review the potential sources of measurement gaps in your own account:

• • Consent: Which orders are eligible to be measured?
• • Browser restrictions: Which signals depend on cookies or browser tags?
• • Data quality: Are order values, refunds and duplicates handled correctly?
• • Attribution: Which sales would have happened without an ad?

Do not equate a percentage of missing conversions with the same percentage of missing ad spend or profit. These are different measures.

Impact on Smart Bidding

Smart Bidding (tCPA, tROAS, Maximise Conversion Value) relies on conversion data to learn. When data is missing:

• • The algorithm undervalues high-performing segments (because it can't see their conversions)
• • CPCs are set lower than optimal (because reported conversion value is understated)
• • You lose auctions to competitors with better tracking (they can afford to bid higher because they see more data)
• • Campaign learning periods take longer (fewer data points mean slower learning)

Incomplete signals can affect optimisation, but sending more events is not by itself evidence of better bidding or incremental profit.

Server-Side Explained

Server-side tracking works differently from client-side:

• • Client-side: Browser fires JavaScript tag → data sent from browser to Google
• • Server-side: Browser sends data to your server → your server sends data to Google

A server can use order information available to the business, but it does not automatically know the visitor's ad journey or consent state. Only eligible data should be sent.

This does not circumvent consent. Compare eligible order data with recorded events and check for duplication before treating a change as an improvement.

Implementation Options

The main approaches:

• • Google Tag Manager Server Container: A managed server container with hosting and maintenance requirements.
• • Managed hosting: Third-party hosting can reduce operational work but still requires a consent and data-quality review.
• • Shopify native: Shopify's Customer Events API with server-side dispatch. Simplest for Shopify stores.
• • Custom API integration: Direct uploads require engineering resources and careful eligibility and deduplication checks.

Consent Mode Interaction

Server-side tracking and Consent Mode v2 work together. Consent Mode handles the legal framework (what you're allowed to track). Server-side tracking handles the technical framework (how you capture data that users consent to share).

With both in place, you maximise data capture within legal boundaries. Without server-side, even consenting users' data gets lost to technical blockers. That's data you have permission to use but can't capture.

Next Steps