What server-side tracking is
Most analytics starts in the browser: a script on the page sends a request each time something happens. Server-side tracking moves that request to your backend. When your server creates an account, takes a payment or cancels a subscription, it sends the event itself with a secret key.
Ad platforms offer the same idea for conversions. Meta's Conversions API, for example, lets advertisers send web, app and offline events from their server straight to Meta, so the ad platform still learns about sales the browser pixel missed.
Server-side tracking vs a first-party proxy
The term is used for two different setups, and it helps to keep them apart:
| Setup | Who sends the event | What it fixes |
|---|---|---|
| Server events | Your backend, with a secret key | Events the browser never sees or can't be trusted with: signups, payments, cancellations |
| First-party proxy | The browser, through your own domain | Ad blockers that block known analytics domains; pageviews still need the script |
Server events can't tell you where a visitor came from on their own: the referrer, UTM parameters and click IDs live in the browser. The strong setups combine both: the browser records the visit and its source, and the server records the outcome against the same visitor or user ID.
How to set it up
- Pick the events your backend is the source of truth for: account created, trial started, payment succeeded, subscription ended.
- Pass an identifier from the browser to your server, such as the analytics visitor ID, or link your user ID to the visitor when they sign in.
- Send the event from the backend with the tool's secret key, never from client code.
- Make sending idempotent: a retry with the same ID shouldn't count twice.
Common server-side tracking mistakes
- Losing the link to the visit. A server event with no visitor or user ID counts but can't be credited to a source.
- Sending the same conversion twice. A browser event and a server event for one purchase double it unless they share an ID.
- Shipping the secret key to the browser. Anyone could then send fake events.
- Expecting it to replace consent rules. Where the data comes from doesn't change what you may collect about a person.
How MetricRaven handles server-side tracking
MetricRaven does both halves. The browser script records visits and their source. Your backend sends verified goals with the site's secret key through server events or metricraven.track() in the Node.js SDK, identified by the visitor ID or your own user ID. Payments come server-side too: from Stripe, Paddle, Lemon Squeezy and other provider webhooks, or the Payments API for custom billing.
To keep browser events from being blocked, the managed proxy serves the script and receives events on a subdomain of your site with one CNAME record. Server events and payments are never affected by ad blockers or the data-center filter. See Link users to visits for joining the two.
Questions
- Is server-side tracking better than client-side tracking?
- For outcomes such as signups and payments, yes: your server knows they happened and ad blockers can't stop them. For where a visit came from, you still need the browser, because the referrer and campaign tags only exist there.
- Does server-side tracking get around ad blockers?
- Events sent from your server do. Browser events still pass through the visitor's browser; serving the analytics script from your own domain reduces how often they are blocked.
- What is the Meta Conversions API?
- Meta's way to receive conversion events from an advertiser's server, website platform, app or CRM instead of only from the browser pixel. It helps Meta count sales the pixel missed.