EngineeringScore
Performance Medium

Reduce third-party request volume

Once a site is served over HTTP/2 or HTTP/3, the number of requests a browser makes stops being the dominant cost, because multiplexing over one connection removed the old per-host connection limit that made bundling and sprite sheets worthwhile. What still matters is how many of those requests go to origins the site owner does not control: every analytics tag, ad pixel, chat widget, and web font from a third-party CDN adds its own DNS lookup, connection, and script-execution cost that no amount of first-party bundling can remove.

Why it matters

On sites already served over HTTP/2 or HTTP/3, first-party bundling stops being the dominant performance cost: third-party tags (analytics, ads, chat widgets, fonts) usually are, and unlike first-party code the owner cannot bundle them at build time, only choose to run fewer of them.

How to fix

  1. 1 List every distinct third-party origin the page loads a request from, and identify which tag or embed each one belongs to.
  2. 2 Remove tags that are no longer used, and consolidate overlapping tools where one tag manager can load what several separate snippets currently do.
  3. 3 Defer non-critical third-party scripts (chat widgets, marketing pixels) so they load after the page is interactive instead of blocking it.

Common mistakes

  • Adding a new marketing or analytics tag through a tag manager without ever removing the old one it replaced, so both keep firing on every page load.
  • Treating "our site is fast" as proven by a Lighthouse run on a page with ad blockers or consent banners suppressing the third-party scripts that real visitors actually load.
  • Assuming a single tag manager script counts as "one request", when the tags configured inside it can each pull in several more third-party origins of their own.

How to verify

  1. 1 Open devtools network panel, filter by domain, and count how many distinct third-party origins the page actually loads a request from.
  2. 2 Check each vendor script against what it is currently used for (dashboards, reports, or a marketing team change log) to confirm it is still needed.
  3. 3 Re-run the scan after removing or deferring a tag and confirm both the third-party request count and total page weight actually dropped.

Related checks

Performance Request CountPerformance Page WeightPerformance Render Blocking

Ready to see where you stand?

Scan your site and get your Engineering Score with a prioritized roadmap in under a minute, no signup required.

Scan your site