Cloudflare RUM: choose a problem worth testing
Turn landing-page performance observations into a testable task with comparable periods, one hypothesis and a separate review of accepted leads.
On October 13, 2026, Cloudflare announced LCP and INP breakdowns in RUM. Event and publication date: October 13. Primary source.
For publishers and advertisers using Cloudflare Web Analytics on their sites. This is a global update; verify actual service and data access in your account and country. It is not a new AffiliateCore dashboard feature.
The metric documentation describes the selected site's Core Web Vitals section. LCP concerns main-content loading; INP concerns interface responsiveness.
Our recommendation: one testable task
Choose an important landing page, device and comparable periods. Save the observed sample size and current page version. Ask a specialist to formulate one hypothesis from available report details. Do not change server, images and form simultaneously; that makes the outcome hard to explain.
Teaching situation: the page feels slower after a cover change. That warrants checking its contribution, not assuming causation. Compare a reproducible path before and after one change, keep a control version and test the form on mobile. If there is no improvement, restore the verified state and reconsider the hypothesis. A small sample is not proven growth.
What the report does not establish
A better speed measure alone does not prove more sales or better leads. Evaluate accepted outcomes separately with consistent definitions and reversals. If visits cannot be joined to requests, do not construct a sequential conversion rate from separate totals.
Give the implementer a short brief: URL, period, device, observed problem, one change, technical validation and rollback condition. That is more useful than a promise to speed up everything. We do not claim the new measurement automatically improves a site or is already connected inside AffiliateCore.