ApogeeWatcher Shifted INP Monitoring to Post-Load Interactions

ApogeeWatcher, a web performance consultancy, overhauled its client monitoring strategy after Interaction to Next Paint became a Core Web Vital. The firm moved beyond initial page load metrics to track user experience across the full visit.

When Interaction to Next Paint (INP) became a Core Web Vital in March 2024, many web performance strategies, including those at ApogeeWatcher, faced immediate obsolescence. The firm had spent years optimizing for Largest Contentful Paint (LCP), ensuring client homepages achieved green LCP scores. This focus, however, meant neglecting the critical user interactions that occur after the initial page load, where true responsiveness issues often manifested.

INP measures responsiveness across an entire user visit, not just a single loading moment. ApogeeWatcher discovered their existing audit practices were misaligned, leading them to audit incorrect URLs, rely on misleading proxies, and celebrate fixes that failed to improve user experience on interactive elements like menus, cart steps, or filter drawers. The shift required a fundamental re-evaluation of what constituted a 'fast' user experience.

Old Measurement Missed Post-Load Interactions

ApogeeWatcher's previous default audit pack concentrated on loading performance. This typically involved evaluating the homepage and a top landing page using tools like PageSpeed Insights and Lighthouse. Metrics like LCP and Cumulative Layout Shift (CLS) were primary concerns. First Input Delay (FID) rarely failed during these runs, leading to a false sense that interactivity was resolved once the main thread cleared its initial tasks.

This approach left significant gaps. ApogeeWatcher frequently skipped, or measured only once, crucial areas. These included post-load routes on single-page applications (SPAs), such as category filters, account areas, and multi-step checkout processes. They also overlooked templates shared across numerous URLs, like global navigation, consent banners, or chat widgets, where a single slow handler could degrade INP across many pages. Furthermore, mobile performance was often not measured separately, despite its business importance, and field data for URLs with modest traffic was ignored due to thin CrUX buckets and reassuring lab scores.

Identifying Where Poor INP Manifests

ApogeeWatcher observed a consistent pattern: the first user interaction after a page loaded often felt acceptable, but subsequent interactions did not. The fifth interaction, for example, frequently exposed responsiveness issues. Specific actions that revealed poor INP included opening sticky navigation elements, applying faceted search filters, submitting multi-step forms, and hydrating client-rendered blocks after an API round trip. These gestures, critical to user flow, were precisely what INP aimed to measure, and FID had not flagged them as problematic.

Third-Party Tags as Recurring Villains

Third-party tags emerged as another persistent source of INP issues. These were not always found on the homepage. A change to a tag manager script on a blog template or a new A/B testing snippet on a checkout page could introduce significant INP degradation on URLs that had not been re-tested, under the assumption that

Pull quote: “”