Feature Request: Speed Index for User Interactions (Post-Interaction Visual Completeness Metric)

[hbpatel142]

Problem Statement

Currently, Speed Index only measures visual completeness during initial page navigation. However, modern web applications increasingly rely on interactions that trigger significant visual changes (e.g., filtering products, loading search results, expanding accordions with lazy-loaded content).

The gap: INP measures responsiveness (time to visual feedback), but doesn't capture how long it takes for the complete visual change to render after an interaction.

Real-World Scenario

Consider an e-commerce site where a user clicks "Filter by Price":

• INP: 150ms ✅ (Good - initial feedback is fast)
• Visual Reality: Loading spinner appears at 150ms, but filtered products + images take 3+ seconds to fully render
• User Perception: The interaction feels slow despite good INP

Business stakeholders understand "the page visually loaded in X seconds" (Speed Index) better than technical metrics. This same intuitive understanding would be valuable for interactions.

Proposed Solution

Introduce Interaction Speed Index (or similar) that measures visual completeness after a user interaction, similar to how Speed Index works for navigation.

Measurement approach:

1. User triggers interaction (click, tap, etc.)
2. Start measuring visual progression from the interaction timestamp
3. Calculate Speed Index-style metric until visual completeness stabilizes
4. Report as "Interaction Speed Index" in Lighthouse user flows

Use Cases

1. E-commerce filtering/sorting - Products loading after filter selection
2. Search interactions - Results rendering after search submission
3. Infinite scroll - Content loading as users scroll
4. Tab/accordion interactions - Lazy-loaded content in expanded sections
5. SPA route changes - Visual completeness of client-side navigation

Why This Matters

• Complements INP: INP tells you when feedback starts; Interaction Speed Index tells you when the change completes
• Business alignment: Non-technical stakeholders understand "visual loading time" metrics
• Identifies hidden issues: Catches cases where INP is good but perceived performance is poor
• User flow optimization: Helps prioritize which interactions need performance work

Comparison with Existing Metrics

Metric	What it measures	What it misses
INP	Time to first visual feedback	Complete visual rendering
TBT/TTI	Main thread availability	Visual progression
LCP	Largest element paint	Post-interaction changes
Interaction Speed Index	Visual completeness after interaction	N/A - fills the gap

Implementation Considerations

• Could leverage existing Speed Index calculation methodology
• Would work within Lighthouse user flows (already supported)
• Could use similar filmstrip analysis post-interaction
• Threshold guidance needed (what's "good" vs "needs improvement")

Community Interest

This would be particularly valuable for:

• E-commerce sites with heavy filtering/search
• SPAs with client-side routing
• Content-heavy sites with lazy loading
• Teams measuring perceived performance vs technical metrics

Related Work

• Core Web Vitals
• User Flow API
• Speed Index documentation

---

Would the Lighthouse team consider this addition to better capture post-interaction visual performance?

Our Experience

We've built an internal site speed analysis service using Lighthouse user flows and consistently see cases where INP metrics are excellent, but users report slow interactions. The missing piece is measuring visual completeness after the interaction - essentially "Speed Index for interactions."