How do you choose the right speed metrics for your website?

Why do all speed metrics not carry the same weight across different sites?

The idea seems obvious but is often misapplied in practice: not all sites provide the same experience. A WordPress blog behaves differently than an online booking interface or a complex SaaS application. Google's Core Web Vitals metrics measure different aspects of performance, and not all weigh the same according to what the user is looking for.

FCP (First Contentful Paint) measures the time it takes for the first content element to appear on the screen. For an editorial site where the user expects text, seeing that first paragraph appear quickly determines their perception of performance—even if everything else takes a few more milliseconds. Conversely, FID (First Input Delay), now replaced by INP (Interaction to Next Paint), measures the delay before a user interaction receives a response. For a web application where clicking, typing, and selecting are essential, this responsiveness is what matters.

Does this distinction really change optimization priorities?

In practice: yes. If you optimize an editorial site to improve FID at the expense of FCP, you risk missing the point. A visitor to an article expects to read content—if they see a blank page for 3 seconds, they leave, even if the interaction JavaScript is ultra fast. The opposite is true for a SaaS tool: if the content displays but each click takes 500ms to respond, the experience is broken.

This statement from Martin Splitt reminds us of a reality that many overlook: copying and pasting an optimization strategy from one type of site to another leads to dead ends. A media outlet following e-commerce recommendations on JS responsiveness is wasting time and resources. The first challenge is to accurately identify your site type before even touching the code.

How can I determine if my site is editorial or interactive according to Google?

The boundary is not always clear. A site can mix both—think of an e-commerce platform with an integrated blog, or an application with documentation pages. The practical approach is to analyze the dominant user journey. If 80% of traffic goes to static content pages, you are primarily an editorial site. If most engagement occurs through forms, filters, dashboards, you are on the interactive side.

To decide, look at your behavioral data: bounce rate, time spent, interaction events recorded in GA4 or Matomo. A site where users scroll and read is different from a site where they click, fill out, and validate. This distinction determines which CWV metric weighs most in Google's evaluation of your user experience.

• FCP is prioritized for editorial sites: blogs, media outlets, documentation, product pages focused on description
• FID/INP is prioritized for interactive sites: SaaS, online tools, e-commerce with complex filters, web applications
• LCP remains universal but is not sufficient to characterize the experience according to site type
• A site may require a hybrid approach if the types of pages are mixed—segment by template in your analysis
• Before any optimization, identify your dominant type to avoid investing in the wrong metric

Is this recommendation consistent with what we observe in practice?

Absolutely. Sites that have invested heavily in improving FCP without addressing JS responsiveness—typically media that optimized critical rendering—have seen measurable gains in engagement even though FID was not excellent. Conversely, SaaS platforms have maintained an average FCP as long as INP (the successor of FID) was performing well, without any observable negative impact on ranking or experience.

What Splitt does not explicitly state but is essential: Google does not apply a uniform weight to all CWV metrics depending on the site context. If your site is clearly editorial, a mediocre FCP carries more weight than an average FID in the overall evaluation. This aligns with what we know about the Core Web Vitals mechanics: the three metrics (LCP, FID/INP, CLS) are not simply added together—their relative importance varies.

What nuances should be added to this distinction?

First, the boundary between editorial and interactive is never as clear as in a theoretical diagram. A modern e-commerce site mixes product pages (more FCP focused) with interactive shopping experiences (more INP focused). A media site with a subscription area incorporates application zones. In such cases, you need to segment by template: optimize FCP for product sheets, INP for the checkout flow.

Furthermore, LCP remains the reference metric for all sites combined—it appears in all studies correlating with ranking. FCP and INP complement each other depending on the type, but do not replace LCP. If you have a catastrophic LCP, no matter how good your FCP is: the overall experience is degraded. [To be verified] whether Google really adjusts the weight of metrics depending on the automatically detected type, or if this recommendation remains a prioritization aid for SEOs—no official documentation details the exact algorithmic weighting.

In what cases does this rule not apply or become misleading?

If your site radically changes nature across sections, applying a single strategy is a mistake. A portal with a public editorial area and a private application area needs to handle both types. Prioritizing FCP everywhere because 60% of the site is content ignores the experience of 40% of users—potentially those who convert.

Another case: modern hybrid sites where static content is enhanced with interactivity (filters, maps, embedded tools). A real estate site displays listings (editorial) but mainly thrives on advanced search and the interactive map. Here, INP is strategic even on content pages. Don’t let yourself be boxed in—analyze real behavior before deciding.

What concrete steps should be taken to adapt your performance strategy?

First step: map your site by page type. Identify the main templates and classify them as ‘primarily content consumption’ or ‘primarily interaction’. A spreadsheet is sufficient: typical URL, template, dominant user behavior, priority CWV metric. This mapping guides all your optimization decisions.

Next, set up your monitoring accordingly. In Google Search Console and PageSpeed Insights, filter data by page group. If you are on a CMS like WordPress, segment by post type (article vs. page). On a custom site, use GA4 segments to cross-reference performance and content type. The goal: measure separately the metrics that matter for each type.

What mistakes should you avoid in applying this recommendation?

Do not fall into the trap of “all or nothing”. Just because your site is editorial does not mean that FID/INP doesn’t matter—it just matters less. If you completely neglect responsiveness because “Google said FCP was enough”, you will create friction points (buttons that do not respond, choppy scrolling) that degrade the experience. The idea is to prioritize, not ignore.

Another common mistake: optimizing one metric at the expense of another. Reducing FCP by lazy loading everything can ruin your INP if the main thread remains saturated. Improving INP by reducing JS may slow down initial rendering if you delay it too much. Seek balance—the CWV metrics are interdependent, not silos.

How can I check if my site is optimized for the right metric?

Use PageSpeed Insights and the CWV report from Search Console, but go further with real-time monitoring tools like CrUX (Chrome User Experience Report) accessible via BigQuery or public dashboards. Compare your field metrics (real users) with lab metrics (simulations). If your editorial site has a poor field FCP but an excellent INP, you know where to invest.

Then, correlate performance and business metrics. A good FCP should translate into a lower bounce rate on articles and improved engagement time. A good INP should enhance conversion rates on forms or interactive journeys. If optimizing one metric does not change user behavior, it’s either because it wasn’t the real bottleneck or another issue is masking the improvement.

• Map your site by page type (content vs. interaction) before any optimization
• Prioritize FCP for editorial templates: blogs, articles, descriptive product pages
• Prioritize INP (formerly FID) for interactive templates: dashboards, tools, conversion funnels
• Segment your CWV reports by page type in Search Console and your analytics tools
• Never neglect LCP, which remains the universal reference metric for all sites combined
• Ensure that your optimizations translate into measurable gains in engagement or conversion