Blog 8 min *

* Publication times use Tampa, Florida local time (Eastern Time).

A softly blurred analog clock face behind a sharp horizontal amber line bisecting the frame.

The Speed Advice You Got Two Years Ago Is Now Wrong

By Gabriel Lopez-Seco

In March 2024, Google replaced a metric.

It is not a headline that sounds like a business problem. But here is what it means in practice: a website that passed all three Core Web Vitals in 2023 can fail them in 2024, without a single change to its code, without a new plugin, without touching anything. The standard changed. The site did not. The site now fails.

If your last speed audit was in 2022 or 2023, you have not had a speed audit by current standards. You have a report graded against a test that no longer exists.

The Metric Change That Most Business Owners Missed

Core Web Vitals are the three technical performance metrics Google uses to classify website quality in its search ranking signals. There is LCP: Largest Contentful Paint, a measure of how fast your main visible content appears. There is CLS: Cumulative Layout Shift, a measure of whether your page jumps around while loading. And there is the interactivity metric, the one that changed.

Until March 2024, that third metric was FID: First Input Delay. FID measured one specific thing: how long the browser took to respond to the very first interaction on a page. Tap a button once, measure the response time. Done.

In March 2024, Google replaced FID with INP: Interaction to Next Paint. INP is a materially harder test. It does not measure just the first interaction, it measures all interactions throughout a session and reports the worst one. If a visitor fills out a form, navigates through tabs, and finally clicks “Submit,” INP captures every one of those interactions and grades on the slowest.

A site can pass FID comfortably, because the first interaction was fast, while having multiple later interactions that are slow. Under FID, those slow later interactions did not count against the score. Under INP, they are exactly what counts. The Web Almanac 2024 confirmed this: INP replaced FID in March 2024, and the mobile INP gap versus desktop performance remains a significant and documented challenge across the web (HTTP Archive / Web Almanac 2024 Performance chapter, 2024).

The 2023 speed report on your desk tested FID. It is a passing grade on a cancelled exam.

What the Year-over-Year Numbers Actually Show

The Year-over-Year Shift Table built from the HTTP Archive Web Almanac series tells the story in specific terms.

In 2022, only 31% of mobile origins passed all three Core Web Vitals when measured on the INP basis, the same basis that Google uses today (HTTP Archive / Web Almanac 2024, 2024). On the now-retired FID basis, the 2022 mobile pass rate was 39%. The gap between those two numbers: 8 percentage points, is the difficulty premium of INP over FID, measured at the same point in time on the same sites.

By 2025, the directional trend from the Web Almanac series suggests roughly 48% of mobile origins pass all three Core Web Vitals, with desktop at approximately 56% (HTTP Archive / Web Almanac 2025, per secondary summary, these figures are a directional trend anchor from an aggregated source, not directly extracted from the primary chapter text; treat as indicative of the improvement direction rather than a precise count).

That 17-point rise from 31% to approximately 48% is genuine improvement across the web. More sites are clearing the bar.

But “the bar has moved upward” does not mean “a bar-clearing performance in 2022 still clears the bar today.” It means the opposite. The sites now in the passing category are better than the 2022 passing category, because the test is harder. A site that barely cleared the FID-based CWV threshold in 2022 is not likely to clear the INP-based threshold in 2025. It is likely to fail by more than it would have in 2022, because the passing threshold has risen while the site’s underlying performance has not.

The mobile-desktop gap, meanwhile, has persisted. The Web Almanac 2024 confirmed that mobile performance significantly lags desktop, and the approximately 8-point gap between mobile and desktop pass rates (48% vs 56%) reflects a structural challenge that has not closed despite overall improvement (HTTP Archive / Web Almanac 2024 and 2025, 2024/2026).

The Economic Standard Also Moved

Technical standards did not shift alone. The cost of acquiring a visitor rose at the same time pass thresholds moved upward.

Contentsquare’s 2025 Digital Experience Benchmarks report, covering 90 billion sessions across 6,000 websites globally comparing Q4 2023 to Q4 2024, found that the cost per digital visit rose 9% in 2024 and 19% over two years (Contentsquare 2025 Digital Experience Benchmarks, 2025). At the same time, conversion rates fell 6.1%.

Read those two numbers together: you are paying 19% more per visit than you were two years ago, and those visits are converting at a lower rate. The performance bar for what that money needs to accomplish, getting a paying customer, has risen in both the technical and the financial sense.

This is not a comment on consumer patience. It is a comment on the economics of paid and organic acquisition. Every dollar spent on Google Ads in 2026 buys fewer clicks at higher CPCs than the same dollar bought in 2022. The clicks that do arrive land on a site that now fails CWVs it once passed, and for the majority of those clicks arriving on mobile, they encounter a site with an INP gap versus desktop that has been confirmed as a persistent, significant challenge (Web Almanac 2024, 2024).

A Worked Example: The Professional Services Firm

Take a Tampa Bay professional services firm, an accounting practice, a law office, spending $3,000 per month on Google Ads in 2022. At a hypothetical $8 CPC, that budget bought roughly 375 clicks per month. With a mobile conversion rate reflecting the mid-level performance of a site that passed 2022 CWVs on FID, that might have produced 22 inquiries per month.

In 2026, the same $3,000 at a higher CPC, costs per click in professional services have risen alongside overall digital visit costs, might buy 270 clicks. Those clicks land on a site that now fails INP on mobile, meaning the form interactions that convert inquiries are slower than Google’s “good” threshold. The mobile traffic, which accounts for 72% of ecommerce visits and is similarly dominant in professional services search (Attribute / Monetate Q4 2024, 2025), is converting at a lower rate than the desktop minority.

The firm’s web developer ran a report in 2023. It passed. Nothing has been touched since. The 2023 report tested FID on desktop field data. INP on mobile is a different test entirely, and it is the one that determines whether those 270 paid clicks produce inquiries.

Same budget. Fewer clicks. Harder test. More mobile traffic. Lower conversion rate. The 2022 optimization is now working against you.

Why the 2023 Report Cannot Be Trusted as Current

The specific ways the 2023 report became obsolete:

It tested FID, not INP. Any report from before March 2024 tested the retired metric. INP performance on the same site may be materially different, worse or better, but the 2023 report cannot tell you which.

It likely emphasized desktop field data. The mobile-desktop gap in INP is well-documented. A passing desktop INP does not indicate a passing mobile INP.

It used the 2023 passing threshold as the reference point. The population of passing sites has expanded since 2022, meaning the baseline Google’s ranking signals are calibrated against is more competitive. Barely passing in 2023 is a weaker position in 2025, even with the same performance.

Lighthouse scores are lab tests. If the 2023 report was a Lighthouse score or a PageSpeed Insights result, it is a simulated lab measurement, not the Chrome User Experience Report (CrUX) field data that reflects what your actual visitors experienced. A Lighthouse score of 90 on a 2023 desktop simulation does not describe your mobile INP under current CrUX measurement.

What a Current Audit Looks Like

A current field-data audit pulls your CrUX data, the real-user measurements from your actual visitors in Chrome browsers, and maps them against the 2024 and later CWV thresholds: INP under 200 ms for “good,” LCP under 2.5 seconds for “good,” CLS under 0.1 for “good.”

It looks specifically at mobile performance, because the mobile-desktop gap is documented and persistent. It identifies whether failures are in the first interaction or in subsequent interactions (which FID-era audits could not distinguish). And it maps the current performance against the cost-per-visit reality: what is each failing session costing you, given the 19% increase in visit cost over two years?

Gabriel Lopez-Seco at iServU runs this assessment directly, no account manager, no handoff to a junior analyst. It produces a current-standard field-data read against 2024 and 2026 thresholds, not the 2022 playbook.

If the last report on your desk has a 2022 or 2023 date on it, it is measuring the wrong thing on the wrong standard. The assessment is the only way to know where you actually stand today.

Schedule a free assessment at iServU.

Sources

  1. HTTP Archive / Web Almanac (2024). Performance chapter: CrUX real-user data, millions of origins, including INP/FID transition documentation. https://almanac.httparchive.org/en/2024/performance
  2. HTTP Archive / Web Almanac (2025, via aggregated secondary summary). Performance chapter: CWV pass rates, CrUX data, July 2025 snapshot. https://almanac.httparchive.org/en/2025/performance
  3. Contentsquare (2025). 2025 Digital Experience Benchmarks: 90 billion sessions, 6,000 websites globally, Q4 2023 vs Q4 2024. https://contentsquare.com/press/2025-digital-experience-benchmarks/
  4. Attribute / Monetate E-Commerce Quarterly Q4 2024 (2025). Mobile vs Desktop Checkout 2025 Data. https://www.getattribute.com/blog/mobile-vs-desktop-checkout-2025-data