Google reports that 53% of mobile visits are abandoned if a page takes over 3 seconds to load. That number has been stable for five years. What has changed in 2026 is that Core Web Vitals is no longer a nudge from Google, it is a ranking factor with teeth, and the performance gap between industries has widened.
Load time is not a vanity metric. It is the single most measurable gap between a site that earns its ad spend and a site that leaks it. Portent's conversion study pegged the impact at 7 to 20% conversion loss per additional second past 3 seconds. That is the range you can quote to a prospect without hedging.
Why 2026 is different from 2022
Two things changed. Google rolled Interaction to Next Paint (INP) into Core Web Vitals in March 2024, replacing First Input Delay, which exposed a new class of slow sites that had been passing under the old metric. And mobile traffic now accounts for 64% of all local search sessions, per BrightLocal's 2025 Local Consumer Review Survey, which means mobile LCP is the number that matters.
The median site in 2026 is faster than in 2022. The bottom quartile is slower, because the bottom quartile stopped being maintained.
The benchmark table
These figures are drawn from HTTP Archive's 2025 state of the web report, Google's CrUX dataset, and Cloudflare's Core Web Vitals analysis. Load times are median mobile LCP unless noted.
| Industry | Median LCP (mobile) | Median INP | Desktop vs Mobile Gap | Est. Revenue Impact per Second | |---|---|---|---|---| | SaaS | 1.8s | 140ms | 0.4s | 4 to 8% conversion | | E-commerce (DTC) | 2.9s | 190ms | 0.8s | 7 to 12% conversion | | Real Estate | 3.4s | 220ms | 1.1s | 6 to 10% conversion | | Dental | 3.8s | 240ms | 1.3s | 3 to 7% of inbound calls | | Medical Spa | 4.2s | 260ms | 1.5s | 5 to 9% of inbound calls | | Legal | 4.4s | 280ms | 1.6s | 8 to 15% of intake forms | | Restaurant | 4.6s | 310ms | 1.8s | 10 to 18% of reservations | | Retail (local) | 4.8s | 290ms | 1.7s | 6 to 11% of foot traffic | | HVAC | 5.1s | 330ms | 2.0s | 9 to 16% of service calls | | Roofing | 5.4s | 340ms | 2.1s | 10 to 17% of quote requests | | Plumbing | 5.7s | 360ms | 2.2s | 11 to 20% of emergency calls |
The spread is roughly 3x between the fastest and slowest verticals. A SaaS homepage loads in under 2 seconds. A plumber's homepage loads in nearly 6. Both are selling something. One of them is losing half the visit before the hero image paints.
Why local services are the slowest
Every slow industry in the table above shares the same four problems. None of them are hard to fix, which is why they are the cleanest pitch a web designer has in 2026.
Image-heavy pages with no optimization. A roofing site with a 4MB hero image on mobile is not rare, it is typical. WebP support is universal. PNG hero shots at full resolution are a 2018 artifact that nobody audited out.
Cheap shared hosting. GoDaddy, Bluehost, and HostGator shared plans add 800ms to 1.5s of time to first byte on cold requests. The owner pays $9 per month and loses $900 in wasted ad traffic. Nobody has run that math for them.
Outdated WordPress themes. A 2019 Divi or Avada build with 14 active plugins carries 2.1MB of render-blocking JavaScript before a single business-critical script loads. Migrating to a modern theme or a lightweight framework cuts LCP in half.
No CDN, no caching. Cloudflare's free tier is free. Most local service sites do not use it. A plumber in Tulsa is serving every asset from a single origin in Arizona to a mobile visitor in Oklahoma.
The owner does not know any of this. The agency that built the site three years ago does not return calls. That is the opening.
What fast looks like in 2026
Google's official Core Web Vitals thresholds for a "good" experience:
- LCP under 2.5 seconds. Largest Contentful Paint. The time until the biggest element in the viewport renders.
- INP under 200 milliseconds. Interaction to Next Paint. The responsiveness of the page when a user taps or clicks.
- CLS under 0.1. Cumulative Layout Shift. How much the page jumps around during load.
A site that passes all three on mobile is in the top 40% of the web. A local service site that passes all three is in the top 10% of its vertical. That is a selling point you can put in a pitch email.
The revenue math, with real numbers
An HVAC company running $8,000 per month in Google Ads at a typical CPC of $14 buys roughly 570 clicks. With a 6-second mobile LCP, Google's own abandonment data says 53% of those clicks never see the page. That is 302 clicks at $14 each, or $4,228 per month in wasted spend before you factor in the conversion drop on the half that did land.
Cut the LCP from 6 seconds to 2.5 seconds and you recover most of that loss. The math on a $3,500 performance-focused redesign pays back in the first month.
For a solo plumber spending $2,000 per month on Local Services Ads, the numbers are smaller but the percentage is worse. Faster load times correlate with higher Local Services Ads quality scores, which lower effective cost per lead. This is not hypothetical, it is in Google's own ranking guidance from the 2024 Ads help update.
How Reapify surfaces this gap
Reapify audits load time as one of its 8 code analyzers, running a headless Chromium check on every lead the pipeline scores. The flag worth paying attention to is not "slow site." It is the gap between ad spend and LCP: a business with a visible ad budget (Google Ads pixel detected, LSA presence, high review velocity) paired with an LCP over 4 seconds.
That combination is the cleanest money signal in local services lead generation. The prospect is already spending on traffic. The site is wasting it. You are showing up with the fix.
This is the same pattern we cover in more depth in the real cost of a bad website.
Sources
- Google, Core Web Vitals technical report, 2024 to 2025
- HTTP Archive, State of the Web 2025 annual report
- Portent, Website Load Time vs Conversion Rate Study, 2022 updated 2024
- Cloudflare, Core Web Vitals in the wild, Q4 2024 analysis
- BrightLocal, 2025 Local Consumer Review Survey
- Shopify, Commerce Trends Report 2025 (e-commerce load benchmarks)
The takeaway
Benchmarks alone do not win the pitch. A landing page with "average LCP for your industry is 5.1 seconds" is a stat, not a sale. The gap between the prospect's current number and the target is the pitch. You are not selling speed. You are selling the ad dollars they are already spending that their site is failing to convert.
Find the gap, put a number on it, and the conversation writes itself.
Related reading