Get more clients · Every business
How Fast Your Website Loads, and What Slow Pages Cost
A study of 37 brand websites and more than 30 million visits found that shaving a tenth of a second off mobile load times moved people measurably further toward a purchase or an inquiry, and fewer than half of mobile sites meet Google's own speed thresholds.
At a glance
- In a study commissioned by Google and run by Deloitte and the agency 55, a 0.1-second improvement in mobile site speed was followed by 8.4% more conversions on retail sites and 21.6% more visitors reaching the form page on lead-generation sites.
- Only 43% of mobile websites met all three of Google's Core Web Vitals thresholds in 2024, compared with 54% on desktop, according to the HTTP Archive's Web Almanac.
- For an illustrative service business with 1,200 site visits a month, a lift of that size in visitors reaching the booking page would be worth about $11,700 a year.
The number and where it comes from
The figure most worth knowing comes from a study titled Milliseconds Make Millions. Researchers monitored the mobile load times of 37 European and American brand sites hour by hour for 30 days at the end of 2019, covering more than 30 million user sessions, audited each site's tracking and confirmed that no site was redesigned during the study; they then compared what visitors did when pages were faster. When every page improved by a tenth of a second across four speed measures, retail conversions rose 8.4%, travel conversions rose 10.1%, and on lead-generation sites, the closest match to a service business taking quote requests, progression to the form submission page rose 21.6%.
Earlier Google research points the same way. A 2017 analysis published by Google's mobile team, built on a neural network trained on bounce and conversion data, reported that as mobile load time grows from one second to seven seconds, the probability of a visitor bouncing rises 113%, and that the average mobile landing page then took 22 seconds to load fully. Those figures are older and came from large advertisers, so they set direction rather than a precise forecast for a salon or a plumbing company.
Google's own definition of "good" is specific. Its Core Web Vitals guidance says the main content of a page, measured as Largest Contentful Paint, should load within 2.5 seconds, the page should respond to a tap within 200 milliseconds, and the layout should not jump around more than a score of 0.1, all measured at the 75th percentile of real visits on mobile and desktop.
The worked calculation
The study's lead-generation figure translates into a simple chain for a service business: visits × share reaching the booking page × share who submit × share who book × average job. Every link in that chain is an assumption below except the 21.6% benchmark, which is applied here as if a small site would see the same effect, something the study did not test.
| Website visits per month (Assumption) | 1,200 |
|---|---|
| Share reaching the booking or quote page (Assumption) | 10% |
| Share of those who submit a request (Assumption) | 30% |
| Share of requests that book (Assumption) | 50% |
| Average first job (Assumption) | $250 |
| Booked jobs per month today (Calculated (1,200 × 10% × 30% × 50%)) | 18 |
| Visitors reaching the booking page after a 21.6% lift (Calculated (120 × 1.216)) | 145.9 |
| Booked jobs per month after the lift (Calculated (145.9 × 30% × 50%)) | 21.9 |
| Added revenue per month (Calculated (3.9 more jobs × $250)) | $972 |
| Added revenue per year (Calculated ($972 × 12)) | $11,664 |
Illustrative example; replace every assumption with your own site analytics and booking records.
The calculation counts only first jobs. A customer who books once and returns brings repeat revenue that a slow page never gets the chance to earn.
How you compare
- Against the web: run your home page and your booking page through Google's free PageSpeed Insights; if your mobile Largest Contentful Paint is above 2.5 seconds, you are in the 41% of mobile pages that miss Google's "good" mark for that measure.
- Against your own pages: booking widgets, large photo galleries and chat pop-ups are common slow spots worth testing on a service business site, so check the page customers book from, not only the home page.
- Your own baseline: write down this month's mobile load time and the number of booking requests before changing anything, then compare after 30 days.
About this research: This article draws on the Milliseconds Make Millions study of 37 brand websites, the HTTP Archive Web Almanac 2024, Google's Core Web Vitals documentation and a 2017 Google analysis of mobile landing pages. The studies covered large brands and advertisers, and the worked example is illustrative.
Sources
- Deloitte and 55, commissioned by Google. Milliseconds Make Millions, 2020, summarized in Milliseconds make millions by O. Demidova, web.dev, 2020. 37 brand sites, more than 30 million user sessions.
- HTTP Archive. Web Almanac 2024: Performance. 2024. Chrome User Experience Report data.
- Google. Web Vitals. web.dev.
- An, D. New industry benchmarks for mobile page speed. The AMP Blog, excerpt of a Think with Google article, February 2017.
Related reading: What Every Extra Step in a Booking Form Costs · Online Booking for Small Service Businesses: What to Turn On and What to Leave Off · What Customers Learn About Your Business Before They Ever Call · Local Search Basics: How Google Decides Which Businesses to Show
Work with osNOVA on a website that loads fast and books customers
We help owner-run service businesses find the slow pages and extra steps between a search and a booked visit, then fix the ones that cost the most, starting from what you want more of, whether that is more first visits, more quote requests or fuller weeks.
Email hello@osnovaops.com about your business
Or leave your email and we will reach out
Tell us what kind of business you run and what you want more of.
Not ready to talk? osNOVA runs free live events for owners.