A shopkeeper who watched customers walk in, glance around, and walk straight back out would know within a week that something about the entrance was wrong. A website owner never gets that view. The visitor who gives up while your page is still loading leaves no trace, no complaint, no abandoned form, just a blank line in your analytics where an enquiry should have been. Slow speed is one of the few marketing problems that actively hides its own evidence.
That is precisely why it is worth measuring deliberately rather than trusting a gut feeling. The good news is that, unlike most invisible problems, this one has a precise, well-documented cost.
How do I know if slow speed is actually costing me leads?
You know by the numbers, not the feeling, because a slow site rarely feels slow to the people who built it. Per Google and SOASTA Research's mobile speed benchmark, cited in Think with Google, 53% of mobile visitors abandon a page that takes longer than three seconds to load, and Portent's analysis of ecommerce performance found conversion rates drop by an average of 4.42% with each additional second of load time across the first five seconds. If half of your mobile traffic never sees the page at all, that is not a rounding error in your lead numbers, it is very possibly the largest single lever you are not pulling.
The honest test is simple: load your own website development project on a real phone, on mobile data rather than office WiFi, and time it with a stopwatch, not a memory of how fast it felt the day it launched.
What exactly counts as "slow" in 2026?
Google now judges this with three specific, named metrics called Core Web Vitals, and treats them as a direct ranking signal, not just a user-experience nicety. Largest Contentful Paint, how long the biggest visible element takes to render, has a "good" threshold of 2.5 seconds. Beyond that line, both your visitors and Google's own ranking algorithm start treating your page as underperforming, quietly, without any warning email arriving in your inbox.
Where does the leak actually happen?
We have written before about the four leaks that account for most lost bookings, an unclear headline, competing calls-to-action, forms that interrogate, and speed. This post is a deep dive into that fourth leak specifically, because "make it faster" is not a task, it is a diagnosis that hides several different, separately fixable causes:
- Unoptimised Images: A single unresized product photo straight off a phone camera can weigh several megabytes; the same image, properly compressed and served in a modern format, weighs a fraction of that with no visible quality loss. This is the single most common cause we find on client sites, and usually the cheapest to fix.
- Third-Party Scripts You Forgot You Added: Chat widgets, old analytics snippets, marketing pixels installed for a campaign that ended a year ago, each one is a separate network request your visitor's phone has to wait on before the page is usable, whether or not the tool is still in use.
- Render-Blocking Code: Fonts, stylesheets and scripts that must fully load before the browser will paint anything at all turn a fast connection into a blank white screen for several seconds, which reads to an impatient visitor as a broken site rather than a loading one.
- Cheap or Distant Hosting: A server physically far from your actual customers, or a shared hosting plan straining under other sites' traffic, adds delay before your page has sent a single byte, no amount of front-end optimisation can fix a slow starting gun.
Example in Action: the six-second homepage
Example in Action: picture a clinic whose homepage takes six seconds to become usable on mobile, three unoptimised hero photographs, a chat widget from a trial they never activated, and a booking form that only appears after everything else has finished loading. Applying Portent's per-second conversion figures, the gap between a six-second load and a two-second one is not a marginal improvement, it is the difference between roughly a fifth of visitors bouncing before the page even finishes rendering and the large majority actually reaching the booking form. Nothing about the clinic's service changed in that comparison, only whether visitors survived long enough to see it.