Findings · 2026-11-18 · By VSNARY | Emmanuel Orta
The quote form was the slow part: 980 ms cut to 50 ms
A tree service home page froze for almost a second while it loaded. We blamed our own scripts first. Google's test showed the quote form was doing it.
On a tree service site we run, Google's PageSpeed test measured 980 ms of blocking time on the home page: the page was on screen but could not respond. 16 of the 18 slow tasks in Google's report came from the third-party quote form and the captcha it loads, not from the site's own code. Loading the form when a visitor first touches, scrolls or types, instead of 1.5 seconds after the page opens, cut the reading to 50 ms.
What was wrong #
The home page of a tree service site we run looked finished within a couple of seconds, but for almost a full second it could not react to a tap. Google's PageSpeed test calls this Total Blocking Time and wants it under 200 ms. The page measured 980 ms.
Blocking time is not about how long a page takes to appear. It is the time the browser is too busy running scripts to respond. On a phone, that is the moment someone taps Call and nothing happens.
Our first fix was aimed at the wrong thing #
We assumed the site's own scripts were the cause, so we trimmed them: the script file the home page loads went from 434 KB to 106 KB. On our own test machine the blocking time fell from 480 ms to 175 ms. That looked like a win.
Google's test did not move. It still read 980 ms.
The difference was the machine. Google's test simulates a mid-range phone. Our test ran on a fast computer, which gets through the same scripts so quickly that the slow ones barely show. A speed test on a fast machine can hide the problem your visitors have.
What Google's report actually said #
PageSpeed lists every long task and the script behind it. We read that list instead of guessing:
| Long tasks on the home page | Count |
|---|---|
| Total | 18 |
| From the quote form and the captcha it loads | 16 |
| From everything else on the page | 2 |
The quote form is a third-party widget. It loads its own scripts, and those load a captcha. Together they were doing nearly all of the blocking, on a form that sits below the first screen on both phone and desktop. Visitors were paying for it before they could see it.
The fix: load the form when someone shows up for it #
The form used to start loading 1.5 seconds after the page opened, for every visitor. Now it starts when the visitor first touches the screen, scrolls, presses a key or moves the mouse. If none of that happens, it loads on its own after 15 seconds.
Nothing visible changed. The form is below the first screen, and it is ready about a second after the first touch or scroll, which is before anyone can reach it.
The result #
| Site | Measured by | Before | After |
|---|---|---|---|
| Tree service home page | Google PageSpeed, blocking time | 980 ms | 50 ms |
| Tree service home page | Google PageSpeed, largest content paint | not recorded | 2.2 s |
| Fencing home page, same form | CrawlCheck lab reading, blocking time | 1.26 s | 289 ms |
The fencing site is better and not finished: 289 ms is still above the 200 ms mark, and we have not yet read Google's task list for that page. We are not calling it fixed.
How to check your own site #
- Run your home page through Google's PageSpeed test on the mobile setting. Read the number from that test, not from a test on your own computer.
- Open the long-task list and the third-party section. Each row names the script responsible.
- For every outside widget in that list — a form, a chat box, a map, a review feed — ask where it sits on the page. If it is below the first screen, it does not need to load before the visitor moves.
- Load it on first touch or scroll, with a timer as a fallback. Then run the same test again.
A slow image is the other common cause, and it shows up differently. We wrote up one of those in a slider that took 8 seconds.
What this does not show #
- Real visitors. These are lab readings. The numbers from actual visits are a rolling 28-day average and will take weeks to reflect the change.
- More leads. We verified the form still loads and still submits. We have not measured whether more people fill it in.
- That every form is slow. This was one widget on two sites. Read your own task list.
Published by CrawlCheck. Both sites are run by the person who builds CrawlCheck.
CrawlCheck’s dataset when this page loaded: 570,959 sites measured.
Share this finding:LinkedInXEmailFollow Emmanuel on LinkedIn
Questions this post answers
What is Total Blocking Time?
The time during loading when the browser is too busy running scripts to respond to a tap or click. Google's PageSpeed test treats 200 ms or less as good on mobile.
Why did our own test show a better number than Google's?
Our test ran on a fast computer. Google's test simulates a mid-range phone, where the same scripts take much longer. The phone number is the one visitors feel.
Does delaying a form lose leads?
The form here sits below the first screen and is ready about one second after the first touch or scroll. We verified it still loads and submits. We have not measured a change in lead volume.
What if the visitor never touches the page?
The form loads on its own after 15 seconds, so it is always there for someone who waits.
Can I do this with any third-party widget?
Any widget below the first screen can wait for the first touch or scroll: forms, chat boxes, maps and review feeds. Something visible at the top of the page should not be delayed this way.
Related findings
The main product
Found this on your own site? We fix it for $100.
Scan free to see where you stand. The fix is one site, every finding implemented and re-measured, with a sealed before and after.
Free browser extension
Check this on any site you open, without leaving the page.
CrawlCheck for Chrome grades every site you visit A–F and shows, from your own browser, what search engines and AI assistants are actually served. Free, no account.
Write for the library
Measured something yourself? Submit an article.
Send your finding, the data and the method. A person reads every submission and replies by email; accepted articles are published under your name with a link to your study. Free to submit. Articles are accepted on merit; no one can pay to have one published.
How it works · a reply within about a week
Spotted a mistake or a bug on this page? Report an error. Confirmed errors are fixed and logged on the corrections page.
Comments
Comments are read before they appear. Nothing is published automatically, and no account is needed. An AI agent with a published signing key can post without the queue: it signs the request and the signature is checked against its own domain (how). Those appear at once, marked verified.
Writing about this? Facts, live figures and marks — every number on that page is dated and traceable to a scan.