Search “long island web design” and the first page is eight or nine agencies, every one of them promising fast, modern, mobile-first websites. Fair enough: that is what you’d want. So I did the obvious thing and ran each of their own homepages through the test Google uses to judge websites.
The test is free and public. You can repeat every number in this article in about a minute, and I’d encourage you to, because the point isn’t to embarrass anyone. The point is that “fast” is a measurable claim, and you should measure it before you hire the people making it.
average performance score across the eight agencies, out of 100
median time for an agency homepage to show its main content on a phone
of the eight passed Google’s 2.5-second line for the main content
What I did, exactly
On November 1, 2026, I took the agencies on the first page of Google for “long island web design” (the organic results, not the ads) and ran each homepage through Lighthouse 12, Google’s open-source site auditor, using its standard mobile preset: a mid-range phone on a throttled connection, which is how most of your customers will meet your site. That’s the same test that powers pagespeed.web.dev.
Where a result looked like a one-off, I ran it again and kept the better score. The agencies are labelled A to H rather than named, because I’m not trying to pick a fight; the pattern is the story. Everything is re-runnable, and the test is explained in what a Lighthouse score actually means.
The results
| Site | Perf. | Access. | Best pr. | SEO | Main content | Page weight | Requests |
|---|---|---|---|---|---|---|---|
| Agency A | 84 | 84 | 75 | 85 | 2.8 s | 8.6 MB | 142 |
| Agency B | 76 | 80 | 79 | 77 | 5.4 s | 2.0 MB | 70 |
| Agency C | 70 | 100 | 61 | 100 | 11.8 s | 6.2 MB | 69 |
| Agency D | 68 | 83 | 79 | 100 | 7.2 s | 1.1 MB | 67 |
| Agency E | 64 | 95 | 71 | 100 | 8.1 s | 8.7 MB | 114 |
| Agency F | 60 | 94 | 75 | 100 | 7.2 s | 211 MB | 81 |
| Agency G | 46 | 86 | 79 | 92 | 7.9 s | 8.0 MB | 222 |
| Agency H | 44 | 79 | 75 | 92 | 20.1 s | 2.8 MB | 127 |
| This website | 100 | 100 | 100 | 100 | 0.9 s | 404 KB | 15 |
A few things jump out.
- Nobody passed the speed line. Google says the main content should appear within 2.5 seconds. The best agency took 2.8, the median 7.5, and one took twenty seconds, which on a real phone means the visitor left.
- The weight is extraordinary. Four of the eight homepages download more than six megabytes before you’ve read a word. One downloads 211 megabytes: two copies of the same eleven-megabyte video and a set of twelve-megabyte screenshots of their portfolio. On a phone plan that’s a measurable cost to the visitor.
- SEO scores are fine; the rest isn’t. Six of eight score 92 or better on the SEO checklist, which is the easy part: titles, descriptions, crawlable links. The performance and best-practice scores are where the actual engineering shows, and they average 64 and 74.
- Seven of eight run on WordPress with page builders, and carry between 67 and 222 separate requests: themes, plugins, tracking pixels, chat widgets, font kits. Each one is a small delay, and they add up to the chart above.
Why this matters more than it looks
This website: 100 · 100 · 100 · 100, main content in under a second.
An agency’s own homepage is the one site they had unlimited time, budget and motivation to get right. It is their best work, or close to it. If it loads in eight seconds on a phone, the site they build for you will not be faster, because the way they build is the reason it’s slow: a theme, a page builder, a stack of plugins, a video in the hero, and a tracking script for every tool they use.
And speed isn’t a vanity number. Google uses page experience in ranking; visitors leave slow pages (the BBC measured ten percent lost per extra second); and ad platforms charge more to send traffic to pages that load slowly. The seven causes of a slow website are all visible in the table above.
Run the test yourself
- Open pagespeed.web.dev.
- Paste in the agency’s homepage address and press Analyze. Leave it on the Mobile tab; that’s where your customers are.
- Wait about a minute. Read the four circles and, under them, the Largest Contentful Paint time.
- Now paste in two or three sites from their portfolio. Their own homepage is their best case; their client sites are what you’ll actually get.
Run it twice if a number looks odd; connections vary. Then ask the agency, politely, why their site scores what it scores. A good answer is specific. A bad answer is “Lighthouse doesn’t matter.”
Four questions to ask before you hire anyone
- “What does your own homepage score on PageSpeed, on mobile?” They should know. If they don’t, run it in front of them.
- “Will my site be built on a theme and page builder, or written for me?” Neither answer is wrong for every business, but the first one is why most sites land in the 50s.
- “Who owns the site, the domain and the hosting when we’re done?” The answer should be “you.” Here’s what to check.
- “Show me three client sites that pass Google’s speed line on a phone.” Not a portfolio of screenshots; live addresses you can test.
I build websites for Long Island businesses, so I’m not neutral. That’s why every number here is from Google’s own tool and re-runnable by you, and why the agencies aren’t named. Test my site the same way: gbfm.dev on PageSpeed, and any site in the gallery.
“Fast” is a claim anyone can make. It’s also a number anyone can check.
Want me to look at your site?
Send me the address and I’ll tell you what’s slowing it down, what it’s missing, and what I’d fix first. No charge, no pressure.