Websites for local businesses
27% of local business websites do not work
We checked 28,046 websites belonging to local businesses in eight US metros. More than a quarter of them did not work at all. Here is exactly what we measured, how we measured it, and what it means if you own one of them.
What we checked
We checked the websites of local businesses in eight US metros: Charlotte NC, Raleigh NC, Tucson AZ, Colorado Springs CO, Richmond VA, Chattanooga TN, Boise ID and Des Moines IA.
Those sites belong to 35,788 business listings. The gap between those two numbers matters: chains and multi-location businesses share a single website, so listings and websites are not the same thing. Every headline number below is per distinct website.
Where we checked, and how much of it was broken
| Metro | Listings checked | Broken |
|---|---|---|
| Charlotte NC | 8,368 | 20.7% |
| Raleigh NC | 7,449 | 18.0% |
| Tucson AZ | 4,709 | 20.9% |
| Colorado Springs CO | 4,141 | 19.8% |
| Richmond VA | 3,782 | 22.2% |
| Chattanooga TN | 2,588 | 19.9% |
| Boise ID | 2,558 | 19.1% |
| Des Moines IA | 2,114 | 20.5% |
Broken share is calculated per listing in this table, and per distinct website in the headline numbers. Both come from the same checks.
What counts as broken here
Broken means a customer who typed the address got nothing usable. Four things cause almost all of it, and we counted each one.
- The domain does not resolve at all: 2,603 sites. The name points nowhere and nothing loads.
- The domain has lapsed and is no longer registered: 1,579 sites. Anyone could buy that name today.
- The server refused the connection or failed to answer: 1,451 sites. The site exists but nobody can reach it.
- The site loads an error page: 943 sites. Visitors see a 404 or a server error instead of the business.
We did not count a site as broken when it simply refused to answer an automated check. 3,877 sites (13.8%) blocked us outright, and those were removed from the total rather than being called broken. That exclusion is the difference between a real number and a scare number.
The headline numbers
One in four local business websites we could check was dead. Another one in nine loaded, but failed at the things a customer actually needs.
Everything we measured, in plain numbers
| Measure | Number |
|---|---|
| Business listings examined | 35,788 |
| Distinct websites examined | 28,046 |
| Websites that blocked our checks | 3,877 (13.8%) |
| Websites we could verify | 24,169 |
| Hard-broken websites | 6,576 (27% of verified) |
| Live but unusable on a phone, or visibly abandoned | 2,817 (11.6% of all) |
| Healthy websites | 14,776 (52.7% of all) |
Which trades are worst
It is not evenly spread. The trades that depend most on someone calling right now have the worst websites, which is the opposite way round from what you would want.
Broken rate by trade (trades with 150+ listings)
| Trade | Listings | Broken |
|---|---|---|
| Car dealers | 1,253 | 39.5% |
| Auto repair | 2,378 | 27.8% |
| Spas | 1,162 | 25.4% |
| Nail salons | 1,240 | 25.2% |
| Beauty salons | 2,247 | 24.7% |
| Hair salons | 2,727 | 24.2% |
| Contractors | 1,994 | 24.2% |
| Barbers | 990 | 21.6% |
| Accountants | 722 | 20.8% |
| Massage therapy | 1,409 | 20.5% |
| Real estate agents | 5,696 | 18.6% |
| Landscaping | 577 | 18.5% |
| Lawyers | 1,482 | 18.3% |
| Gyms | 1,536 | 17.4% |
| Dentists | 1,433 | 10.4% |
| Dermatologists | 227 | 3.1% |
Real estate is understated here: 32.6% of those listings blocked our checks, the highest of any trade, so fewer of them could be verified either way.
The sites that still load are often failing too
A site that responds is not the same as a site that works. We measured the ones that did load, and a large minority fail the things customers actually notice.
- 11.8% of sites that load are not usable on a phone, which is where nearly all local searches happen.
- 1.7% still have no HTTPS at all, so browsers warn visitors that the connection is not secure.
- Median load time was 686 ms. The slowest tenth took 2.4 seconds or more, which is long enough to lose the visitor.
- 541 sites still show an under construction page.
- 6,245 sites carry a visibly stale year in their content, some of them years out of date.
How we measured this
Every site was checked once, automatically, over 11 and 12 September 2026. The checks are the same ones a normal visitor makes, and nothing was submitted, logged into or changed.
- A DNS lookup to see whether the domain still exists and resolves.
- A full page request over both http and https, recording the status code and the final address.
- A mobile viewport render to see whether the page is usable on a phone.
- Load timing for the page as delivered.
- Read-only throughout: no forms submitted, no accounts accessed, no data written.
This is our own pipeline. We run these checks to find local businesses whose website is broken, so we can rebuild it before we contact them. The study is the same data, aggregated.
What we are not claiming
The numbers are worth more if you can see their limits, so here they are.
- These eight metros are the ones we work in. They are not a random sample of the United States.
- Every figure comes from a single check on a single day. Sites break and get fixed constantly.
- Classification is automated. Hard-broken means exactly what is listed above, nothing subjective.
- Listings are not businesses. Multi-location companies share one website.
- We cannot see what a broken site cost its owner, and we do not guess at it here.
- No business, domain, owner or contact detail appears anywhere in this study.
If your business is one of the 27%
We rebuild sites like these. When we find one, we build the whole replacement first and send the owner a link to look at, free, before anyone is asked for anything.
The rebuild is five pages: home, services, about, FAQ and contact, built from your own services, your area and your opening hours, with a real map and a tap-to-call button. It goes on the domain you already own and it is live within 24 hours.
It costs $600 once. No retainer, no monthly fee for the first year, no contract and no sales call. You see the finished site before you pay anything, and you keep the preview either way.
Real rebuilds are open right now at preview.localsiteshop.com. Open one on your phone rather than a laptop. That is the only test that matters.
Questions we get about this
Can you tell me whether my business is in this data?
No. This study is aggregate only. No business, domain, owner or contact detail is named or identifiable anywhere in it, and we do not share the underlying list.
How do I know if my site would have counted as broken?
Open it on your phone using mobile data, not the shop wifi. If it will not load, shows a parked page, a for sale notice or an error, it would have been counted as broken.
Do you fix the sites you find?
Yes. We rebuild the whole site and send it to the owner as a free preview before asking for anything. It is $600 flat if they want it live on their own domain, within 24 hours.
Why publish this?
Because we measure it anyway to find businesses that need a working site. The numbers are more useful out in the open than sitting in our own notes.
Will you run this again?
Yes. The next edition will cover more metros and a bigger sample, so the numbers can be compared over time.
If your website is one of the 27%, we have probably already rebuilt it.
$600 flat, live on your own domain within 24 hours. Look at the finished site before you pay anything.