I checked 4,258 small business websites across the US. Here's what I found.

I build websites for small businesses, and I kept seeing the same few problems. I'd already counted them in South Florida. This time I wanted to know if it was a local thing or a national one, so I checked the whole country.

I drew the businesses from Overture Maps, a free, open map of places that anyone can download, so anyone can check my work. I took ten kinds of business and drew 650 of each, spread across the 50 states and DC in proportion to where those businesses are. I left out chains, franchises, corporate offices and health systems: these are independent, owner-run businesses. On 19 October 2026 a script read each homepage. 4,258 loaded properly and got checked.

• Can a customer reach you?

  A contact form, a booking link, or a phone number you can tap.

• Can Google read you?

  A page title, a description, and the structured data that says what your business is.

• Can you see who visits?

  Any sign of analytics in the page.

• Ten trades, every state

  Dentists, salons, med spas, chiropractors, law firms, real estate agents, contractors, auto shops, restaurants and accountants.

828 of the 4,258 sites (19.4%) had no contact form, no booking link and no tap-to-call link. I looked on the homepage and, if it linked to one, the contact page too.

356 of those still showed a phone number as plain text, so a customer could copy it. The other 472, 11.1% of every site I checked, showed no phone number in the page at all. Some of those may have an email address or a number inside a picture. Either way, the customer has to work for it, and most won't.

Restaurants need a word of fairness. Plenty are reached by walking in or through an ordering app, and I didn't count ordering links. But nearly a third of them (108 of 359) didn't show a phone number in the page either.

The kind of site matters even more than the kind of business. On WordPress sites it was 6.5% (95 of 1,457). On sites I couldn't match to a known builder, it was 29.3% (405 of 1,383).

A lot of people look up a local business on their phone. When your number is a link, they tap it and it rings. When it's plain text, they have to press, hold, copy, open the dialer and paste. That's where you lose them.

Of the 3,392 sites that showed a phone number, 1,016 (30.0%) didn't make it tappable. Across all 4,258, only 58.7% had a tap-to-call link.

It's the cheapest fix on this page. On any builder, it's one link.

Google says it uses your page content and your meta description to write the short summary under your name in search results, and it looks at your page title for the headline. Leave them out and Google guesses.

• Half don't say what they are

  2,135 of 4,258 (50.1%) had no LocalBusiness or Organization structured data, the code that tells Google your type of business, address and hours.

• 27.5% had no meta description

  1,169 of 4,258. That's the line of text under your name in Google.

• Only 36.0% had LocalBusiness data

  1,534 of 4,258. Wix and Squarespace add it on their own; WordPress sites mostly didn't (26.7%).

• Titles are mostly fine

  Only 148 of 4,258 (3.5%) had no usable page title.

5.2% (221) were also missing the small tag that tells a phone how to size the page.

1,740 of the 4,258 sites (40.9%) showed no sign of analytics. No Google Analytics, no tag manager, nothing my scan could see. The owner can't tell how many people came, what they looked at, or whether the site brings in a single customer.

Salons and barbers were the worst group: 212 of 389 (54.5%) showed no analytics. Dentists were the best, and still one in five (109 of 500, 21.8%) had none.

• 27.5%

  no meta description

  1,169 of 4,258

• 40.9%

  no analytics

  1,740 of 4,258

• 15.1%

  missing all three: description, LocalBusiness data and analytics

  641 of 4,258

Here are all ten. Each cell is the number of sites that missed the check, and the share of that trade.

Dentists did best on almost everything. They also lead on online booking: 60.8% (304 of 500) link to it, with med spas (55.2%) and chiropractors, PT and optometrists (52.6%) close behind. Contractors (8.5%) and real estate agents (9.8%) almost never do, which is fine if you don't take appointments.

I expected the regions to differ. They barely did. The share of sites with no way to reach the business ran from 16.8% in the South to 21.7% in the Northeast. Between trades it ran from 9.2% to 40.4%. By website builder it ran from 0.7% on Duda sites (1 of 152) to 29.3% on sites I couldn't match to a known builder.

So if your site is missing these things, it isn't because of where you are. It's how it was built.

117 of the 4,258 sites (2.7%) had a Spanish version or a switch to Spanish that my scan could find. Law firms did most at 8.1% (39 of 481); restaurants least, 1 of 359.

It depends hugely on the market. In my South Florida study, 15.8% had one, about six times the national rate. Florida as a whole sat in between, at 6.2% (23 of 372) in this sample. If a lot of your customers would rather read Spanish, a Spanish page is still rare enough to set you apart.

Some states make contractors put their licence number on their advertising. Florida's statute says it has to appear in every advertisement. California's licensing board says the same rule covers websites. Rules differ by state and by trade, and I wrote about it in my guide to roofing and contractor websites.

50 of the 457 contractor and home service sites I checked (10.9%) showed a licence number on the homepage. In Florida it was 16 of 41; in California, 9 of 42. Plenty said “licensed and insured” without the number.

Not everything is broken, and it would be wrong to say it was.

• 92.0% are secure

  3,917 of 4,258 ended up on https. That still leaves 341 that didn't.

• 94.8% are set up for phones

  4,037 of 4,258 had the tag that sizes the page to the screen.

• Almost none hide from Google

  Only 77 of 4,258 (1.8%) told search engines not to list the page.

• Most footers are current

  Of the 2,627 sites that show a copyright year, 72.7% say 2025 or 2026.

The quickest way: run my free website check. Type your address and it runs the same checks on your homepage and shows you how you compare with these 4,258 sites. Or do it by hand. Grab your phone.

• Tap your phone number

  On your own site, on your phone. If it doesn't ring, it's text.

• Send yourself a message

  Fill in your own contact form and see if it reaches you. Then ask who else gets it.

• Google your business name

  Read the line under your name. If it's nonsense or a jumble, you have no meta description.

• Ask who can see your visitors

  Ask whoever built your site where your analytics are. If nobody knows, you have none.

• Check your map listings

  In the map data I drew from, 1 in 6 listed websites no longer existed. Make sure Google, Apple and Yelp point to the right address.

• Contractors: find your licence number

  Can a homeowner see it on the page they land on?

Every site I build has all of this from day one: a number that rings, a form I test all the way to your inbox, a description and structured data for Google, and analytics you can read. It's never more than $999 to build. Care is $99 a month, it's required while I host your site because your hosting is part of it, there's no minimum term, and if you leave you get a full export, free. If you want to know how your site did on these checks, run the free check or ask me and I'll go through them with you.

I'm putting the limits next to the numbers, because a number without its limits isn't worth quoting.

• Ten trades, about equal. I drew 650 of each, so every trade counts about the same in the totals. The totals describe these ten trades, not every business in America.
• Spread by state. Within each trade, each state got its share in proportion to how many of those businesses it has. The biggest single city was Houston, at 83 of 6,500.
• Independents only. I left out anything Overture tags as a brand, any website shared by three or more listings, location pages on bigger sites, and a list of chains, franchises and health systems. A few edge cases still slip through.
• The homepage only, plus the contact page when I was looking for a form. A site might do something right on another page.
• The page as the server sent it. My script didn't run JavaScript, so a form or a phone number added that way wouldn't be seen. That's why I don't publish form figures for Squarespace, Weebly or Wix, where forms usually load that way.
• Not “dead”. 1,135 of the listed websites didn't exist any more. That says the map listing is out of date, not that the business closed, so I don't claim any share of businesses is gone.
• Not checked: speed. There are no speed numbers here.
• How precise. For all 4,258 sites, a figure is good to about 1.5 points either way; for one trade, about 5 points.

The same checks, the same code, as my South Florida study. I'll run it again and date the new numbers, so you can see whether things get better.