Field Notes
The Contractor Website: What Actually Matters on 785 Sites
This summer I pointed one diagnostic at 785 general contractor websites and let it run. Same eleven checks on every site, no exceptions, no judgment calls. Does the phone number tap. Does Google know what kind of business this is. Is anyone measuring traffic. When was the last page published. Can the domain receive email. Is there a proposal system. Is there a client portal.
The frame came from OpenStreetMap, which means it is a set of contractors somebody bothered to put on a map, not a random draw of the whole trade. Treat every number below as roughly plus or minus five points. The shape is the point, not the decimal.
Here is what came back.
| What the check looked for | Missing or broken |
|---|---|
| A proposal or project-management platform a client can reach | 95.8% |
| A client portal | 93.8% |
| LocalBusiness structured data | 68.7% |
| Google Analytics | 49.8% |
| A phone number the phone can tap | 43.3% |
| Any analytics at all | 27.6% |
| Built on a DIY or template builder | 25.6% |
| Nothing published since 2024 or earlier | 12.4% |
| Cannot receive email at its own domain | 8.4% |
| No mobile viewport | 3.6% |
| Broken certificate | 3.4% |
Read it bottom to top. The things everybody worries about are nearly solved. The things that win jobs are nearly absent.
Run the same eleven checks on yours tonight
None of this needs a developer. A browser and about twenty minutes.
- Tap the phone number. Open your site on your phone and tap the number in the header. If it does not start a call, you are with the 43.3% whose number is a picture, not a link.
- Ask Google what you are. Paste your homepage into Google's free Rich Results Test. If it finds no LocalBusiness or GeneralContractor data, Google is guessing at your trade and your service area. So is every AI answer engine. 68.7% of the sites left it guessing.
- Look for analytics. View the page source and search for "gtag" or "G-". Nothing there means nobody is counting. Half the sites had no Google Analytics. About a quarter had nothing at all, not even the site builder's own counter.
- Check who built it. The first twenty lines of source usually say. A template builder is not a sin. It is a ceiling, and you should know you are standing under one.
- Find the newest date. Load /sitemap.xml and read the most recent lastmod. Or find the newest project or post. If the answer is 2024, the site is a monument.
- Confirm mail works. Run your domain through a free MX lookup. If there is no mail record, the form on your contact page is sending into a wall.
- Pinch the screen. On the phone again. If the text is microscopic and you have to zoom, there is no mobile viewport. Rare now, under 4%, but fatal when it happens.
- Look at the padlock. A broken certificate throws a full-screen warning at every visitor. Also rare. Also fatal.
- Find your proposal system. Where does a bid live after you send it? If the answer is "a PDF in their inbox," you are with 95.8% of the trade.
- Find your client login. Is there a place a signed client goes to see their own job? 93.8% of sites had none.
- Count what you can prove. Could you tell me, today, which of your last five jobs came from the website? If not, the site is unmeasured, whatever the analytics tag says.
Most sites I have looked at pass the first eight and fail the last three. That pattern is the whole story.
Your site is probably fine. That is the problem.
As a brochure, the contractor website is in decent shape. 96% of the sites were mobile-ready. 97% had a valid certificate. The trade fixed the things it was told to fix around 2018 and then stopped.
So when a builder asks me about construction website design, I hear the wrong question. Design as an aesthetic exercise, a nicer hero photo, a cleaner font, will not move the number. The site already looks fine. Looking fine is table stakes, and the table is full.
The right question is what jobs the site does. I count five.
It is where the bid lives. The most important document your company produces is the proposal. On most sites, the proposal has no relationship to the site at all. It goes out as a PDF, gets forwarded, gets printed, gets lost. A bid that lives on your own domain, with the client's actual house rendered on it, is a different object. It is your site doing its highest-value job.
It carries the brand to the second reader. Two people decide who builds the house. You met one of them on the site walk. The other only ever sees paper and a screen. The site, the packet, and the proposal are the only version of you that reader gets. I wrote about that reader in the branding post.
It feeds a CRM, not an inbox. A form that emails the owner is a form that gets answered when the owner gets around to it. A form that lands in a pipeline with a stage, a source, and a follow-up timer is a system. Same form. Different company.
It is the front door to the portal. After the signature, the client's relationship with your company moves from "what will you build" to "what is happening at my house." A client login on the site is where that year lives. Without it, that year lives in text messages.
It can be measured. Not "traffic is up." Which jobs, from which channel, against which spend. If the site cannot answer that, it is not part of the business. It is a poster.
A construction company website that does those five jobs will outperform a prettier one that does none of them. Every time.
Cheap is not the problem. Abandoned is.
Earlier this year I traced one low-cost web vendor's template across 109 small trade-business sites in 34 states, contractors among them. Same skeleton, different logo, different town.
I am not going to name the vendor or describe how I found them. The pattern is what matters.
Of the 109, 104 had no analytics at all. 68 could not receive email at their own domain. 35 had not published anything since 2024 or earlier. And as far as anyone could tell, the clients were still paying for them.
The instinct is to blame the price. I do not. A cheap site that someone maintains is a perfectly good tool for a two-truck outfit. The failure here was not the invoice. It was that the vendor built once and left, and the owner never had anyone whose job it was to notice.
That is the real risk in the contractor website market. Not ugly. Not slow. Orphaned. A site with no owner on the business side drifts until the phone number is wrong, the mail bounces, and the last project photo is from a house you finished two winters ago.
If you cannot name the person accountable for your site this quarter, that person does not exist. Fix that before you fix anything else.
What the best home builder websites have in common
The good ones are not the ones with the biggest photos. They share a shorter list.
They are fast. Quick to paint on a phone, on a cell connection, standing in a framed-out kitchen. The best home builder websites I have seen load like a text message.
Each page has one job. The homepage says who you are and where you build. The portfolio shows the work. The contact page gets the lead into the pipeline.
The photos are theirs. Real houses they built, shot properly, in the towns they actually work in. Stock imagery on a builder's site is a confession.
There is something behind the brochure. A proposal system. A client portal. The site is the front of a machine, not a standalone object. This is the trait that separates the top few percent from everyone else, and the survey says it is rare. Fewer than one site in ten had either.
Machines can read them. Structured data tells Google and the AI answer engines exactly who this business is, what trade it works, and where. When a homeowner asks an assistant for a custom builder on the Outer Cape, the sites that get named are the ones that told the machine who they were.
None of that is design in the visual sense. It is design in the engineering sense. The site is built to do work.
The order to fix it in
If you scored badly on the eleven checks, do not start with a redesign. Start here.
- Name the owner. One person on your side of the table is accountable for the site. Could be you. Could be a seat you rent. Cannot be nobody.
- Fix the plumbing. Tap-to-call, working mail, a real analytics tag, structured data. The cheapest work on the list, and the checks where 43.3% and 68.7% of the trade is still failing.
- Put the proposal on the domain. This is the biggest single move. The bid stops being a PDF and starts being a page. I built The Bid Room to do exactly this for custom builders, and I compared it against the off-the-shelf tools in the proposal software post.
- Open the portal. Give signed clients a login. The year after the signature is where referrals come from, and the portal is where that year gets managed.
- Then, and only then, worry about design. Once the site does the five jobs, make it look the way the company deserves to look. Cosmetics on top of a machine compound. Cosmetics on top of a brochure evaporate.
That order is the same one I use when I build the whole stack for a client, which is laid out on the services page. Site, then CRM, then proposals, then portal. Each layer starves without the one under it.
The website was never the product. It is the front door to a business that either has the machinery behind it or does not. 785 sites told me most do not. The eleven checks will tell you which side you are on. Where the site fits in the rest of a builder's marketing is in the full 785-site study, and what to do with a season's budget once the door works is in the marketing plan.