What a contractor website needs above the fold

Contractor homepages fail in a predictable way. A full width photo, a slogan across it, and a menu. The visitor learns that you are family owned and committed to quality, which is what every competitor also says.

Meanwhile the two questions that actually decide the visit go unanswered. Do you fix my thing, and do you come to my house.

What are the four things, exactly?

What you do, in the words a customer uses. "Plumbing, drain cleaning, water heaters" beats "residential and commercial solutions." If you specialize, say the specialty. If a homeowner has to interpret your headline, you have lost the four seconds.

Housecall Pro's 2026 Home Services Report, which pairs a homeowner survey with Q1 2026 platform data, found that the most common trigger for calling a contractor is a sudden problem, at 58%, followed by a smaller issue getting worse at 51%, routine maintenance at 47% and planned projects at 38%. Roughly five in nine visitors arrive because something broke. They are scanning for the name of their broken thing, not for your positioning statement.

Where you do it, named. Not a map graphic, not a radius, a list of towns in text. "Serving Santa Rosa, Petaluma, Rohnert Park, and Sonoma." This one line prevents the most common bounce on a contractor site, which is somebody leaving because they could not tell if you drive to them.

The phone number, in text, tappable. Not inside an image. Not only in the header on desktop. Text, large, at the top on a phone.

A second action for people who will not call. An estimate request, a text option, a "get a price" button. Planned work often starts with somebody who does not want to talk yet, and if the only action is a phone number, those visitors leave without a trace. The same Housecall Pro survey found 68% of homeowners get additional quotes when pricing comes in higher than expected, which tells you the comparison behavior is real and it happens in writing.

That is the whole list. Adding a fifth thing makes the four weaker.

What should not be up there?

The slogan. The stock photo of a smiling family. The rotating slider. The chat bubble that covers the phone number on a small screen.

Sliders deserve a specific mention because agencies still sell them. Visitors do not wait for slide two. You have chosen to show your most important message to a fraction of your traffic for no reason.

Awards and badges are a real trust signal but they are not a first screen signal. They work after the visitor knows you do their job in their town, and they compete for attention before that.

If an element does not help a homeowner answer "do they fix my thing in my town," it belongs below the fold.

How much is the load time actually costing you?

More than the design, and this is the part nobody measures because the lost visitor never appears in a conversation.

Google's analysis using a neural network trained on bounce and conversion data, with a stated 90% prediction accuracy, puts it in one sentence: as page load time goes from one second to 10 seconds, the probability of a mobile site visitor bouncing increases 123%. The same page adds a second finding that is more relevant to a bloated hero section than the first one: as the number of elements on a page goes from 400 to 6,000, the probability of conversion drops 95%.

Be careful with that 123% number, because you will also be quoted 113%, and they are the same finding. Google's original February 2017 edition said one second to seven seconds and 113%. The February 2018 update said one second to 10 seconds and 123%. It is one study, re-run and re-published, not two independent confirmations. Anyone presenting both to you as separate evidence has not read either. A third version, "one to three seconds, bounce probability increases 32%," does not appear in either primary document at all.

What does fixing it actually look like?

Does this change for paid traffic?

Yes, and this is where the money is.

Organic visitors arrive with a general question. Paid visitors arrive with a specific one, because you paid for a specific search term. Sending a search for "tankless water heater installation" to a homepage that lists eleven services wastes the click you bought, and in this category the click is expensive. LocaliQ's 2026 benchmarks put the median cost per click for Home and Home Improvement at $8.33 and the median cost per lead at $90.92, second only to attorneys.

The rule is one page per thing you paid for. If you cannot build a page per service, you should not be bidding on that service, which is the same underlying gap as offering a service you never wrote a page about.

This is also where cost per lead starts to mislead people. A cheap click that lands on a vague page produces cheap leads that never close, and the difference between cost per lead and cost per customer is the number that decides whether the channel is working.

The national lead aggregators understand this better than you do and they build a page per service per city. You cannot outspend them, but you can be the local company with a real page, real photos, and a phone that gets answered. How to compete with the national aggregators in your own market is mostly about the things they structurally cannot do.

What about the seasonal version?

Change the first screen when demand changes. Most contractors never do.

In August, an HVAC company's first screen should be about air conditioning. In October it should be about heating. The rest of the site can stay put. This is a fifteen minute edit that most companies skip for years, and the traffic arrives before the weather does, which is why preparing the site before the seasonal spike is a calendar problem more than a design one.

Same logic for capacity. If you are booked three weeks out on installs but have same day capacity on repairs, say so up there. It sorts your traffic for you.

Where does this advice break down?

Three places, and the third one is where money gets wasted.

When your traffic is too small to test any of it. A local business at 300 to 800 monthly visits with a 3% conversion rate produces nine to twenty-four leads a month. Split across two variants, that is single digits each. Work out what a real test would need rather than trusting a quoted floor. Sample size follows the metric's variance and the sensitivity you want, which is the principle Kohavi and colleagues set out in KDD 2014, and it is never a flat traffic number. At 95% confidence and 80% power the two-proportion calculation carries a constant of 2 × (1.96 + 0.84)², about 16, and the conversions required per variant is that constant divided by the square of the relative lift. A 40% improvement in quote requests, which would be enormous, still needs 100 conversions per variant (derived), and 20% needs 400 derived. At twenty-four leads a month, that is years. You are not going to A/B test your hero section. Make the change on judgment, then measure over a long window and say plainly that it was not a controlled test.

When the evidence you are borrowing came from somewhere else. Baymard Institute's checkout research, which found the average flow carries 11.3 form fields and that field count matters far more than step count, is the best available work on form friction. It is also e-commerce checkout research. It supports "fewer fields, better" as a direction. It does not tell you what a plumber's quote form should contain.

When somebody quotes you an industry standard price for the build. There is no survey with a disclosed sample behind any of the "$2,000 to $10,000 to build" ranges in circulation. They trace to agency pricing pages quoting other agency pricing pages. An agency's price is a fact about that agency, not a benchmark.

What to do this week

Open your homepage on your own phone, on cellular, not office wifi. Start a timer. Screenshot what you can see without scrolling.

Show that screenshot to somebody who does not know your business and ask two questions. What does this company do, and would they come to my house. If they hesitate on either, the first screen is failing and no amount of traffic fixes it.

Then run the five step weight check above and record your Largest Contentful Paint before and after.

Then check the number itself. Tap the phone number on your own site and see whether it dials. Broken tel links are more common than anyone expects, particularly after a redesign.

Last, decide what happens to the calls that first screen produces after 5pm. A better homepage means more calls, at hours you did not plan for. Automation can carry some of that, but not all of it, and the limits of what an AI can answer show up faster in trades than most vendors admit.

Questions about your first screen?

Email me at eric@seod.com with your website address. I will open it on a phone, screenshot the first screen exactly as a homeowner sees it, and send it back with the four things marked present or missing. No call, no proposal, and if it already does the job I will tell you what to spend the money on instead.

There is more on contractor sites and lead flow in the home services library.