The math contractors actually need to run
Forget "websites are important." Run four numbers you already have sitting in your head or your invoicing software: average job value, jobs booked per month from the web, close rate on web leads versus referral leads, and what you are currently spending on the site (hosting, a marketing retainer, a DIY builder subscription). Everything else is noise.
Start with average job value. A roofer replacing shingle roofs at $9,000 to $14,000 a job needs a lot fewer web leads to justify a rebuild than a handyman doing $250 service calls. An HVAC contractor selling $8,000 to $12,000 system replacements is in the same bracket as roofing. A plumber doing mostly $150 to $400 service calls with the occasional $6,000 repipe has a blended number that matters more than either extreme.
Next, be honest about your current close rate on web leads. If a lead fills out your form and you close 1 in 4, and your average job nets $8,000 in revenue, one extra qualified lead a month is worth $2,000 a month in booked work, whether or not it shows up as "profit" on a spreadsheet. Multiply that by 12 and a rebuild that costs a few thousand dollars pays for itself inside a year on volume most owners underestimate because they never counted the leads that bounced off a broken form or a page that never loaded on a jobsite phone.
Then look at what you are already spending. A lot of established contractors are paying $99 to $300 a month for a DIY builder or a "we'll host it for you" arrangement that has not produced a form submission in months. That is not a free site. That is a recurring bill for something that does not work. Compare that ongoing cost, over 24 to 36 months, against a one-time build cost for something hand-coded and fast. Often the math is not "spend more," it is "stop bleeding money on something broken and put it toward something that works."
- Average job value times plausible missed jobs per year = the upside
- Current monthly site cost times 24-36 months = the sunk cost of doing nothing
- Load time and form reliability on a jobsite phone = the leak you cannot see in a spreadsheet
The five signs your current site is losing you jobs
Owners rarely know their site is a problem until someone tells them. Here are the signs that show up over and over in this trade, regardless of whether you are roofing, HVAC, plumbing, electrical, or landscaping.
- It is slow on a phone. Homeowners are calling from a driveway after a storm, or from a kitchen with a leak already running. If your site takes more than a couple seconds to load on a phone connection, a meaningful chunk of those people bail before they ever see your number. Under 2 seconds is the target, not a nice-to-have.
- It has one page for everything you do. A single "Services" page that lists 12 trade lines in a bulleted paragraph does not read as credible to a homeowner searching for "emergency water heater replacement" or to the AI tools now answering that search. Each real service and each city you actually run trucks to deserves its own page.
- The quote form does not work, or does not exist. This is the single most common thing found on a contractor site that has not been touched in years: a contact form wired to an email address that bounces, or a form with no confirmation, so the owner has no idea leads are vanishing.
- It was built by a relative, an employee's cousin, or a template five years ago. Nothing wrong with how it happened. But software rots, browsers change, and a site frozen in 2021 design and 2021 information is not representing a business that has grown since then.
- You cannot find your own service pages by searching your own trade and city. Search "[your trade] [your city]" from your phone right now. If you are not on page one, or if what shows up is a directory listing instead of your own site, that is the gap costing you jobs.
Any one of these alone might not justify a rebuild. Two or three together usually do. The pattern worth watching for is compounding: a slow load makes a homeowner impatient right as they hit a page with no clear service match, and by the time they reach a form, they have already half-decided to call the next name on the list instead. Each weak point on its own loses a few visitors. Stacked together, they lose most of them.
One more sign worth naming separately because it is easy to miss: a site that looks fine on a desktop monitor but was never actually tested on a phone in real conditions. A lot of sites built five or more years ago were designed desktop-first, with mobile treated as an afterthought that "should just work." Pull it up on your own phone standing outside, not on office wifi, and see what actually loads.
When the current site is fine and a rebuild is not the answer
We say no to bad fits, and this is one of the honest ones: if your site already loads fast, already books jobs, and already has dedicated pages for your services and cities, a rebuild is not what you need. Spending money to replace something that works is not strategy, it is just spending.
Signs your current site does not need replacing: it loads in under 2 seconds on a phone, it has real pages for your actual services (not one page trying to cover everything), the quote form reliably lands in an inbox someone checks, and you can point to jobs that came from it in the last few months. If that describes your site, the better spend is usually on the work that happens after the build: ranking those pages higher, tightening up your Google Business Profile and map pack presence, or getting picked up as an answer in AI search tools. That is a different kind of work than rebuilding the asset itself, and it is where a site that is already solid should be putting its next dollar.
The other case where a rebuild is premature: if you are not yet running the business at a volume where a handful of extra jobs a year moves the needle, or if you are testing a new service line before committing to it, a lighter touch (a single new page, a fixed form) can close the gap without a full rebuild. Not every problem needs the biggest tool in the shed.
The test is simple. Pull up your site on your phone standing in a driveway with one bar of signal. If it loads fast, looks current, and the form works, you are probably fine. If you are staring at a spinner, or squinting at a page built for a desktop in 2015, you already have your answer.
What changes for AI search, and why it moves the math
Here is the variable that did not exist five years ago and now tips the math for a lot of established contractors: homeowners are increasingly asking ChatGPT, Perplexity, and Google's AI Overviews "who should I call for a roof leak in [city]" instead of scrolling ten blue links. Those tools do not read a site the way a human skims it. They pull structured, specific information: what you do, where you do it, what it costs in general ranges, and whether the page answers the actual question being asked.
A site built as a single page-builder template with vague paragraph copy gives an AI answer engine almost nothing to quote. A site with real, separate pages for each service and each city, built with clean structure instead of bloated page-builder code, gives it plenty. This is not a separate marketing campaign, it is a property of how the site itself is built and coded. A hand-coded static site loads faster and structures information more cleanly than a stack of page-builder plugins, and that structure is exactly what both search engines and AI tools are reading.
To be clear about the boundary: getting quoted by AI search on an ongoing basis, tracking how often it happens, and running that as a continuing visibility program is a separate service. Building the site itself so it is AI-readable from day one, the way the pages are structured, the way service and location content is separated instead of jammed into one page, is part of the build. If your current site cannot be read cleanly by these tools because it is one page of paragraph soup, that is a build problem, and it is part of why a rebuild moves the math in your favor right now in a way it might not have three years ago.
What a rebuild actually costs versus what it returns
Real ranges, not a sales pitch: a hand-coded contractor site with dedicated service pages, service-area pages, and a working quote form runs into the low thousands for a single-trade contractor and scales up from there with the number of services and cities involved. Compare that one-time number to two things: what you are currently paying for whatever exists now, and what a single missed job costs you.
| Scenario | Average job value | Jobs to break even on a rebuild |
|---|---|---|
| Roofing (full replacement) | $9,000-$14,000 | Often 1 job |
| HVAC (system replacement) | $8,000-$12,000 | Often 1 job |
| Plumbing (blended service + repair) | $300-$2,500 | 2-4 jobs |
| Electrical (blended service call) | $250-$1,200 | 3-6 jobs |
| Landscaping (design/build project) | $3,000-$10,000 | 1-2 jobs |
These are not promises about what your site will produce. They are the arithmetic that matters: how few jobs a rebuild needs to close the loop compared to how many jobs you are likely losing right now to a form that does not work or a page that never shows up. For higher-ticket trades like roofing and HVAC, one job often covers the entire cost of a rebuild. For lower-ticket, higher-frequency trades like plumbing and electrical service calls, it takes a few more, but those trades also tend to get more total web traffic, so the volume is usually there.
The other side of the ledger: ongoing hosting and maintenance on a hand-coded static site is minimal compared to a WordPress site with plugins that need constant updates, security patches, and a developer on retainer just to keep it from breaking. That is money you stop paying every month, not just a one-time expense you take on.
How to decide without guessing
You do not need a marketing degree to make this call. You need three answers, and you can get all three in about 15 minutes.
- Load your site on your phone, on cellular data, not wifi. Time it. If it takes more than 2 to 3 seconds to show the page, you are losing mobile visitors before they see anything, and mobile is the majority of contractor search traffic.
- Search your own trade and city from your phone. "[Your trade] [your city]" and a couple of your actual service names. See where you land, and see what shows up when you ask ChatGPT or another AI tool the same question conversationally.
- Submit your own quote form. Actually fill it out and see if a confirmation shows up and where the lead lands. If you cannot verify it works, assume it does not.
If all three come back clean, fast load, you show up, the form works, put your next dollar into the work that happens after the build (ranking, map pack, AI search visibility) rather than a rebuild you do not need. If any one of the three fails, run the job-value math from the top of this guide. Most established contractors find the answer is not close: a site that is not working costs more in invisible lost jobs than a rebuild costs to fix it.
Do this test on more than one page, not just the homepage. A homeowner rarely lands on your homepage first, they land on whichever service or city page a search engine or an AI tool decided answers their question, which might be three levels deep in your site. If your homepage loads fine but your actual service pages are slow, missing, or thin, the homepage test alone will give you false confidence.
An audit takes the guesswork out of all three steps at once. We look at load speed on real mobile conditions, whether your services and cities each have their own page instead of one catch-all, whether the form actually delivers a lead to an inbox someone checks, and whether your site gives an AI search tool anything specific enough to quote back to a homeowner. You get a plain answer back: rebuild, patch, or leave it alone, with the reasoning behind it instead of a sales pitch. Audit delivery runs 1 to 3 business days, and there is no obligation attached to reading it.