The Five Questions That Actually Decide It
Owners ask us to "just redesign the site" the way they'd ask a painter to "just redo the trim." Sometimes that's the right job. Sometimes the trim is the least of it and the sheathing underneath is rotten. Before anyone quotes a rebuild, we run the same five questions on every site.
- Does it load under 2 seconds on a phone? Test it on a job site with two bars of signal, not your office wifi.
- Is it built on something you can still edit? A dead page builder, an abandoned Wix trial, or a WordPress install nobody's patched since 2022 is a liability, not an asset.
- Does every trade and service area you actually work have its own page? One "Services" page listing eight trades in a bullet list doesn't get quoted for any of them.
- Does the quote form actually deliver leads to an inbox someone checks? We've found forms silently failing for months. Nobody noticed because nobody was watching.
- Can an AI answer engine read the page and figure out what you do, where you work, and how to reach you? Flashy homepage sliders and stock-photo hero banners are invisible to that kind of read.
Score two or fewer "yes" answers and you're not looking at a redesign. You're looking at a rebuild. Score four or five and the fix is probably targeted: a new form, a speed pass, a couple of missing pages. That's a different, smaller job, and we'll tell you that on the call instead of quoting you a full rebuild you don't need.
The mistake we see most: an owner spends money re-skinning the same broken foundation. New colors, same slow page builder, same missing job pages. Six months later the phone still isn't ringing and they've paid twice for the same result.
Signal: The Site Is Slow
Speed isn't a nice-to-have. It's the first filter a homeowner (and increasingly, an AI search tool crawling for an answer) applies before they ever read your copy. If your site takes four, five, six seconds to load on a phone, a chunk of your traffic is gone before the hero image finishes painting.
Page-builder platforms (Wix, Squarespace, most WordPress theme-and-plugin stacks) load a mountain of third-party script to render a page that could be plain HTML and CSS. Every plugin is another request, another script, another point of failure. Contractor sites we've inherited from these platforms routinely load in the 4-8 second range on mobile. That's not a tweak-able number. It's architectural.
We build hand-coded, no-WordPress, no page-builder sites specifically because that's the only reliable way to hit under 2 seconds consistently. There's no plugin update that breaks it at 11pm. There's no page-builder bloat to strip out later. If your current site is on one of these stacks and it's slow, redesigning within the same platform usually buys you a smaller number, not a fixed one. That's a rebuild conversation, not a patch.
How to check it yourself before you call anyone: load your homepage on your phone with wifi off, using cell data. Time it with a stopwatch. If it's over 3 seconds, that's your answer regardless of what else is wrong.
- Under 2 seconds: speed isn't your problem. Look elsewhere on this list.
- 2-4 seconds: borderline. Combined with other red flags below, lean rebuild.
- 4+ seconds: rebuild. No patch fixes a platform-level speed problem.
Signal: It's Built on a Page Builder or Dead Platform
Ask who built the site and what it runs on. If the answer is "a guy I found on Facebook Marketplace" or "a $99/month builder I've been meaning to cancel" or "WordPress, but I don't know the login," that's diagnostic. The platform itself is the problem, not any one page on it.
These builds tend to share a pattern: a template pulled from a stock library, generic stock photography, one "Services" page instead of dedicated pages per trade, and a contact form wired to an email address that changed two office managers ago. None of that is fixable with a color change. It's fixable by starting over on a foundation you actually own and can edit.
WordPress itself isn't the enemy. A poorly maintained WordPress install is. Every plugin update is a chance for something to break at 11pm on a Sunday, and most contractors don't have someone watching for it. We've inherited sites where a plugin conflict took the contact form down for weeks and nobody noticed until a homeowner mentioned it in passing. That's the maintenance tax a page-builder or CMS stack carries that a hand-coded static site doesn't.
A hand-coded static site has no login to lose, no plugin to go stale, no monthly builder fee that quietly doubles. It's a file. You own the file. That's the trade-off we make explicit on every strategy call: slightly less drag-and-drop convenience in exchange for a site that still works, and still loads fast, five years from now without anyone touching it.
If your current site is on a platform you don't control (no login, no export, agency holds the keys), that alone is grounds for a rebuild conversation even if the site looks fine today. You can't redesign what you can't access, and you shouldn't have to beg a former web guy for a password to update your own price list.
Signal: Missing Trade Pages, Missing Service-Area Pages
This is the one owners underestimate most. A single "Services" page listing roofing, siding, and gutters in three bullet points does not get quoted by Google, and it does not get quoted by ChatGPT or an AI Overview either. Both need a dedicated page per job type, with specifics: what the job involves, what it costs in ranges, what the process looks like, and proof.
Same logic applies to geography. A roofer working three counties needs pages for each city or region he actually services, not one "Service Area" paragraph buried in the footer. Search engines and AI answer tools match intent to specific pages. "Metal roof replacement Kissimmee" needs a page that answers that exact question. A generic homepage can't hold that many specific answers at once.
Run this check: count how many distinct job types and how many distinct service areas your business actually covers. Now count how many dedicated pages your site has. If the gap is wide, the fix isn't editing your existing pages. It's building the missing ones, which in practice means a rebuild of the site architecture, not a repaint of what's there.
We build out from a trade-specific hub structure so each trade gets its own money page, with real job pages and service-area pages branching from it. That's a page-count and architecture decision made at build time, not something you bolt onto an existing three-page template after the fact.
| What a thin site has | What gets quoted |
|---|---|
| One "Services" page, bulleted list | A dedicated page per trade, per job type |
| One "Service Area" paragraph | A page per city or county actually served |
| Stock photo, generic copy | Specific process, specific proof, specific numbers |
Signal: Forms Fail Silently, Content Is Stale
Sometimes the site is fine and the problem is narrow. We've opened up sites where the lead form has been broken for months and nobody knew, because the failure is silent: the visitor sees a "Thank you," the email never lands, and the owner just thinks the phone's been slow. That's not a rebuild. That's a broken wire, and it should get fixed this week regardless of what else you decide.
Stale content is the other narrow case. A price list from 2019, a team photo missing three current guys, a "now serving" list that hasn't caught up to the counties you added last year. None of that requires new architecture. It requires someone to go in and update the copy, which is a content job, not a redesign job.
The test for "leave it alone, just fix this": is the underlying platform fast, editable, and structured with real pages per trade and area? If yes, and the only complaints are a broken form or old content, you don't need us to touch the build. You need five minutes with whoever manages the site, or a short fix-it engagement instead of a full rebuild quote.
- Test your own form: submit it right now and confirm the email actually arrives.
- Check your service-area list against the counties you actually invoice in this year.
- Check your pricing or process copy against what you tell homeowners on the phone today.
Be honest with yourself here. It's tempting to blame the whole site when the real culprit is one dead form. That mistake costs money twice: once for a rebuild you didn't need, and again because the actual bug (the form) doesn't get looked at until someone finally opens the code.
Signal: AI Search Can't Read It
This is the newest signal and the one most existing contractor sites fail without anyone noticing, because nothing looks broken. The site loads. It looks fine on a laptop. But when a homeowner asks ChatGPT or an AI Overview "who does metal roof replacement near me," the site never gets mentioned, because the page is a wall of marketing copy with no structured facts an answer engine can lift.
AI search tools favor pages that state things plainly: what the service is, what it costs in a real range, what's included, who it's for, who it's not for, how long it takes. A hero slider with "Quality You Can Trust" over a stock photo gives an AI tool nothing to cite. It has to guess, and it usually guesses wrong or skips you entirely.
This doesn't mean re-running your SEO campaign or chasing map-pack rank (that's ongoing work, and it lives with our SEO and local-search teams, not here). What it means at the build level is structure: does each page carry a plain-facts block, clear headings, and specific answers, or is it all soft marketing language wrapped around a contact form? That structural question is answered at build time. If your current site's pages are all soft-sell copy with no factual scaffolding, a redesign that keeps the same page structure won't fix it. The pages need to be rebuilt to carry the facts, not just repainted.
Check one page yourself: read it out loud and ask whether it actually states a price range, a timeline, and a service area in plain sentences, or whether it's all adjectives. If it's all adjectives, that's a build problem, and it's the kind we solve when we build the site, not the kind an ongoing ranking campaign can patch around.
What a Rebuild Actually Costs and Takes
Once you've run the signals above and landed on "rebuild," the next honest question is what that means in time and money. We won't quote a number here because every trade, page count, and scope is different, and we don't post pricing until we've scoped your actual job on a strategy call. What we can tell you is the shape of the engagement, and what drives the range up or down.
A rebuild means starting from a hand-coded, no-WordPress foundation: a hub page for your business, dedicated money pages per trade you actually run, service-area pages for where you actually work, and working quote forms wired to an inbox you check. It's built to load under 2 seconds and structured so both Google and AI answer engines can read it plainly. That's the deliverable. What happens after launch (ranking work, map-pack work, an ongoing AI-visibility campaign) is separate work that lives with our SEO, local-search, and AI-search teams respectively. This silo is about the asset you own at handoff, not the ongoing campaign.
The two biggest cost drivers are trade count and service-area count, because each one is a real page, not a line item on a spreadsheet. A single-trade contractor working one county is a smaller build than a multi-trade outfit covering six counties, and the quote should reflect that honestly rather than a flat rate that overcharges the small guy or underbuilds the big one. If you want a fuller breakdown of price ranges and what moves them, that's covered in its own guide rather than repeated here.
A targeted fix, by contrast, is a scoped repair: a new form, a speed pass, a handful of missing pages added to an otherwise sound structure. It's a smaller job with a smaller footprint, and if that's genuinely what your site needs, that's what we'll tell you. We'd rather quote the right job than the bigger one, even when the bigger one pays more.
Either way, the process starts the same: a straight look at your current site against the five signals above, on a strategy call, no pressure either direction. Bring the site, bring your trade list, bring your service-area list. That's enough for us to tell you which job you actually need.