13 reasons roofing websites don't convert (and the fix for each)
Most roofing companies do not have a traffic problem. They have a leak problem — homeowners land on the site, find nothing that answers the question in their head, and go back to the search results. This is the diagnostic list we run on every roofing site we take over: the thirteen leaks, the fix for each, the effort it takes, and the number that proves it worked.
A homeowner notices a stain on the bedroom ceiling. They pick up a phone, search roof leak repair plus their town, and open three or four contractor sites in tabs.
They will spend somewhere between eight and forty seconds on yours. In that window they are answering four questions: do you work here, do you do this exact job, have you done it for people like me, and how fast can you come. Every reason on this list is a version of the same failure — the site does not answer one of those four questions fast enough, so the visitor answers it by leaving.
What makes roofing brutal is the arithmetic. A roofing lead bought through paid search runs $350–$500, the sales cycle is roughly thirty days, and demand arrives in weather-driven spikes you cannot reschedule. A site that converts at 2% instead of 6% is not "underperforming" — it is paying three times the market rate for every job it books, in the weeks when it can least afford to.
Do not read it and nod. Open your own site on your phone, on cellular data, standing outside, and work down the diagnostic index below. Each reason is written as symptom → why it costs you → the fix → the metric that proves it. Fix the four marked high first; they carry most of the money.
The diagnostic index
Thirteen leaks, ranked by how much revenue they typically move for a 5–50 person roofing company. Effort assumes you have a WordPress site and no developer on staff.
| # | The leak | Impact | Effort |
|---|---|---|---|
| 1 | The headline is about you, not the roof | High | 1 hour |
| 2 | Stock photography of somebody else's roof | Medium | Half a day |
| 3 | The phone number is not tappable | High | 30 minutes |
| 4 | The site is slow on a phone in a driveway | High | 1–2 days |
| 5 | The form asks nine questions | Medium | 1 hour |
| 6 | One contact page for every kind of job | Medium | 2–3 days |
| 7 | The reviews live on Google, not on the site | Medium | Half a day |
| 8 | No real service area pages | Medium | 1 week |
| 9 | Price and financing are hidden | Medium | Half a day |
| 10 | Insurance claims are not addressed | Medium | 1–2 days |
| 11 | Nobody answers what the site produces | High | 2 days |
| 12 | The gallery proves nothing | Lower | Half a day |
| 13 | You cannot tell which pages produce jobs | Medium | 1 day |
The headline is about you, not about the roof
Open almost any roofing site and the first line of text is a variation of the same sentence: "Family owned and operated since 1998. Quality workmanship. Free estimates." It is true, it is well meant, and it answers a question the visitor never asked.
A homeowner with a stain on the ceiling is not shopping for a company. They are trying to work out whether the person who can stop the water is on this page. "Since 1998" is a trust signal — useful, but it belongs in the third position, not the first. The first position belongs to the job.
Quality Roofing You Can Trust
Family owned since 1998 · Licensed & insured · Free estimates
Roof leak in Frisco? We'll be on your roof today.
Emergency tarping within 4 hours · 1,900 roofs repaired in Collin County · Licensed since 1998
- Write the headline as problem + place + timeframe. All three, in one line.
- Move credentials into a single sub-line under it. They support the promise; they are not the promise.
- Put one specific, countable proof point in the hero — roofs completed, years, or a named neighbourhood — not an adjective.
- Say the next step out loud in the button: "Book a free roof inspection", never "Learn more" or "Submit".
Cover your logo and read the hero to someone outside the company. If they cannot tell you what you do, where, and how fast — the headline has failed, regardless of how good it looks.
Stock photography of somebody else's roof
The smiling model in a clean hard hat on a suspiciously Mediterranean tile roof is the single fastest way to tell a homeowner that this site is a brochure, not a business. Homeowners in your county know what roofs look like there. They notice when yours are not from there.
This is not vanity. Photography does specific conversion work in roofing: it proves the crew is real, proves the work is local, and lets a homeowner picture their own street. A slightly imperfect phone photo of your crew on a real roof in a recognisable neighbourhood outperforms a perfect stock image every time, because it is evidence rather than decoration.
- Give every crew lead a two-minute job: three photos per job — before, during, after — plus one wide shot with the house and street visible.
- Replace all stock imagery on the homepage, the service pages and the gallery. All of it.
- Name the location in the file name and the alt text:
shingle-replacement-mckinney-tx-after.jpg. It helps local search and it helps screen readers. - Include at least one photo of the actual humans who show up — in branded shirts, faces visible.
- Compress every image to WebP under 200KB before uploading; see reason four for why this matters more than it sounds.
The phone number is not tappable — and disappears when you scroll
Most roofing traffic is mobile, and most roofing leads are calls, not forms. A homeowner with an active leak does not fill in a form. They tap a number. If your number is baked into a header image, written as plain text, or parked in the footer, you have put a wall between the emergency and the money.
- Make every phone number a real link:
<a href="tel:+19725550134">. Every instance, every page, header and footer included. - Add a sticky mobile bar pinned to the bottom of the viewport with two buttons — Call now and Text us. Roughly 56px tall, always visible, never covering the form's submit button.
- Add "Text us" for real. A homeowner standing in a wet hallway will often text a photo of the ceiling before they will talk. That photo is a qualified lead.
- Use a tracked number so you can attribute calls to pages (reason thirteen), but display the same number consistently in your Google Business Profile and citations.
Do not put the tracking number in your Google Business Profile's main field and a different number on the site. Inconsistent NAP data across citations damages local ranking. Use call tracking's dynamic number insertion, which swaps the number for visitors only.
The site is slow on a phone in a driveway
You test your site on office wifi on a desktop. Your customer opens it on a four-year-old Android on two bars of signal, standing in the rain. Those are different websites.
Google's analysis of mobile page load found that as load time goes from one second to three, the probability that a visitor bounces rises 32%; at five seconds it is up 90%. Nothing else on this list fails as silently — these visitors never appear in your form submissions, so the leak is invisible unless you go looking.
- Run PageSpeed Insights on your three highest-traffic pages and read the mobile tab only. Ignore the desktop score entirely.
- Convert every image to WebP and serve responsive sizes. On most roofing sites this alone is the majority of the problem.
- Audit plugins. Deactivate anything that is not directly earning a lead — sliders, unused page builders, four analytics scripts doing the same job.
- Add page caching and a CDN. On shared hosting, move to managed WordPress hosting; it is usually cheaper than one lost job a month.
- Remove the hero video. It is almost always the single heaviest asset and it almost never converts better than a real photo.
Largest Contentful Paint under 2.5 seconds on a mid-range Android on 4G. Not "green in the lab" — green in the field data, which is what Google actually uses.
The form asks nine questions before it earns one
Name, email, phone, address, city, ZIP, roof type, roof age, "how did you hear about us", and a message box marked required. Every one of those fields is a small decision, and every decision is a chance to close the tab.
The instinct behind the long form is real — you want qualified leads, not tyre-kickers. But qualification is a phone call's job, not a form's job. The form has exactly one job: get you permission to make that call.
Gets you fewer leads, each carrying data you would have gathered on the call anyway — and skews toward people with time on their hands rather than people with an active problem.
Name and phone, plus one tap: Leak / repair · Full replacement · Storm or insurance · Not sure. That single choice routes the lead and tells your estimator how to open the call.
- Cut to name and phone. Make email optional, or drop it — for roofing, phone is the channel that books work.
- Add one tap-to-select field for job type. It doubles as routing logic and as intake data.
- Add an optional photo upload. "Send us a photo of the damage" converts anxious homeowners and pre-qualifies the job better than six text fields.
- Replace "Submit" with what happens next: "Get my free inspection". Then say when: "We call back within 15 minutes, 7am–8pm."
- Test the form end-to-end from a phone on cellular data, and confirm the notification actually reaches a human — not a shared inbox nobody opens.
One contact page for a leak, a replacement and a storm claim
These are three different people with three different emotional states, three different budgets and three different timelines. A homeowner with water coming through the ceiling and a homeowner planning a replacement next spring should not land on the same page — and definitely should not land on the same generic "Contact Us".
Build one dedicated page per intent, each with a headline that repeats the visitor's own words back to them:
Emergency leak repair
Response time above the fold, tarping promise, "we're on call tonight", phone-first. No long copy — this person is scanning.
Roof replacement
Material comparison, price ranges, financing, warranty, timeline, photos of completed jobs on similar homes. This person is researching.
Storm & insurance claim
What the adjuster process looks like, what you do and don't handle, deadlines, a free inspection offer. This person is anxious and confused.
- Point each ad group and each Google Post at its matching page — never at the homepage.
- Keep one primary action per page. Two competing buttons halve the strength of both.
- Reuse the same nine-element skeleton on all three so they stay fast to build and easy to maintain.
Your reviews live on Google, not on your website
You have 190 reviews at 4.8 stars. It took years to earn them. And on your website there are three unattributed quotes in italics that could have been written by anyone, because they probably were.
Roughly eight in ten consumers use Google to read reviews of a local business. That means the homeowner comparing you against two competitors is doing that work anyway — the question is whether they do it on your page, where you control the framing, or in a Google tab where your competitors are sitting one scroll away.
- Put the live rating and review count in the hero: 4.8 ★ · 190 Google reviews. Link it to your profile — sending them to proof you control is not a leak.
- Pull real reviews in automatically with a widget so the wall stays current without anyone remembering to update it.
- Show first name, town and date on every quote. "Sarah M., Plano — March 2026" is credible; "S.M., satisfied customer" is not.
- Place one relevant review beside every form and every price block — the moment of hesitation is where proof does its work.
- Add
LocalBusinessandAggregateRatingschema, and keep it consistent with what is displayed. - Reply to every review, including the bad ones. Prospects read the replies more carefully than the reviews.
"We serve the greater metro area" — and no page proves it
Roofing is bought locally and searched locally. The homeowner types their town. If your site has no page that says that town, Google has nothing to match and the homeowner has nothing to recognise.
Both fail, for opposite reasons. A list is too thin to rank. Thirty spun duplicates are thin and risk being treated as doorway pages. The version that works is fewer pages with real local substance in each.
Build five to eight genuine pages — one per town that actually earns revenue — and give each one:
- Three or more completed jobs in that town, with photos and a one-line description of the work.
- Two or more reviews from customers in that town, named.
- Something only a local would write: the common roof types on those streets, the hail history, the permit office, the HOA rules that matter.
- Response time and coverage for that specific area, stated plainly.
- A unique title, meta description and H1 built around roofing contractor in [town].
Do not generate thirty pages from a template with the town name swapped in. It is the most common roofing SEO mistake, and the cheapest one to be penalised for. Five real pages beat thirty hollow ones.
Price and financing are hidden, so the visitor assumes the worst
The first question in every homeowner's head is what this is going to cost. Roofing sites almost universally refuse to answer it, on the theory that every roof is different. Every roof is different. Every homeowner still needs a range before they will let a stranger onto their property.
When you give no number, the visitor fills the gap with a guess — and the guess is usually too high, which kills the lead, or too low, which produces a wasted estimate visit. A published range does not cost you negotiating room. It filters.
- Publish honest ranges by scope: repair, partial, full replacement — by material, for a typical home size in your market.
- Explain the four things that move the number: square footage, pitch, layers to tear off, material. Now the range is a lesson, not a dodge.
- Put financing where the sticker shock lands, not in the footer: "From $X/month" beside the replacement range, with the term and the approval time.
- Add a simple instant estimator if you can support it. Even a rough one converts, because it trades a number for a phone number.
- Say what is included — tear-off, disposal, permits, warranty — so the range is comparable to the competitor's quote.
The insurance claim — the biggest question in roofing — is not addressed
After a hail or wind event, a large share of roofing revenue runs through an insurance claim. The homeowner has never done this before. They do not know whether to call you or the carrier first, what a deductible means for them, how long they have to file, or what an adjuster meeting involves.
Whoever answers those questions clearly becomes the trusted party — usually before any estimate is given. This is the highest-leverage content on a roofing website and it is missing from most of them.
- Write one page: How a roof insurance claim works in [state], in plain language, step by step, from damage to final payment.
- Cover what homeowners are actually afraid of: will my premium rise, what is depreciation, what is my deductible responsibility, how long do I have to file.
- State clearly what you do and do not do. Legal and ethical lines vary by state — describe your role honestly rather than implying you can negotiate the claim for them.
- Offer the low-commitment step: a free documented storm inspection with photo evidence they can use, whether or not they hire you.
- Add a short FAQ block with schema so these questions can surface directly in search results.
Storm work is timed. When a hailstorm hits, thousands of homeowners in your county search the same five questions inside 72 hours. A page that already ranks for those questions is the cheapest lead source in roofing — but only if it was published before the storm.
The website works. Nobody answers what it produces.
This is the one that hurts, because it is not a website problem at all — and it invalidates every fix above it. You can double your form fills and book fewer jobs, if the response happens tomorrow.
The research here is old, cross-industry and consistent: roughly 78% of customers buy from the business that responds first, and a lead contacted within an hour is far more likely to qualify than one contacted a day later. In roofing the effect is sharper still, because the competing tab is already open.
- Set a five-minute rule for every inbound lead during working hours, and write it where the whole team can see it.
- Turn on an instant automated SMS the second a form is submitted: "Got it — [Name] is calling you in the next few minutes." It buys you the window.
- Route leads to a phone, not an inbox. Push notification to the person who calls, with the job type from the form already attached.
- Solve after-hours properly — on-call rota, answering service, or an AI voice agent that can take details and book an inspection slot. Then test it yourself at 9pm.
- Add missed-call text-back so no unanswered call dies silently.
- Log median response time weekly. Not "we're pretty fast" — the number.
The gallery is forty photos and proves nothing
A grid of roofs with no captions is wallpaper. The homeowner cannot tell which one was a repair and which was a replacement, what it cost, how long it took, or whether any of it happened near them.
Turn the gallery into short project records. Six well-documented jobs beat sixty photos:
- Town, roof type and material — "Architectural shingle replacement, 2,400 sq ft, Allen TX".
- The problem, in the homeowner's words. The fix, in yours. Days on site.
- Before and after in the same frame and the same light.
- The customer's review attached to the job it belongs to.
- Whether it ran through insurance — this is the detail storm-damage visitors search for.
Then link each project from the matching service area page (reason eight). The same asset does two jobs: proof for the visitor, local relevance for search.
You cannot tell which pages produce jobs — so you keep guessing
Ask most roofing owners which page brings in the most revenue and the honest answer is a shrug and a theory. Without measurement, every fix on this list is a matter of opinion, and the next twelve months of marketing spend is allocated by feel.
- Install GA4 and define real conversion events:
call_click,form_submit,text_click,estimate_start. - Add call tracking with dynamic number insertion, so calls attribute to pages and sources without breaking your NAP consistency.
- Connect leads to your CRM or job board so you can follow a lead through to booked and to invoiced.
- Build one weekly scorecard: leads by source, cost per lead, lead-to-booked rate, median response time, revenue per source.
- Review it for twenty minutes every Monday. Kill what does not book work. Fund what does.
Revenue per source, not leads per source. A channel producing forty cheap leads that book three small repairs is worth less than one producing six leads that book two replacements — and only this metric shows it.
Bonus: your site is invisible to the AI answer engines
A rising share of "who should fix my roof" research now happens inside AI assistants and AI overviews rather than in a list of ten blue links. Those systems quote pages that state facts plainly, in structured form, with clear entities — and skip pages built entirely from marketing adjectives.
- Answer real questions in question form, with the answer in the first two sentences beneath the heading.
- Keep entities explicit and consistent: legal business name, service area towns, materials, certifications, license numbers.
- Use
LocalBusiness,Service,FAQPageandAggregateRatingschema, and keep it truthful. - Publish original local facts nobody else has — your hail history, your permit timelines, your average job durations. Original data is what gets cited.
- Check monthly what an AI assistant says when asked for a roofer in your town, and fix the gaps it reveals.
Bonus: every visit is a one-off, because you sell nothing recurring
A roof is bought roughly once every twenty years. That makes roofing structurally dependent on new leads forever — unless you sell something that keeps the relationship alive between roofs.
- Package an annual inspection and maintenance plan: two visits, gutter check, photo report, priority scheduling after storms, a small discount on repairs.
- Give it its own page with a price, and offer it on every completed job — replacement and repair alike.
- Sell priority scheduling as the real benefit. After a hailstorm, "we come to plan holders first" is worth more than any discount.
- Automate the reminders, the report delivery and the renewal. Two people cannot run this manually at scale; software can.
Recurring revenue does two things at once: it smooths the winter, and it makes every future replacement a warm conversation rather than a cold lead you have to buy again.
What the leaks are costing you a year
Conversion work is not a design opinion, it is a revenue calculation. Put your own numbers in. "Revenue per lead" is your average job value multiplied by the share of leads you close — if you close a third of leads at an average of $9,000, that is $3,000.
6% is used as the target because it sits inside the 3–7% band reported for roofing and below the 7.8% home-services average — an achievable ceiling for a well-built site, not a best case. Nothing here is a guarantee; it is the size of the prize, which is what decides whether the work is worth doing.
Run the first seven checks now
Everything above, condensed into a working audit. Open your site on a phone, on cellular data, and tick as you go. This is the first of the audit's five sections — the other thirty-three checks are on the audit page.
That is the first section of five. The full forty checks — mobile and speed, forms and intent paths, proof and price, response and measurement — live on their own page, where the list scores itself as you tick and remembers where you got to.
Tappable phone number with a sticky mobile bar · a hero headline that names the problem, the town and a timeframe · a two-field form · and a five-minute response rule that survives after 5pm. Those four carry most of the money on this page, and none of them takes a week.
Want to know which of these thirteen is costing you the most?
We'll run this audit on your roofing site and send you a short recorded teardown — the leaks we find, ranked by what they are worth, with the fix for each. No pitch deck, no obligation, and you keep the findings whether or not you work with us.
Sources
- WebFX — Home Services Marketing Benchmarks (7.8% average conversion rate; 3–7% for roofing; $350–$500 cost per lead for premium services; ~30-day roofing sales cycle)
- Think with Google — Mobile Page Speed Industry Benchmarks (bounce probability +32% at 3s, +90% at 5s, +123% at 10s)
- BrightLocal — Local Consumer Review Survey (share of consumers using Google to read local business reviews)
- LocaliQ — Home Services Search Advertising Benchmarks
- Lead response research: MIT / InsideSales Lead Response Management Study, and The Short Life of Online Sales Leads, Harvard Business Review (2011)
Response-time and page-speed figures are cross-industry benchmarks, not roofing-specific measurements. The funnel diagram, price examples and calculator defaults are illustrative. Use them as structure, and validate every number against your own analytics before making a spending decision.