Skip to content

We Analyzed 50 Small Business Websites: Here Are the 7 Most Common Mistakes

We audited 50 real small business websites and found the same critical mistakes over and over. Here is what we found, and how to fix each one.

An audit of 50 real small business websites turned up the same mistakes repeatedly, from hero sections that never say what the business does to contact forms that quietly lose leads. This post lists each mistake found and the specific fix for it, in the order worth doing them.

Adam WhitakerBy Adam Whitaker · Tips & Trends · 7 min read · Published · Updated

The interesting thing about auditing fifty small business websites in a row is how quickly it stops being interesting. The same seven problems keep appearing. Different industries, different budgets, different platforms, same failures. None of them are exotic. Most were introduced by a well-meaning builder who was never asked to think about what happens after the site launches.

Here they are, in the order that costs the most money, with the fix for each.

1. The homepage never says what you do or where you do it

The single most common defect is a hero section that could belong to any company on earth. 'Quality You Can Trust.' 'Your Partner in Excellence.' 'Building Better Futures.' A visitor lands, spends about five seconds deciding whether they are in the right place, and finds no service and no geography.

This is not only a conversion problem. Search engines and AI answer engines use that same above-the-fold text to work out what the page is about. A hero with no noun and no place name gives them nothing.

The fix: your H1 should contain what you sell and roughly where you sell it. 'Residential Roof Replacement in Salisbury and Wicomico County' is not poetry, but it tells a human and a crawler the same true thing in one line. Put the sub-headline underneath for the differentiator, and keep the aspirational language for the About page where it belongs.

2. There is no obvious next action

Sites routinely offer a visitor six equally weighted options and no primary one. Or the phone number sits in the header as plain text on mobile, so tapping it does nothing. Or the only call to action is 'Learn More', which asks the visitor to do work rather than telling them what happens next.

The fix is boring and effective. Pick one primary action per page. Make it visually dominant. Repeat it at natural decision points rather than once at the bottom. On mobile, wrap the phone number in a tel: link so a tap dials, and consider a persistent bottom call bar for service businesses where phone is the dominant channel. Replace 'Learn More' with the specific outcome: 'Get a free roof inspection' or 'Book a 15-minute call'.

One more detail that gets missed: every page needs a next action, not just the homepage. Blog posts and service pages that dead-end are where a large share of qualified traffic quietly leaves.

3. The site is slow, and it is almost always the images

Google's Core Web Vitals thresholds are public and specific. Largest Contentful Paint should be under 2.5 seconds. Interaction to Next Paint should be under 200 milliseconds. Cumulative Layout Shift should be under 0.1. These are measured at the 75th percentile of real visits, on real devices, which for most local businesses means mid-range Android phones on cellular connections rather than the designer's laptop.

The usual culprit is a hero image exported straight from a camera or a stock library at three or four megabytes. Second most common is a page loading five or six third-party scripts before anything renders: a chat widget, two analytics tags, a review carousel, a font service and a cookie banner.

  • Serve images as WebP or AVIF, sized to the largest slot they actually occupy, and use responsive srcset so phones do not download desktop assets.
  • Set explicit width and height attributes on every image so the browser reserves space and layout does not shift.
  • Preload the LCP image and lazy-load everything below the fold.
  • Defer non-critical third-party scripts, and audit whether you still need all of them. Chat widgets in particular are frequently installed, forgotten and never staffed.
  • Self-host fonts, subset them, and cap yourself at two families and three weights.

4. One page is trying to rank for everything

A very common structure is a homepage listing eight services, one thin Services page listing the same eight, and nothing else. The business then wonders why it does not rank for any of them.

Search engines rank pages, not businesses. If drain cleaning, water heater installation and sewer line repair all live in one paragraph each on a shared page, no page is the definitive answer for any of those queries, and your own pages compete with each other for the same terms.

The fix is a real information architecture: one substantial page per service, 700 words minimum, with its own title, meta description, FAQ block and internal links. Where a service and a market both have real search demand, add service-by-location pages beneath them. The discipline that matters is that each page has to earn its existence with genuinely different content. Spinning up thirty near-identical pages that differ only by a town name is how sites get classified as doorway pages and dropped.

5. Metadata and indexation are broken in ways nobody checked

This category is dull and it is responsible for an enormous amount of invisible traffic loss. The recurring patterns:

  • Title tags that are missing, duplicated across pages, or set to the page name plus the site name and nothing else. Aim for roughly 50 to 60 characters, unique per page, with the primary term near the front.
  • Meta descriptions absent or auto-generated. Target 140 to 158 characters and write them as ad copy, because that is what they are.
  • Multiple H1s per page, or heading levels chosen for font size rather than hierarchy.
  • A canonical tag on every page pointing at the homepage. This is startlingly common on template builds and it tells Google that none of your inner pages deserve to exist.
  • No XML sitemap, or a sitemap that omits the very pages the business is investing in.
  • A staging noindex tag that shipped to production and was never removed.

The fix is a one-hour audit and then a permanent habit. Verify the site in Google Search Console and Bing Webmaster Tools, submit a sitemap generated from the actual route list so it cannot drift, and check the Pages report monthly for anything in 'Crawled, currently not indexed' or 'Duplicate without user-selected canonical'.

6. The contact form loses leads

Nothing on this list is more expensive per incident. A form that silently fails, or that delivers to a mailbox nobody monitors, converts marketing spend directly into nothing.

The specific failures worth checking: submissions routed to a personal address that later got filtered to spam; a plugin that stopped sending after a hosting change and produced no error; twelve required fields on a mobile form for what should be a name, phone and one sentence; no confirmation message, so the visitor submits twice or assumes it broke; no spam protection, so real leads drown in bot noise and the owner stops reading the inbox.

  • Cut required fields to the minimum that lets you follow up. Name, phone or email, and a short description. Qualify on the call, not on the form.
  • Send submissions to two destinations, ideally a monitored inbox plus a CRM or database record, so a single mail failure does not lose the lead.
  • Redirect to a real thank-you page rather than swapping in an inline message. That gives you a clean conversion event to track.
  • Add invisible spam protection rather than a puzzle. Honeypot fields plus a scoring service beats making customers identify traffic lights.
  • Test the form yourself on a phone, on cellular, once a month. Put it in the calendar.

7. There is no proof that any of it is real

The last mistake is the one owners resist fixing because it takes effort outside the website. The site says the company is trusted, experienced and local, and offers nothing to support any of it. Stock photography of models in hard hats. No reviews. No named clients. No address. No photographs of actual completed work.

Buyers discount unsupported claims almost completely, and AI answer engines have nothing to corroborate. Proof is also the cheapest differentiator available to a small business, because most competitors are equally lazy about it.

  • Put real photos of real jobs on service pages, with descriptive filenames and alt text. Phone photos beat stock imagery.
  • Pull live reviews onto the relevant page and mark them up with valid review schema. Never fabricate ratings markup.
  • Publish a full name, address and phone number in the footer, formatted identically to your Google Business Profile.
  • Name clients where you have permission, and describe the work in specifics rather than adjectives.
  • Show licence numbers, insurance status, manufacturer certifications and association memberships. In trades, these are trust signals and also backlink opportunities.

What to do with this list

Run through the seven in order and be honest. Most sites we look at fail at least four. The encouraging part is that the fixes are cumulative rather than competing: clarifying the hero helps conversion and retrieval, fixing images helps Core Web Vitals and mobile bounce, adding proof helps buyers and AI citation at the same time.

If you can only do one thing this week, do number six. Submit your own contact form from your phone and see whether the lead actually arrives. It is a ten-minute test and it fails more often than anyone expects.

Frequently asked questions

How do I know if my website is fast enough?
Run your URL through PageSpeed Insights and look at the field data section, not the lab score. You want Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1 at the 75th percentile. The Core Web Vitals report in Google Search Console shows the same data across your whole site.
How many pages should a small business website have?
Enough for one substantial page per service you actually want leads for, plus home, about, contact and any location pages that have real search demand. For most local service businesses that lands between eight and fifteen pages. Page count is the wrong target; the right target is that every page answers a distinct question and could rank on its own.
Is a template website always a mistake?
No. A well-chosen template with fast hosting, clean metadata and real content beats a bespoke build with none of those things. The problem is that templates ship with generic copy, decorative stock imagery and heavy scripts, and most owners never replace them. The template is not the failure. Leaving it as delivered is.
Which of these seven mistakes should I fix first?
Test your contact form first, because a broken form makes every other fix pointless. Then fix indexation problems such as a canonical pointing at the homepage or a stray noindex tag, since those stop the site being found at all. Speed, structure, proof and copy follow in that order.

Web design and local SEO for the Eastern Shore

Calls are answered around the clock. After hours, an AI agent books appointments and takes messages.

Last updated