A small business website has one job. Someone who does not know you lands on it, understands what you do, and can take the next step without hunting. Speed, a clear offer, a conversion path, and a page that works on a phone are how that job gets done. Animation, a clever headline, and a stock photo of a handshake are not.
We write this because the same four things fail on the sites we are asked to fix. They fail in Marianna on a stretched connection. They fail in Tallahassee on a phone in a parking garage. They fail when a Google Business Profile is right and the page it opens is not.
Speed is a closed tab, not a score
If the first screen takes more than a couple of seconds on a mid-range Android, a lot of people leave. They do not wait to see that the brand is “premium.” They go back to the map and tap the next pin.
We still look at Lighthouse. We do not design for a 100. We design so the largest thing in the first viewport is a real heading and a way to act, not a 2MB hero that has to decode before the text appears. That means:
- Images with a width and a height, in AVIF and WebP, sized for the slot they sit in. A 4000-pixel photograph of the storefront, served into a 640-pixel column, is not craft. It is a tax on the customer’s radio.
- Type that is self-hosted and limited. Every extra family, weight, and italic is another file on a slow network. Personality can come from scale and tracking. It does not require a display face plus a body face plus a “fun” accent.
- JavaScript that answers a click, not a scroll. A site that cannot render its offer until a bundle downloads is a site that is late. Our own marketing site is held under a small JS budget for this reason. Client sites get the same rule: if a library is not doing work the customer asked for, it does not ship.
- Hosting that is close to the visitor and that can be cached. A cheap shared box in another region, with no CDN in front of it, will lose to a competitor who did less design and more plumbing.
PageSpeed is a report. The customer’s phone is the test. We watch both. If they disagree, the phone wins.
Rural Panhandle connections are not an edge case for the businesses we serve. A four-second homepage on a stretched LTE hop is a closed tab. If your trade is HVAC, dentistry, or a clinic off I-10, assume the search happened in a truck, not at a desk.
The offer has to be on the first screen
Most underperforming sites hide the work. The homepage opens on a slogan, a slideshow, and a paragraph about being passionate. The services are in the third scroll. The town is in the footer. The person who typed “Marianna dentist” or “Panama City boat repair” already knows the category. They need to know it is you, here, and what happens if they call.
A homepage that performs says, in the first viewport:
- Who you are.
- What you do, in the words the customer uses, not the words the trade group uses.
- Where you do it, if place matters.
- What to do next.
That is not a brand exercise. It is a retrieval exercise. If the Google Business category is “Plumber” and the homepage says “flow solutions for the built environment,” you have already lost the person who is standing in water.
Do not accent a single word in the headline in a different color and call it strategy. Do not write “in today’s digital landscape.” Write the service and the town.
Inside pages have the same job at a smaller scale. A service page is for one service. If two URLs fight for the same query, pick one. If a service is only offered in season, say so. If you do not do residential, say so before they fill out a form.
Conversion is a path, not a button style
A conversion path is the sequence from “I think this is the right firm” to a call, a form, or a booked slot. It fails in boring ways:
- The click-to-call number on mobile does not match the Google Business number. The customer calls, someone else answers, or nobody does.
- The form has ten fields, a “how did you hear about us,” and a budget dropdown that feels like a qualification trap. People in a hurry want a name, a number, a sentence. You can ask for the rest on the phone.
- Hours in the footer disagree with Google. You are closed on the listing and open on the site, or the reverse. Either way the customer does not trust the next step.
- There is no next step on the page they landed on. A blog post about “ten tips” that never links to a service, a phone number, or a form is a dead end you paid to write.
- The thank-you page is a shrug. After a form, say what happens and when. “We reply within two business days from hello@…” is a path. “Thanks!” is not.
We do not mean every page needs a pop-up. We mean every template that can receive a visitor from search or from the map has a primary action, and that action works. On a phone, that is often tap-to-call. On a desktop, it is often a short form. Both should be visible without opening a menu.
Track the path you actually have. If you cannot tell whether a call came from the site, you will argue about the website instead of the offer. Call tracking is a tool. It is also how numbers on listings get duplicated. If you use it, the public number on the profile and the site still have to be one story.
Mobile is the site
Most of the local searches we care about happen on a phone. Designing desktop first and “making it responsive” is how you get a 44-pixel tap target that is actually 12, a map that the visitor cannot pinch, and a hero that pushes the offer below the fold on a small screen.
Check the templates at 375px, not as a screenshot on a 27-inch monitor. Check:
- The heading still states the offer when it wraps.
- The primary button is a button, not a text link in the paragraph.
- Sticky headers do not cover the heading when someone taps a jump link.
- Forms do not zoom the page because the input font is 12px.
- Images do not overflow. Overflow is how you get a horizontal scrollbar and a visitor who thinks the site is broken.
If the business lives in the map pack, the page that opens from the profile is often a service or location URL, not the homepage. That URL has to carry the name, the address, the phone, and a reason to act. A homepage that performs and a landing page that does not is a leak. Google will send people to whatever URL you put on the profile. Put the one that can close.
What we do not count as performance
A count-up of “happy clients.” A logo strip. A testimonial card with an oversized quotation mark. A video that autoplays. A chat widget that asks “how can I help” before the page has rendered. Those can be added later if they earn their place. They do not make a weak offer faster or clearer.
We also do not treat a redesign as a performance strategy if the listings are wrong. A new coat of paint on a site that still has two Google profiles, a suite that emptied in 2019, and a tracking number in the footer will look better in a screenshot. It will not get more of the right calls.
What to do on Monday
If you run the site yourself, do this in order:
- Open the homepage on your phone, off Wi-Fi. Time how long until you can read the offer and tap a number. If you cannot, stop decorating and start subtracting.
- Read the first heading out loud. If it does not name the work, rewrite it.
- Tap every phone number and submit a test form. Fix whatever fails before you write another paragraph of blog copy.
- Compare the name, address, phone, and hours on the site to the Google Business Profile. Make them identical.
- Look at the URL on the profile. Load it on a phone. If it is a thin inner page, give that page the same care as the homepage.
If you want us to do that work, that is a website project, or a care plan if the templates can take it. We will tell you which after we look. We start from the page list rather than a mood board.
Performance is not a mystery. It is whether a stranger on a phone can see what you do and act before the next listing loads.
