Six checks.
One afternoon.

01

Does it work on a phone?

Not "does it load." Does it work? Can somebody standing in a parking lot with one hand free actually use it?

For a local service business the mobile share is usually higher still, because the search that finds you happens away from a desk. If your site is awkward on a phone, that isn't an edge case you can get to next year. It's most of your traffic.

The common failures are boring and fixable: text you have to pinch to read, buttons too small or too close together to hit, a phone number that's an image instead of a tap-to-call link, forms that push the submit button off-screen when the keyboard opens, and pop-ups with a close button you can't reach.

02

Can a stranger tell what you do?

You have a few seconds, and the visitor is deciding one thing: am I in the right place? Not "is this attractive." Not "do I like the logo." Just whether this business does the thing they came looking for, for people like them, somewhere near them.

This is the check business owners fail most often, and it's the one they're least able to see. You know what you do. You cannot un-know it, which makes you the worst possible judge of whether your homepage says it.

Four things a visitor should get from your first screen without scrolling:

• What you do: in the words a customer would use, not your industry's words.
• Who it's for: homeowners, restaurants, other agencies, people in Houston.
• Why you: one specific thing that isn't "quality service and competitive prices."
• What to do next: one obvious action, not five competing ones.

03

Can search and AI find you?

There are two versions of this question now, and most businesses only check the old one.

The old one: when somebody types what you sell into Google, do you appear? Still worth knowing, still worth fixing.

The new one: when somebody asks for what you sell (in an AI Overview at the top of Google, in ChatGPT, in Perplexity), does your business get named? Because increasingly that answer is the only thing they see.

These answers get assembled from sources the model trusts: your site, your listings, your reviews, and anywhere else your business is described consistently. A business can rank perfectly well in blue links and still be completely absent from the answer sitting above them.

This is also the one check on the list that didn't exist three years ago, which is why almost nobody has done it.

04

How many visitors actually contact you?

Traffic is the number everyone watches. It's also the expensive one to change. The cheaper number is the share of the traffic you already have that does something.

If 1,200 people visited last month and 18 got in touch, that's 1.5%. The other 98.5% you already paid for, in ad spend or in the years of work that built your reputation, and they left.

Ignore benchmarks. A "good" conversion rate depends entirely on what you sell. 0.5% is excellent if a customer is worth $8,000, and catastrophic if they're worth $40. The number that matters is yours, and whether it moves.

The usual reasons it's low, roughly in order of how often we find them:

• No single obvious next step, or five of them competing on the same page.
• The phone number isn't visible without scrolling, and isn't tappable.
• The form asks for eight things when it needs three.
• Nothing on the page proves anyone has ever hired this business before.
• Prices are a total mystery, so the visitor leaves to find someone who'll say.

05

How fast is it on real mobile data?

Speed gets measured wrong almost every time. Someone loads the site on office wifi on a new laptop, decides it's fine, and moves on. Your customer is on a three-year-old phone with two bars.

Google publishes the thresholds it actually uses, and they're the only speed targets worth caring about:

Ignore the letter grades and percentage scores that speed tools lead with. Plenty of sites score an A and still feel slow, because those scores are averages of things that don't all matter equally. The three numbers above are what Google measures on real visits to your site.

When a site is slow, it's nearly always one of four things: images uploaded straight from a camera at ten times the size they display, cheap hosting that takes a second to respond before anything starts, embedded third-party content like social feeds and chat widgets that load last and block everything, or a page that's simply too heavy, more than a couple of megabytes for what is mostly text.

06

Is it current, and is it safe?

Two problems that look different and cost you the same thing: credibility.

Current. A copyright date from three years ago, a "latest news" post from 2022, a team page with two people who've left, a service you stopped offering. None of it breaks anything. All of it tells a visitor the business might not be paying attention, and they have no way to know whether that's true of the work too.

Reviews are part of this, and they're not optional any more:

Recency matters more than volume. Forty reviews with the newest one two years old reads worse than eight from this quarter.

Safe. The baseline is short: the site loads over HTTPS with a valid certificate, the software behind it is updated, someone other than your hosting company holds a backup, and logins use a password manager rather than a shared password in an email thread.

One piece of old advice worth dropping: forcing everyone to change passwords every 90 days. Current guidance from NIST is the opposite. Scheduled rotation drives people toward weaker, more predictable passwords. Use long unique passwords, store them in a manager, and change them when there's a reason to.