Your website isn't bringing in clients. It's rarely the design.

Two gaps, not one problem

The story gap. Your business kept moving after the site launched. Offers sharpened, prices changed, your best clients got bigger, your process matured. The site still tells the story from the year it was built. A prospect lands, reads a description of a company that no longer exists, and can't match their need to your words. They don't email you to clarify. They leave.

The machine gap. More of your buyers than you think never land at all. They ask a search engine or an AI engine, and the answer is assembled from what machines can read about you. If your site is hard for machines to read — content painted in by JavaScript, no structured data, headings in the wrong order, nothing that answers a real question — you're missing from answers you should own.

The story gap loses the visitors you get. The machine gap loses the visitors you never see. A redesign fixes neither, because both live below the paint.

Why this happens to good businesses

This isn't neglect. It's structure. In most founder-led firms, the real story of the business lives in the founder's head: how the work actually gets done, who it's really for, what makes it different. The website was written once, by someone translating that story secondhand, and nobody owns keeping it true.

So the site drifts. Every quarter the gap between what you say and what you do gets a little wider. Meanwhile the founder keeps winning work in person, on referrals and reputation — which hides the cost. The site quietly filters out everyone who checked you out before the call, and you never hear about the deals that died there.

A quiet website isn't evidence that websites don't matter in your business. It's usually evidence that yours is answering the wrong questions, or none at all.

Five checks that find the real problem

1. The stranger test. Give your homepage to someone who doesn't know your business. Sixty seconds. Then ask: who is this for, what do they sell, what does it cost? Wrong answers here mean the story gap is live, and every visitor is failing the same test silently.
2. The founder-gap test. Record yourself describing your business to a friend for two minutes. Compare it to your homepage. The distance between the recording and the page is the exact size of your story gap. Most founders find their spoken version is sharper, more current, and more honest than anything on the site.
3. The AI answer test. Ask ChatGPT, Claude, or Perplexity the question your prospects ask before they find you. If the answer is wrong about you, vague about you, or names competitors instead of you, the machine gap is live.
4. The source test. Right-click your homepage, view source, and search for a sentence you can see on the page. Missing? Machines may be reading a nearly empty page. Present? Check whether there's a block starting with application/ld+json — structured data that tells machines what your business is. No block means they're guessing.
5. The price test. Can a buyer learn what working with you costs, even roughly, anywhere on the site? Buyers who can't find a price assume they can't afford you or that you're hiding something. Machines can't cite a price you never published.

What to do with what you find

If the story gap dominates: rewrite before you redesign. Take that two-minute recording and make the homepage say what you said. Publish your current offers with current prices. Cut every page that describes work you no longer want.

If the machine gap dominates: fix structure before you write another word. One H1 per page. Structured data that names the business, the services, the prices. Content in the document, not painted on after. Then every page you publish afterward inherits the fix. Our guide on whether AI can read your website walks through each check.

If both are live — which is the common case — fix them in that order: story, then structure, then volume. Publishing more content on a broken foundation just gives the machines more of the wrong thing to repeat.

What the gaps are actually costing

Run the napkin math before you decide this can wait. Take your average engagement value. Now assume the gaps cost you one deal a quarter — a buyer who checked the site before the call and quietly didn't book, or asked an AI engine and got pointed somewhere else. For a firm whose average engagement is $40,000, that's $160,000 a year, against a fix that costs a rounding error of that.

One deal a quarter is a conservative guess, and that's the point: you can't measure the deals that die on the site, because they die silently. No bounce report shows you the prospect who read a stale services page and crossed you off. No analytics tool shows you the AI answer that named your competitor. The absence of evidence reads as "the website doesn't really matter for us" — right up until you fix it and the inbound changes.

There's also an asymmetry worth naming: your referrals are checking the site too. A warm referral doesn't need the site to sell them, but a stale site can unsell them. The strongest pipeline in the world still passes through the same broken surface.

When to get an outside read

You can run every check above yourself. What's hard to do yourself is judge the cost of what you find — which gaps are actually losing you money and which are cosmetic. You're too close to the story to see which parts have drifted.

That judgment call is a bounded, buyable thing. It shouldn't require a discovery call or a retainer.