Most websites aren't broken in the way people think. The images load. The navigation works. The phone number is there somewhere. But leads are thin, rankings are poor, and the site isn't doing what it's supposed to do: bring in enquiries from people who didn't already know you existed.

That's a different kind of broken. And a technical crawl won't find it.

What follows is a six-area diagnostic. Run through it honestly. Be prepared to find things you'd rather not find.

‍

Area one: can anything actually reach your site?

Go to your domain followed by /robots.txt. Read what's there.

If you see "Disallow: /" applying to Googlebot, your site is invisible to search. Not poorly ranked. Invisible. This happens constantly on Kenyan sites where a developer ticked "discourage search engines" during build and nobody unticked it after launch. OpenAI's OAI-SearchBot, which powers ChatGPT search, has the same requirement: don't block it or you won't appear in ChatGPT answers. Same file. Five-minute check.

Then open Google Search Console. Coverage report. How many pages are indexed? If you have fifteen pages and four are indexed, something is wrong. Noindex tags left on after development. Orphaned pages with no links pointing to them. URL parameters generating duplicate versions of the same content. All common. All fixable once you know they exist.

This whole area takes about twenty minutes. Do it first. Everything else assumes your pages can be reached.

‍

Area two: does your content answer real questions?

Read the first paragraph of your most important service page out loud. If someone who knew nothing about your business read only that paragraph, would they know what you do, who for, and why they should care?

If no: rewrite it before doing anything else. Research on how AI systems cite content consistently shows citations cluster early in a document. One study found 44.2% of all LLM citations came from the first 30% of the text. Humans decide within seconds whether to keep reading. Both audiences need the same thing upfront: a clear, specific answer to "what is this and who is it for."

Beyond that opening, ask:

• Does each major service have its own page, or are multiple services compressed into one page with bullet points?
• Does the language match how customers actually talk, or is it internal terminology they'd never search?
• Are obvious questions answered anywhere, or left for the prospect to figure out?
• Is there actual evidence anywhere, or only claims?

"We deliver excellence" is a claim. A short paragraph describing a project you did for a Kigali NGO, what the challenge was, what you built, what changed as a result, that's evidence. Claims are cheap. Evidence is valuable. Most Kenyan business websites are heavy on the former and almost empty of the latter.

‍

Area three: what does it feel like on a phone in the real world?

Not in Chrome DevTools. Not on PageSpeed Insights running a simulated test. Put the website on your actual phone, on mobile data, and walk around a bit. See what the experience is like when someone in Thika or Mombasa Road is using it on their lunch break.

The Communications Authority of Kenya reported 48.73 million smartphones in use by late 2025. Almost none are high-end devices on fibre. The site that loads in 1.8 seconds on your office MacBook may take six seconds on a Tecno in Kisumu when the signal keeps dipping. Those are your users.

Google tracks Core Web Vitals: LCP (main content load speed, good under 2.5 seconds), INP (responsiveness to interaction, good under 200ms), and CLS (how much things jump while loading, good under 0.1). Check these in Search Console's Page Experience report. you can read more about this on our previous post: Why Fast, Simple Websites Still Win in 2026. Use field data from actual users, not simulations.

The usual culprits on slow Kenyan SME sites: oversized homepage images, animation libraries loaded for one visual effect, chat widgets, tracking pixels, fonts loading from external servers. Any single one of these can push a site from fast to frustratingly slow. Together they create something that drives people away before they've read a word.

‍

Area four: would a complete stranger trust this?

This is the part most business owners are least honest about. Because you know the business is good, you may not notice that the website gives almost no reason to believe it.

Flanagin and Metzger's work on online credibility showed that users rely on surface cues as shortcuts when evaluating unfamiliar sources. They're not reading carefully and forming rational judgments. They're making quick assessments from things that feel peripheral but actually drive the decision.

Check these specifically:

The About page. Does it name real people, describe actual experience, explain anything honest about how the business works? Or does it say "we are a passionate team committed to delivering exceptional results"? The second version signals there's nothing specific to say.

Testimonials. Specific enough to be credible? "Great to work with, highly recommend!" from "John, Nairobi" is decoration. "They redesigned our labels ahead of a Nairobi Supermarket listing and the buyer said it was the strongest pitch they'd seen that quarter" from a named person at a named company is evidence.

Contact information. Email matching the domain. Physical address or at least a neighbourhood. Phone number consistent across the website, Google Business Profile, Instagram bio, LinkedIn. Inconsistencies read as disorganisation, which reads as unreliable.

Visual coherence. Does the site look built intentionally by one coherent team, or like three different freelancers worked on different sections in different years? Because that second thing happened on a lot of Kenyan SME sites assembled piece by piece over time.

‍

Area five: can people actually find you locally?

For most Kenyan service businesses, local search is where most discovery happens. Someone searching "accountant Westlands" or "event venue Karen Nairobi" is not browsing options. They're about to call someone. The question is whether that's you.

Start with Google Business Profile. Is it claimed? Hours correct? Services listed? A real description that mentions what you actually do? BrightLocal's 2025 data found 85% of local consumers consider hours and contact information important. An incorrect closing time is not a minor detail. It costs you calls.

NAP consistency. Name, address, phone. Identical everywhere: website footer, Google Business Profile, Facebook, Instagram bio, LinkedIn. Not similar. Identical. "Vibur Studio" and "Vibur Creative Studio" are different entities to a search system. Every inconsistency reduces the confidence a ranking system has in your legitimacy.

Reviews. How many, and how recent? A business with forty Google reviews averaging 4.7 looks fundamentally different from one with six at 5.0. Volume signals active trading. Recency signals the business still operates the same way. Both matter. And increasingly, AI systems generating local recommendations are drawing on review signals too.

‍

Area six: when someone's ready to contact you, can they?

Submit your own contact form right now. Does it work? Does it send to an inbox someone actually checks? Does anything confirm what happens next?

This takes two minutes. If the form is broken, you may have been losing enquiries for months without knowing.

Is WhatsApp available as a contact option? In Kenya, Uganda, Tanzania, Ethiopia, a significant share of potential clients will default to WhatsApp over email. If your only contact path is a form, you're adding friction for a substantial portion of your market. A WhatsApp link costs nothing to add and changes the experience meaningfully.

Are there calls to action on service pages, or only buried in the navigation under "Contact"? Someone who lands on a service page from Google search has to actively go looking for a way to reach you if the CTA isn't right there. Put it on the page. Make the next step obvious.

And track which pages generate enquiries. If you can't say which service page got you the most leads last month, you don't have measurement. Without measurement you can't improve the right things.

‍

What to do with what you find

Don't try to fix everything at once.

Fix immediately: crawl errors on important pages, broken contact forms, inaccurate Business Profile data, severe mobile failures, missing or incorrect contact information.

Fix within a month: thin service pages, weak trust signals, missing case studies, generic About copy, vague CTAs.

Fix over the quarter: structured data, FAQ content, local landing pages where justified, conversion improvements.

The audit is only useful if it produces a sequenced list of work, not a report that gets filed.

We've written about why fast, simple websites consistently outperform complex ones and about brand clarity as the engine of business growth. A proper audit connects both to what's actually happening on your site right now.

‍

Frequently Asked Questions

How often should I audit my website?

At minimum once a year. In practice, run a lighter review any time you make significant changes: redesign, new hosting, new pages, a new campaign. The most damaging issues tend to appear during updates, not at original launch.

Can I do this myself?

Most of it. Business Profile review, form testing, mobile testing, Search Console review, reading your own copy critically - all doable without technical expertise. Crawl errors, structured data problems, and indexation issues are harder without tools or experience. A professional audit is worth the cost if you intend to act on what it finds.

What's the single most common critical issue in Kenyan SME sites?

An overloaded homepage paired with empty service pages. All the content, photography, and copywriting effort goes into the homepage. The pages potential clients actually land on from search are skeletal. Fixing this is probably the highest-leverage single change available on most Kenyan business websites.