Why many audits don’t help
Many website audits consist of an SEO tool export: hundreds of warnings sorted into categories, with no connection to the business. A missing alt attribute sits next to a contact form that doesn’t work – flagged in the same color.
For a B2B website, though, only one question counts: What’s keeping the right prospects from getting in touch? A good audit answers it and ranks all findings accordingly.
The four review areas
| Area | Guiding question | Typical checkpoints |
|---|---|---|
| Found | Is the website found on Google and in AI systems – for the right questions? | Indexing, titles and descriptions, search intent, content for expert questions, brand outside the website |
| Understood | Does a visitor grasp within a few seconds what’s offered, for whom, and with what result? | Core message, service descriptions, audience language, proof, structure |
| Chosen | Is there a clear, working next step – and reasons to take it? | Contact options, forms, calls to action, trust signals, mobile use |
| Tech | Does the foundation hold up? | Availability, redirects, load time, accessibility, security, legal pages, measurement |
Google sums up the basis well: its systems favor content written for people; trust is the most important aspect [1]. An audit should therefore check not only whether a page is technically “correct” but whether it’s convincing.
Example: the ditengy.com audit
We reviewed our own website using this framework on October 30, 2026 – shortly after it first went live. We measured from the outside: status codes and redirects, retrieval without JavaScript, Lighthouse, display from 320 to 1,280 pixels wide. Search Console data, web analytics, and user tests weren’t available. Statements about the effect on visitors are therefore expert judgment, not measurement.
The key findings:
| Area | Finding (Oct. 30, 2026) | Severity |
|---|---|---|
| Tech | The domain without “www” still showed an error page from the previous website provider | critical |
| Chosen | No working way to get in touch: the form had no recipient, and no email address was listed | critical |
| Tech | Legal notice and privacy policy with placeholders were live | critical |
| Understood | Core message “Make more of your digital potential” – interchangeable, with no target audience and no result | high |
| Found | No content for expert questions; the most important entry offering had no page of its own | high |
| Chosen | No contact button in the mobile header on very long pages | medium |
| Found | 6 of 9 page titles too long, identical subheadings on four service pages | medium |
| Tech | The host’s preview copy of the website was indexable by search engines | medium |
| Found | Brand and founder barely discoverable online | high, but not solvable on the website |
Already in good shape were, among other things, HTTPS, clean URLs, sitemap, structured data, and static HTML that’s readable even without JavaScript.
How findings become an order of priority
We rate every finding against five criteria: business value, expected effect, effort, risk, and how certain the insight is – verified finding or assumption. This results in three tiers:
- First: what prevents inquiries or poses legal risk. For ditengy.com, that was the contact option, the legal pages, and the broken domain. All three were fixed or bridged within two days.
- Then: what makes inquiries harder. A new core message, clear service descriptions, a dedicated page for the entry offering, a contact button on smartphones.
- After that: what builds visibility. Trade articles, connecting the brand with profiles, measuring AI visibility.
It’s important to recognize prerequisites. Web analytics alone won’t bring in a single inquiry – but without them, no later measure can be evaluated. That’s why they move up the list despite little direct benefit.
Result after two days (as of Nov. 1, 2026): the critical findings have been fixed or bridged. Lighthouse measures 98–99 points for performance on desktop and 86 on mobile, and 100 each for accessibility, best practices, and SEO. Whether this leads to more inquiries can’t be said yet – there isn’t enough data so far.
What an audit can’t do
- It doesn’t replace data. Without analytics and search data, every statement about visitor behavior remains a hypothesis. A good audit says so openly.
- It doesn’t replace users. Five conversations with real prospects often reveal more than any expert review.
- It doesn’t guarantee rankings. It shows what’s within your control – not how Google or an AI system decides.
- It doesn’t solve everything on the website. For ditengy.com, a big lever lies outside: the brand is still barely known online. No code changes that.
Checklist for your own website audit
Found
- Is the website reachable with and without “www,” and does it redirect cleanly?
- Are all important pages indexed, and are there no indexable copies?
- Does every service have its own page with a clear search term?
- Is there content that answers your audience’s expert questions?
Understood
- Does the homepage say in one sentence what you do, for whom, and with what result?
- Is every service explained in two or three sentences?
- Is there proof – methods, examples, people – rather than mere promises?
Chosen
- Does the contact option demonstrably work (test inquiry sent and received)?
- Is the next step visible on every page, including on smartphones?
- Is it clear what happens after the inquiry?
Tech
- Are the legal notice and privacy policy complete?
- Does the website load quickly on mobile, too?
- Is privacy-compliant measurement in place?

