A real audit is a prioritised diagnosis of a specific site. It is not a tool’s PDF export with a score out of 100 — that distinction is the entire difference between an audit worth paying for and one that is not.
What actually gets checked
Crawlability. Whether search engines can reach every page that matters, and whether robots.txt or meta directives are blocking anything unintentionally.
Indexation. Whether pages that should be indexed are indexed, and whether pages that should not be — thin, duplicate, or staging content — are leaking into the index.
Site architecture. How pages link to each other, whether important pages are reachable within a few clicks, and whether the structure reflects how the business actually organises its offering.
Speed on mobile connections. Load time and responsiveness tested against realistic Kenyan mobile conditions, not just a fast office connection.
Structured data. Whether schema markup is present, correctly formatted, and matched to what the page actually contains.
Content overlap. Pages competing with each other for the same search intent, which look like strength but function as internal competition.
Diagnosis versus automated score
An automated score report weights what is measurable, not what matters. A tool can count missing alt text or flag a page as “too short” in seconds; it cannot judge whether that missing alt text is on a decorative image nobody searches for, or whether that short page is short because it is a clean, complete answer to a narrow question. The score treats both cases the same, because it can only count what is easy to count.
A real diagnosis starts from the score’s raw data but adds judgement: which of the flagged issues would actually move something if fixed, given this site’s specific traffic, competition, and goals, and which are cosmetic.
What a good audit deliverable contains
A usable audit deliverable has four parts: the findings themselves, a severity rating for each, an estimate of the effort to fix it, and a prioritised order to work through them in. An audit that lists forty issues with no severity or order is not more useful than one that lists five — it is often less useful, because it gives no guidance on where to start.
Does this actually work?
A genuine audit should be independently verifiable — every finding in it should be something the business can go and check itself, directly. A claim that “the sitemap is missing” can be checked by looking for it. A claim that “page speed is poor on mobile” can be checked with a free public tool. If a finding cannot be checked without trusting the auditor’s word alone, that finding has not actually been demonstrated.
Could this damage my site or waste my money?
Two risks sit specifically at the audit stage. The first is acting on findings out of order — fixing a low-severity cosmetic issue before a high-severity indexing block, because the list had no priority attached. The second is a fix that breaks something else: a redirect change that creates a loop, a canonical tag applied incorrectly, a robots directive edited without checking what else it affects. Both risks are reduced by the same thing — a properly prioritised, specific deliverable rather than a generic checklist applied without judgement.
Where this connects
Cost ranges for an audit are covered in how much does SEO cost in Kenya?. What a business can reasonably check or fix without help beforehand is covered in can you do SEO yourself?