An audit exists to prepare decisions
An SEO audit should answer three questions: what prevents important pages from working in search, what needs to change, and how the team will verify the change after release. A crawler export with hundreds of warnings does not answer them. Each material finding needs an example URL, a cause, affected scope, priority, and a testable target state.
Start with the page route. A crawler must discover the address, receive a stable response, see the primary content, and understand which version is canonical. Only then assess query fit and the user journey. Button colour and title length should not outrank a section blocked from crawling, returning 5xx, or losing content during JavaScript rendering.
A useful report separates observation, interpretation, and recommendation. An observation is repeatable on a URL. The interpretation explains possible impact. The recommendation specifies a change and acceptance criterion. This structure lets a developer estimate the job and shows the client exactly what result they are paying for.
Commission an audit around risk or a confirmed symptom
An audit is most valuable before a migration, CMS change, redesign, or large section launch. It records control URLs and redirect rules before release, then checks server responses, robots.txt, sitemaps, canonicals, internal links, and rendered content afterwards. Finding failures before visibility changes is cheaper than reconstructing the cause later.
The second case is a confirmed symptom: important pages are excluded, traffic changed, or the team cannot explain differences between sections. Diagnosis should combine Search Console, Yandex Webmaster, analytics, and a crawl. A newly published five-page website without a systemic signal may only need a free review of key URLs and measurement setup first.
Choose the format by the decision. A quick check confirms one symptom; a full audit covers templates and data; an implementation audit adds release and follow-up testing. Price alone cannot compare these products. Scope, deliverables, and ownership need to be visible before payment.
What KILENI verified on two live projects
For eco-santeh.ru, a second crawl after template changes confirmed that 509/509 final URLs opened successfully and the checked set no longer contained missing primary tags or exact metadata duplicates. The mobile Lighthouse laboratory score moved from 36 → 57. That is a controlled test result, not a promise of ranking or sales.
For засорсервис.рф, the opening screen shifted during load. After the fix, CLS in two control runs moved from 0.519 to 0.0001, while 575/575 sitemap URLs opened in the final crawl. Both projects used the same proof pattern: record the starting state, change the template, and repeat the identical test.
Do not transfer these numbers to another website as a forecast. They demonstrate an evidence standard: state the method, sample, and limitation. A percentage increase without a date range, source, or baseline cannot be checked or attributed to a specific piece of work.
Accept the audit as a set of testable tasks
Ask for the templates checked, example URLs, evidence for each critical finding, implementation order, and a definition of done. “Fix duplicates” is not enough. “Remove parameter URLs from the sitemap, canonicalise them to clean addresses, and retest ten examples” can be estimated and accepted.
Resolve crawl blockers and repeated template faults before isolated page polish. Release small groups of changes so regressions remain visible and reversible. After each group, repeat the crawl, inspect control URLs, and complete the user journey. Engines need time to reprocess documents, but technical correctness can be checked immediately.
Compare prices only with scope attached: templates and URLs, webmaster-tool access, manual JavaScript checks, developer requirements, implementation, and follow-up testing. A cheap crawler export and an audit with implementation are different products. The client should know what they keep after delivery.
Questions
Key questions
Can an audit guarantee ranking growth?
No. It can verify obstacles and implementation quality; demand, competition, content, and search-engine decisions also affect ranking.
Can we begin with a free review?
Yes. It can confirm primary symptoms and whether deeper work is justified, but it does not cover every template or private data source.
Should an audit run every year?
There is no universal interval. Repeat it after material releases, migrations, new symptoms, or substantial template growth.
