What’s included in an SEO audit, and what to check before you start promotion

What an SEO audit is, and why it comes before promotion

Promotion means changes to the site: new pages, template edits, copy, links. Starting them without a diagnosis makes it easy to spend resources on the wrong things — for example, writing copy for pages that are blocked from indexing, or creating a new page when an existing one already covers the topic.

An audit records the site’s starting state and turns it into a plan: what problems exist, what each one affects, and in what order to fix them. That’s why an audit is the first step, not a substitute for SEO work.

Audit, monitoring and automated checks — how they differ

These three things are often lumped together under the word «check,» but they serve different purposes:

• An automated check is running a tool that compares pages against a set of rules: response codes, meta tags, duplicate titles, broken links. It quickly finds formal deviations.
• Monitoring is ongoing observation of key metrics and statuses. For example, the «Site diagnostics» section of Yandex Webmaster (Yandex’s counterpart of Google Search Console) reports errors on the site and recommendations for fixing them. Search Console helps you learn how the site is represented in Google Search and fix possible issues.
• An expert audit is a one-off, in-depth diagnosis. A specialist uses data from tools and dashboards but evaluates it in the context of the business, demand and site structure, then sets priorities.

Monitoring and automated checks report changes and triggered rules. An audit answers why that matters and what to do about it.

What an SEO audit consists of

The set of areas depends on the site, but a full audit always rests on the same foundation. The summary table below shows what’s checked in each area and where to read more:

Area	What’s checked	What it affects	More
Technical foundation	accessibility and response codes, robots.txt, noindex, canonical, redirects, Sitemap, structured data	whether pages make it into the index and in what form	the articles on indexing and on structured data
Structure	sections, hierarchy, URLs, internal links	whether the crawler finds pages and whether the connections between them are clear	the article on internal linking
Keyword research	demand, the «query → page» map, gaps, cannibalization	whether every query group has its own page	the article on the keyword map
Content	page purpose, overlap, match with intent	whether a page answers the query it’s assigned to	the article on SEO copywriting
Metadata	Title, description, headings	how the page is presented in search results and whether its topic is clear	—
Changes and regressions	what broke after releases, migrations, a redesign	whether the site is losing what already worked	—
Priorities	critical, important, can wait	where to start the work	—

Technical conditions: crawler access, indexing and markup

The technical part answers whether search engines can retrieve and correctly process the pages. It checks that important pages return a 200 code and don’t fall into redirect chains; that the sections you need aren’t blocked in robots.txt or tagged noindex, and that canonical doesn’t point to a different address; and that the Sitemap matches the site’s real structure. More detail — what to check in the sitemap.

A separate question is which pages are actually in the index and why some are excluded. That’s worked out from the Yandex Webmaster and Google Search Console reports; step by step, see the article on how to check page indexing.

The audit checks structured data for two things: whether it matches the visible content of the pages, and whether the theme and plugins duplicate entities. Validators and common errors are covered in the piece on how to check structured data.

If there are many technical obstacles, or they keep recurring after every release, fixing the technical obstacles to search becomes a separate task.

Structure: how sections are organized and pages are linked

This looks at how the site is organized: whether the sections match how people search, whether the hierarchy and URLs are logical, and whether links lead to the important pages. Typical findings: pages with no inbound links, links to redirects and errors, important pages buried too far from the homepage.

How to find such problems using the link graph is covered in the article on how to find pages with no inbound links and check anchor text.

Keyword research: demand and pages, gaps, cannibalization

The keyword research part compares demand with what’s actually on the site. It looks for query groups with no matching page, pages not assigned to any query, and cannibalization — several pages competing for the same intent.

The basis of this check is the query-to-page map. How to build it is described in the article on the keyword set and the «query → page» map. If the audit shows the map needs to be rebuilt from scratch, that’s a task where the structure gets rebuilt and queries get reassigned.

Content: page purpose, overlap, quality

Content isn’t judged by length or keyword density. The main questions: does every page have a clear purpose, does it answer the query it’s assigned to, and does it duplicate neighboring pages?

Criteria for checking an individual piece of copy are covered in the article on how to check copy for a match with intent. When the problem isn’t individual texts but pages competing or lacking a purpose, you need work on page purpose and overlap.

Metadata: Title, description, headings

Metadata is checked for two things: uniqueness and whether it matches the page’s content. Google advises writing a title for each page that reflects its content and avoiding template titles where only a word or two changes. Yandex gives an example: catalog pages all titled «Catalog» don’t let a visitor tell them apart.

The description may be used as the page’s snippet in search results. Google generates a description automatically and uses the meta description when it’s more accurate than the page’s content; it advises giving different pages different descriptions.

The audit looks for duplicate and templated Title and description tags, empty fields, multiple H1 headings on a page, and headings that don’t match the topic. Webmaster also flags duplicate titles.

Quality and regressions after changes

A regression is a deterioration that appears after changes to the site: a page starts returning an error, a template picks up noindex, links disappear from the menu. Such problems often go unnoticed at first.

That’s why an audit checks the aftermath of recent releases and sets a baseline before major changes. The guideline is the search engines’ own recommendations. For a structure change, Yandex advises setting up 301 redirects, making sure the new addresses return a 200 code and aren’t disallowed in robots.txt, and updating the Sitemap. For a design change — checking that pages don’t have noindex or a canonical pointing to other pages.

Priorities: in what order to tackle tasks

Prioritization turns a list of findings into a plan. A convenient three-level scale:

• Critical — the problem keeps important pages out of search or affects the whole template: a blocked section, server errors, canonical pointing to the homepage.
• Important — the problem reduces effectiveness but doesn’t block anything: cannibalization, pages with no inbound links, templated Titles.
• Can wait — improvements with a small effect, or ones for secondary pages.

The level depends not only on the type of error but on the scale and importance of the pages involved. A broken link in an old news item and an error in the product card template carry different priorities, even though formally they’re the same category.

What to check before starting promotion: a short checklist

This list doesn’t replace an audit, but it shows whether the site is ready for work:

1. Important pages load, return a 200 code, and don’t redirect through chains.
2. Sections that should be in search aren’t blocked in robots.txt or tagged noindex; utility pages are blocked.
3. Standalone pages’ canonical points to themselves.
4. The Sitemap contains only canonical addresses that should be indexed.
5. The site is connected to Yandex Webmaster and Google Search Console, and the diagnostic errors have been reviewed.
6. There’s a «query → page» map: each query group has one target page.
7. There are no pages competing for the same intent.
8. Internal links lead to the important pages; there are no pages with no inbound links.
9. Every page has its own Title and description, one H1 that reflects the topic.
10. The site’s current state is documented so there’s something to compare against after changes.

If the answer to several of these is «I don’t know,» that’s the reason to start with an audit.

Automated check or expert audit

These aren’t mutually exclusive options. An expert audit relies on tool data, and an automated check is useful between audits.

Criterion	Automated check	Expert audit
What it checks	formal parameters against a given rule set	the same parameters, plus a match with the site’s goals, demand and structure
Site context	doesn’t know which pages matter and which should be blocked	starts from the business goals and priority sections
Result	a list of triggers, often with no connection between them	findings grouped by cause; each one has an impact, an action and a priority
When it’s appropriate	ongoing monitoring, a quick hypothesis check	when you need to decide what to fix, in what order, and which pages to keep

What automated-check tools offer, and where their limit is

Automated-check services quickly crawl many pages and find recurring technical deviations. The search engines’ own tools also check sites automatically: Yandex, for example, does this daily and shows the result in «Site diagnostics,» under the «Errors» and «Recommendations» tabs.

These checks have one limit in common: a rule doesn’t know the context. A tool will flag noindex as an error even when the page was blocked on purpose. And it won’t notice that two pages with different titles answer the same query. It won’t tell you which of two competing pages to keep. Those decisions are made by a specialist.

How to read audit results

What an audit’s result should include

A list of errors found isn’t yet a result. After an audit, the team should have three things: a report that explains the consequences of each problem; a task list ordered by priority; and recommendations detailed enough for a developer or editor to make the fixes without further questions.

If a report only lists errors without impact and priorities, that’s the output of an automated check, not an audit. How we run an audit and what we hand over at the end is described on the OrganicOps SEO audit page.

Finding → what it affects → what to do

Each finding is best read as a chain of four fields. A sample example:

Finding	What it affects	What to do	Priority
The catalog section template outputs noindex	the section’s pages don’t take part in search	remove noindex from the template, check page status in the webmaster panels	critical
Two blog articles answer the same query	the pages compete for one intent	pick the primary one, and merge the other into it or split them by intent	important
Some cards share the same description	the snippets in search don’t distinguish the pages	make descriptions unique for priority cards	can wait

If a finding is missing even one field, ask the person who delivered it. Without «what it affects» you can’t gauge importance; without «what to do» you can’t hand the task to the team.

Why a long list of errors with no priorities is useless

A long list creates the impression of thorough work but doesn’t help you make a decision. Hundreds of identical triggers often come down to a handful of causes — for example, a single template. And a truly critical problem can get lost among the minor ones.

Signs of a useful report: findings are grouped by cause, not by page; each one has an impact and an action; it’s clear where to start. You can implement the tasks with your own team, or continue with implementation as part of SEO promotion and ongoing support — with a check of the result after the changes.

Limitations of an audit

A snapshot as of the check date; repeat after significant changes

An audit describes the site as of the check date. After a release, a migration, a template change or a major structural edit, some of its conclusions can become outdated. That’s why, after significant changes, it’s worth repeating the audit — fully or for the affected areas.

Also account for how the tools process pages. If content is rendered via JavaScript, check whether the tool sees it the same way a user does — otherwise some of the conclusions may be inaccurate.

An audit doesn’t guarantee a rise in rankings

An audit is a diagnosis and a plan, not the changes themselves. The result depends on which tasks actually get implemented, as well as on demand, competitors and the search engines’ algorithms.

The search engines themselves don’t give guarantees either. Google doesn’t guarantee that a site will be crawled, indexed and shown in search results. It states outright that no one can guarantee the number one spot in its results, and that it has no recipe that will automatically put a site in the top positions. If an audit is sold with a promise of rankings, that’s a reason for caution.

Frequently asked questions

Can you do an SEO audit yourself?

Partly — yes. The Webmaster and Search Console reports, the checklist above, and the articles on each area will help you find the obvious problems. The conclusions are harder: without experience, it’s difficult to gauge scale, find the common cause behind similar errors, and set priorities.

How does an SEO audit differ from a technical audit?

A technical audit is one part of an SEO audit: it checks what’s stopping the crawler from retrieving and processing pages. An SEO audit adds structure, keyword research, content and metadata, and rolls it all into a single prioritized list.

How often should you run an SEO audit?

There’s no universal schedule. The guideline is significant changes on the site and a drop in search traffic with no clear cause. Between audits, the site’s state is tracked through the webmaster panels and automated checks.

What should you prepare for an audit?

Gather what only you know about the site: which sections matter most to the business, what’s changed on the site recently, and what’s planned. If the site is connected to Yandex Webmaster and Google Search Console, grant access to them — report history makes it easier to pin down when a problem appeared.

Does a site that’s just launching need an audit?

A full audit of keyword research and content makes sense once pages already exist. But the technical foundation and structure are worth checking before opening the site for indexing: fixing a template before launch is easier than bringing pages back into search afterward.