Our Audit Methodology

By Sameer Bhat · Published 15 January 2026

This page explains every part of the audit method. You can see what the audit measures and how each finding affects your score.

A score without a method gives little value. A visible method turns the score into a useful guide.

Why we publish the method

Every score must connect to specific evidence. You must be able to check that evidence yourself.

The method also shows the limits of the audit. Some sources can fail or provide no data.

The report marks these cases as unmeasured. It never changes missing evidence into a failed result.

The four signal categories

Crawler access

A crawler is software that reads website pages. Search engines and AI search engines use crawlers.

The audit checks whether crawlers can access important pages. It also checks the following items:

• robots.txt access rules.
• noindex instructions.
• Server response codes.
• Redirect chains.
• Sitemap access.
• Different responses for crawler software.

A crawler block can hide an excellent page. Therefore, the audit checks access before it evaluates other page signals.

Structured data

Structured data labels page information in a standard format. JSON-LD is one common structured data format.

The audit checks whether structured data exists. It also checks syntax and important properties.

Business structured data can include a name, website address, logo, social profiles, address, and contact details.

The audit checks links between related structured data items. These links help software understand one connected business record.

The audit also checks Open Graph and social card tags. These tags control how other services describe a shared page.

Content clarity

AI search engines need direct and accurate statements. Long introductions can hide the main answer.

The audit checks headings, direct answers, and specific facts. It also checks whether important claims use clear language.

The audit does not reward repeated keywords. Repetition can make useful information harder to read.

Trust signals

Public sources can support your business claims. These sources include reviews, directories, and established websites.

The audit checks review activity and consistent business details. It also checks available website authority signals.

The audit only checks public evidence. It cannot confirm a private claim that no public source supports.

How the audit calculates the score

The score starts at 100. Confirmed findings reduce the score.

Each finding has one severity:

• Critical: The problem can block major parts of visibility.
• High: The problem can cause a strong visibility loss.
• Medium: The problem can limit performance or clarity.
• Low: The problem has a smaller effect.

Critical findings reduce the score more than lower-severity findings. One critical problem can matter more than several cosmetic problems.

Penalties do not increase without a limit. Each severity has a maximum total penalty.

This rule prevents many small findings from hiding the effect of one critical problem.

The report shows every finding and severity. It does not show only the final score.

How the audit handles missing data

An external service can fail. A page can return an unrelated error. Some signals do not apply to every website.

The audit marks these results as unmeasured. It also gives the reason when the reason is available.

A failed result identifies a confirmed problem. An unmeasured result identifies missing evidence.

These results need different actions. You fix a confirmed problem. You investigate an unmeasured result when it matters.

Why scores can change

Your website can change between audits. External sources can also add, remove, or update information.

An external service might answer during one audit and fail during another audit. This change can affect measured coverage.

The method and severity weights stay the same between audits. The evidence explains any score change.

What to do with a low score

Start with critical and high-severity findings. These findings usually have the largest effect.

Next, review unmeasured results. Decide whether a manual check can provide useful evidence.

Then review medium and low-severity findings. Do not delay an important fix to chase a perfect score.

The goal is clear visibility, not a decorative number.