What the Brand Entity Consistency Checker does
This checker compares how up to ten of your pages state the basic facts about your organisation - name, legal name, website, logo, telephone, email, address, social profiles and identifiers such as a VAT number - in both structured data and visible text, and lists every place they disagree. It then proposes one canonical Organization JSON-LD built from the values most pages agree on.
Inconsistent details are among the easiest problems to fix and the easiest to miss: a contact page that still has the old phone number, an about page using a nickname, an Instagram link to an account you stopped using. Search engines, maps, directories and AI assistants all have to reconcile what your pages say; consistent facts give them nothing to reconcile.
How to use it
- Paste the HTML of two to ten pages, each starting with a line such as === https://example.com/contact ===, or switch to Fetch URLs and list the addresses.
- Include the pages most likely to state your details: home, contact, about, legal or imprint, and any location pages.
- Select Compare pages. The disagreements list names each fact, every value found, the pages it appears on and where on the page it came from.
- Read the fact matrix for the full picture, including facts that only some pages state.
- Correct the pages, then copy the proposed Organization JSON-LD - after checking each value - and use the same block on every page that describes you.
Reading the results
Conflict means two genuinely different values: two phone numbers, two names, two Instagram accounts. Format differs means the same value written differently - "12 Mill St" and "12 Mill Street" - which is worth tidying but not urgent. Uneven means some pages state a fact and others do not, or list different sets of profiles.
The comparison is deliberately forgiving in stated ways: names ignore case, punctuation, "&" versus "and" and legal suffixes such as Ltd or Inc; phone numbers compare their last nine digits so 0113 496 0321 and +44 113 496 0321 match; web addresses ignore https, www and a trailing slash; twitter.com and x.com count as the same site.
The proposed JSON-LD takes, for each fact, the value stated on the most pages. Minority names are offered as alternateName so you can keep or delete them. Facts no page states are left out rather than guessed.
Worked example: a bakery whose about page is out of date
The bakery example has three pages. The home page's Bakery markup and the contact page's LocalBusiness markup agree on the name ("Rise & Shine Bakery" and "Rise and Shine Bakery" normalise to the same thing) and on the phone number, once 0113 496 0321 and +44 113 496 0321 are compared by their last nine digits. The about page calls the business "Rise N Shine" and gives +44 113 496 0999 - two conflicts, both traced to page 3.
The email addresses differ (hello@ on the home page, orders@ in the contact page's mailto link), the address is the same but written two ways, and the about page links an old Instagram account, so sameAs shows two different profiles on instagram.com. Four conflicts and one format difference in all. The proposed JSON-LD uses the majority name and phone, lists Rise N Shine as an alternateName to review, and merges four distinct profiles into sameAs.
Why consistency is the part you control
Knowledge panels, map listings and AI answers draw on many sources, and no outside tool can see how any of them weighs your pages. What you can control is that your own pages never contradict each other. schema.org's Organization type gives every fact a named property - legalName, telephone, address, sameAs, vatID - so the same values can be stated the same way everywhere.
This is an A2Z heuristic, not a documented ranking rule: consistent facts reduce ambiguity for any reader. It is also good housekeeping - an out-of-date phone number on one page costs customers regardless of search.
Limitations: what the result does not prove
- It reads only what the pages state in HTML and JSON-LD. It does not check business directories, map listings, social profiles or company registers.
- Visible-text extraction is limited to tel: and mailto: links, the page title suffix, og:site_name and a copyright line. A phone number written as plain text without a tel: link is not picked up.
- Majority is not truth. If most pages carry an old address, the proposed JSON-LD will too; check every value before publishing it.
- Fetched pages are read without running JavaScript, so markup added by a tag manager or script is not seen. Paste the rendered HTML instead.
Privacy: where your data goes
The analysis runs in your browser. If you choose to load a public URL instead of pasting, a2z.tools fetches that address read-only and returns the text to your tab; private and internal addresses are refused and nothing is stored. Session recording and tag-manager scripts are switched off on this page.
Standards and sources
- schema.org vocabulary - checked 19 Sep 2026
- schema.org - Organization
- schema.org - sameAs
- Google Search Central - Organization structured data
Frequently asked questions
What is entity consistency in SEO?
It means every page, profile and listing states your organisation's core facts - name, address, phone, website, profiles, identifiers - the same way. For local businesses this is often called NAP consistency (name, address, phone).
What should the sameAs property contain?
URLs of pages that unambiguously represent the same organisation elsewhere: official social profiles, a Wikipedia or Wikidata entry, a company register page. It should not list pages that merely mention you, and every page should list the same set.
Why are Ltd and Inc ignored when comparing names?
A trading name and a legal name with a suffix usually refer to the same organisation, and flagging every "Acme" versus "Acme Ltd" would bury real conflicts. Use legalName for the registered form; the checker still shows every spelling it found.
Is a phone number with and without the country code a conflict?
No. Numbers are compared by their last nine digits, so a national form such as 0113 496 0321 matches +44 113 496 0321. Different endings, such as 0321 and 0999, are reported as a conflict.
Can I paste pages instead of fetching them?
Yes. Paste each page's HTML after a marker line like === https://example.com/about ===. Pasted pages are compared entirely in your browser; fetching sends only the addresses to A2Z's reader, one at a time.
Should every page carry the full Organization markup?
It is common to put the full Organization node on the home page and reference it elsewhere by @id, or to repeat it on key pages. Either works as long as every copy states the same values, which is exactly what this check verifies.
Last reviewed by the A2Z.Tools team against the sources listed above.
