Entity Optimization
Entity optimization is the work of making one business unambiguous to machines: the same name, address and phone number everywhere it appears, structured data that describes the business rather than the website, and links to the profiles that already identify it. The aim is that a crawler or an assistant never has to guess which business a page is about.
What is being optimized
Not a keyword and not a page. The object is the identity itself, spread across every place a machine might read it: the site’s own pages, its JSON-LD, its listings elsewhere, the reviews that name it. Each of these is a claim about who the business is, and the work is making the claims agree. One legal name, or one trading name used consistently. One postal address in one format. One phone number. A LocalBusiness or more specific type in the schema, with those same values inside it, and sameAs pointing at the profiles a machine can already resolve. Once the claims agree, an assistant assembling an answer can attach a review from one page and opening hours from another to the same business without hesitation.
How often the identity is already broken
This site’s audit of 9,483 small-business sites in September 2026 read three pages per site and compared them. 43.1% were inconsistent about who and where they are. 36.9% published more than one phone number across their own three main pages. 34.1% had no postal address a machine could parse anywhere. Structured data made it worse rather than better on most sites: only 17.8% had none at all, but what the rest carried mostly described the website, WebSite and WebPage and BreadcrumbList, not the business. The breakdown is in what 9,483 small business websites serve to an AI crawler. Each of those is a site telling the machine two or three things about itself, and it is the ordinary state, not the exception.
A common question
Does having structured data mean a site's identity is already clear to machines?
No, and the audit of 9,483 small-business sites found the opposite on most of them. Only 17.8% carried no structured data at all, but what the rest published mostly described the website, the page and the breadcrumb trail rather than the business behind it. The identity problems sat elsewhere: 43.1% were inconsistent about who and where they are, 36.9% published more than one phone number across their own three main pages, and 34.1% had no postal address a machine could parse anywhere. Markup that describes the site does not answer either question.
Related terms
Sources
The audit figures are this site’s own, computed from its audit records; sample drawn per metro and trade, not from a random frame, and percentages describe the sample. “Inconsistent” means more than one distinct phone number across the fetched pages or no parseable postal address.