Your name, your address and your phone have to match, byte for byte, everywhere they appear. Spark Noir Creative audits that consistency on every site it builds, because when the three disagree, search engines lose confidence in who you are and AI assistants stop naming you.
NAP stands for Name, Address, Phone. It is the identity record for a business or an agent, and it lives in far more places than most people realize: your website footer, your Google Business Profile, your brokerage bio, every directory you have ever been listed on, your social handles, the PDF flyer you emailed last spring, and the schema markup buried in your page source.
The rule is simple and unforgiving. Every one of those copies should say exactly the same thing, character for character. Not close. The same.
The exact business or agent name, including the brokerage. "Jordan Ellis" and "Jordan Ellis, Sunrise Realty Group" are two different entities to a machine unless you make the relationship explicit.
Street, suite, city, state, zip, formatted identically each time. "Ste 200" on one profile and "Suite 200" on another is a mismatch, and so is the wrong city on a stale listing.
One number. Not a tracking number in ads, a cell on the flyer, and the office line on the website. Pick the canonical one and use it everywhere a machine can read it.
Local search has always cross-checked these details. Google reads your details in many places and treats agreement as a confidence signal. Disagreement is not a penalty exactly. It is worse than that. It is doubt, and doubt gets resolved by showing someone else.
The stakes changed when people started asking assistants instead of typing queries. An AI answer engine does not rank pages the way a results list does. It assembles an entity, a single mental record of who you are, from every source it can retrieve. When those sources agree, the entity is solid and you are nameable. When they disagree, the engine does one of three things, none of them good: it picks the wrong version, it merges you with somebody else, or it declines to name a person at all and gives the reader a checklist instead.
We run fixed prompt panels against answer engines for our clients, the same wording every time so the runs stay comparable. The most counterintuitive thing we have watched is an agent's own website climb to become the single most-cited domain on a panel, ahead of the big portals and directories, while the engine still named other agents in the actual recommendation. Being the source is not the same as being the recommendation. The gap is usually corroboration and a clean entity, not more pages.
Every one of these is a real pattern we have found in audits, written here with a stand-in agent.
| Break | What it looks like | Why it costs you |
|---|---|---|
| Stale brokerage | An old page still says Jordan Ellis is with a brokerage they left two years ago. | Engines cite the old affiliation as current, and a reader who checks finds a contradiction. |
| Wrong city | A directory profile lists the wrong city or zip, picked up from an old sign-up. | You get filtered out of the market you actually serve and into one you do not. |
| Two phone numbers | The site says one number, an older PDF or flyer says another. | The engine may surface either. One of them rings nowhere. |
| Extra domains | A second, unrelated website is listed on the business profile. | Splits the entity in two and dilutes the signal for the site that matters. |
| Retired brand | An old brand name survives in a social handle or in cached page titles. | Assistants describe you accurately under a name you no longer use. |
| Name drift | A misspelling of your surname appears on a profile you never claimed. | Creates a second, weaker entity that competes with the real one. |
Pick the canonical record first. Decide the exact name, the exact address format and the one phone number, write them down, and then make every surface below match it. Do the profiles you control before the ones you have to ask for.
Matching text is the floor. The ceiling is telling machines the answer directly rather than making them infer it. That means structured data on your site: a RealEstateAgent or LocalBusiness record carrying the same name, address and phone, plus sameAs links pointing at every profile you just cleaned up. The sameAs list is what says "these are all me" out loud, so an engine does not have to guess whether two similar profiles are one person or two.
Then keep it that way. NAP decays. Brokerages change, numbers change, directories you forgot about resurface. Re-check it on a schedule, not once.
One name, one address, one number, everywhere, in text a machine can read, with the profiles linked to each other. Almost nobody does this completely, which is exactly why doing it works.
We check every surface above, hand you the list of what disagrees, and fix the ones we can reach. It is usually the fastest win available on a site that is otherwise fine.
Leave a Message