Portals publish inventory. An agent has to publish knowledge

Why does the listing feed not work as content?

Because it is not yours, it is not different, and it disappears.

The feed on your site is the feed on every other agent's site. Search engines have seen it, buyers have seen it, and an AI answer assembling a response about a neighborhood has no reason to prefer your copy of it to anyone else's.

It is also temporary. A listing page exists for six weeks and then describes something that is gone, which leaves a site quietly accumulating dead pages about properties nobody can buy, each one a small argument that the site is not maintained.

What ranks and what gets cited is durable, specific, local writing that only somebody who works there could produce. Portals cannot produce it, because they do not know which block gets fog until two in the afternoon or which HOA has been arguing about a roof for three years.

What does "knowledge" actually mean here?

Answers to the questions buyers ask you on a Saturday.

Not market reports. Every agent publishes a market report, they are assembled from the same data, they all say the same thing in the same month, and almost nobody reads them.

The material worth publishing is the stuff you say out loud in the car between showings. Which streets are quiet and which are cut-through routes. What the parking is genuinely like. Which schools people move for and what the catchment does to the price of the two blocks either side. What a foundation of that vintage usually costs to look at. Which building has a special assessment coming. How long the walk to the station really takes with a bag.

Write one page per question, not one page per neighborhood. A neighborhood page tries to be everything and is nothing. "What buyers get wrong about parking in this part of town" is a page somebody will finish reading.

How is this different from a city landing page?

It is the opposite of one, and the difference decides whether it helps or hurts.

The standard agent build is forty near-identical city pages with the name swapped. It looks like coverage and reads like nothing, and past a certain number those pages start working against the site rather than for it. Where that line is and what happens past it is in local landing pages and how many before they start hurting.

The test is simple and it takes ten seconds. If you could swap the place name and the page would still be true, it is a template. If swapping the name would make it false, it is knowledge.

Aim for fewer pages, each of which could only have been written about one place by one person who has been there repeatedly.

Does this actually affect rankings?

It affects two different things and it is worth separating them.

For ordinary local search, the profile does most of the work for the searches that name you or name a service near a place. Content does not substitute for the profile fields, and the block those searches resolve into is explained in the local pack explained for people who do not work in SEO.

For the neighborhood question, which is where agents actually compete, published specificity is most of what there is to compete on. And there is a second reason it matters now: an AI answer about a neighborhood is assembled from whoever published real detail about it. If nobody local published anything, the answer gets built from portal boilerplate. How that works, and what a business with no published detail can expect, is in getting into AI answers when you have no.

Neither of those is a promise about traffic. It is a description of what the alternative is, which is not competing at all.

What about the geography problem?

It is separate from the content problem and gets confused with it constantly.

Agents often find they appear for one part of town and not the next one over, and conclude they need more pages. Usually they do not. Proximity does a lot of work in local results, and a page cannot move a pin. What is going on and what actually helps is in why you rank in one part of town and not another.

Publishing knowledge about a neighborhood you are not near will not make you appear in the local pack there. It can make you the person somebody reads and then calls, which is a different and slower path, and the honest version of the plan says so.

What goes on the profile rather than the site?

The service list, and it is the field agents most often leave empty.

Buyer representation, seller representation, relocation, probate, trust sales, first-time buyers, multi-unit, whatever you actually do, one entry each in the words a client would use. Most agent profiles carry a category and a phone number. The field-by-field version is in adding services and products to your profile correctly.

Two more profile notes for this category. Your name field is your name, not "Jane Smith, Top Producing Realtor," because taglines in the name are prohibited. And if you work from home, set the service area properly rather than publishing a home address you will regret publishing.

What does one of these pages actually look like?

Short, answerable, and written the way you would say it.

Take a real question. "Is the noise from the freeway bad in this part of town?" A portal cannot answer it. A market report will not. You can, in four hundred words, because you have stood in those rooms with the windows open.

The structure that works: the question as the title, the honest answer in the first two lines, then the detail that makes the answer credible. Which streets, which direction, what time of day, what the double glazing does and does not fix, and what it tends to do to the price. Then one line about who it suits and who it does not.

Finish with what somebody should do next, and make it small. Not "contact me for a consultation," but something like: if you are looking at a specific address, send it over and I will tell you what I know about that block.

The credibility comes from being willing to say something negative. A page that says the noise is fine everywhere is a page nobody believes. The agent who says "these two streets, avoid the west-facing units" is the one the reader trusts on everything else.

Photographs help here more than stock imagery ever does. Your own photographs, taken on your phone, of the actual street. They are worse pictures and better evidence.

How much of this is realistic for one person?

Two pages a month, forever, beats twenty pages once.

The failure mode is not laziness, it is a burst. An agent writes fourteen neighborhood pages in a fortnight, publishes nothing for eight months, and concludes content does not work. What actually happened is that the site got a pile of thin pages and no pattern.

Two a month is a page every other Tuesday. It compounds, it stays accurate because you are close to it, and it is small enough to survive a busy quarter without becoming the thing you feel guilty about. Write from the questions you were actually asked that fortnight, which also solves the problem of what to write about.

Keep a note on your phone of every question a client asks that you had to think about. That list is the editorial calendar, and it is better than any keyword tool because it is made of questions somebody actually asked while deciding whether to spend money.

What should you not do?

Three things, all common, all counterproductive.

Do not publish AI-written neighborhood pages. This is the one we see most and it is self-defeating in a way people do not notice until later. They are indistinguishable from the boilerplate you are trying to beat, they contain the errors you are supposed to be correcting, and the whole argument for publishing was that you know something a database does not.

Do not put a market statistic on a page without saying where it came from and what period it covers. An unsourced number is the fastest way to lose a reader who knows the market, and buyers increasingly do.

And do not turn the pages into listing bait. A page that promises neighborhood insight and delivers a form and three listings has broken the only promise that got the visitor to read it, and that visitor now knows something about how you work.

There is a fourth, quieter one. Do not delete a page because it stopped getting traffic in its first three months. Local knowledge pages accumulate slowly and are read by people who are close to a decision rather than by crowds, so judging them on early traffic is judging them on the wrong number entirely.

Where these numbers come from

There are no statistics in this article. We could not read a primary source for how buyers choose an agent, what neighborhood content is worth in enquiries, or how often an AI answer cites a local publisher, and the figures in circulation come from portal operators and content vendors describing their own products. Where this article recommends a cadence, that is our operating judgment and it is labeled as that.

The name and service area rules, the category test and the prohibition on marketing taglines in a business name are from Google's guidelines for representing your business, read 2026-09-07.

The disclosure that has to appear on solicitation materials, which the statute defines to include electronic media, is from California Business and Professions Code section 10140.6, read 2026-09-09.

Google's own guidance on what a merchant may and may not do when soliciting reviews, which agents ask about constantly, is in its prohibited and restricted content policy, read 2026-09-07.