How to Do SEO for a Real Estate Website

By Guru Editorial | August 18, 2026

When someone asks an AI assistant to recommend a real estate agent, the sourcing pattern is brutal for the biggest names in the industry. A 2026 analysis of 1,567 ChatGPT citations behind local agent recommendations found that 98.1% pointed to exactly two places: the agent's Google Business Profile (57.6%) and the agent's own website (40.5%). Zillow, Redfin, Realtor.com, Trulia, and Homes.com combined for zero citations. If your site isn't built to earn that 40.5%, an AI assistant is naming your competitor instead of you.

That statistic captures both the opportunity and the risk of real estate SEO in 2026. Portals dominate impressions and ad spend, but agent and brokerage websites are what Google and AI engines actually treat as the authoritative, experience-backed source. Most real estate sites aren't built to fill that role. They're wallpapered with thousands of auto-generated IDX listing pages that say almost nothing unique, agent bios that read like form letters, and expired listings quietly rotting into broken links. This guide covers the fixes that actually move rankings and citations: taming the IDX feed, building neighborhood and market hubs, proving E-E-A-T, retiring dead listings the right way, and using RealEstateListing schema for what it's actually good for.

Why Real Estate SEO Plays by Different Rules

Real estate sits in Google's "Your Money or Your Life" category alongside finance and health content, which means the bar for demonstrated experience, expertise, and trust is higher than it is for a typical local service business. A vague "About Us" page and a stock photo of a handshake will not clear that bar, and Google's systems are specifically tuned to notice the gap.

The competitive landscape is also unusual. Most local SEO categories are won or lost against other independent operators. Real estate agents compete for the same head keywords as Zillow, Redfin, Realtor.com, and Homes.com, sites with domain authority most agencies will never match. Trying to outrank a portal for "homes for sale in [city]" is rarely a winning fight. The winning fight is the long tail underneath it, and the pages that prove you actually know the market.

There's a useful wrinkle for the GEO side of this too. Real estate queries trigger a Google AI Overview in only about 4.48% of searches, the lowest rate of any industry category tracked, because most real estate search intent is transactional, hyper-local, and tied to volatile inventory that AI Overviews handle poorly. That doesn't mean AI search doesn't matter here. It means classic organic and the map pack still carry the majority of real estate search demand even as buyers increasingly open a chat window first: a Realtor.com survey found 82% of Americans now use AI tools for housing market information, and ChatGPT is the platform they turn to most, used by 67% of them. Winning real estate SEO in 2026 means treating classic and AI search as two engines you're feeding with the same well-built pages, not two separate campaigns.

Fix the IDX Trap Before It Sinks Your Domain

Almost every real estate site runs on an IDX or MLS data feed, and that feed is the single biggest technical liability on the domain. IDX listing pages typically contain a price, a handful of photos, bed and bath counts, and a boilerplate description pulled directly from the MLS. There's very little unique text on the page for a search engine to reward, and the same listing data is syndicated to hundreds of other agent sites at the same time.

When Google's crawlers encounter the same address, the same square footage, and the same MLS description on hundreds of near-identical URLs across the web, they pick one canonical version to index and largely ignore the rest. If your site is thin on original content and heavy on duplicated MLS data, you're spending crawl budget on pages that were never going to rank, and diluting the topical authority of the pages that could.

The fix isn't to delete your IDX integration. It's to stop treating listing pages as if they were content pages, and to build the correct guardrails around them:

• Noindex the bulk of your auto-generated listing pages. Reserve indexation for a curated set where you've added genuine, agent-written commentary, not the entire feed.
• Set self-referencing canonical tags on any listing page you do keep indexed, so mirrored copies elsewhere don't get credited over yours.
• Exclude bulk IDX URLs from your XML sitemap. A sitemap full of thin listing pages signals low overall site quality and wastes crawl budget that should go to your hub content.
• Add unique, first-person context to any listing you want to rank, why you're highlighting it, what it's really like to live near it, how it compares to similar homes on the market, information the raw MLS feed can't provide.
• Route the SEO investment toward original pages instead of hoping thousands of syndicated listings will eventually rank on their own.

This is the same principle behind building programmatic pages that don't get flagged as thin: volume isn't the problem, undifferentiated volume is. A real estate site with 200 tightly indexed, genuinely useful pages will consistently outperform one with 20,000 noise pages competing with themselves and with every other agent on the same MLS feed.

Neighborhood and Market-Report Hubs: Your Real Moat

If listing pages can't be your differentiation, hyperlocal content has to be. Neighborhood guides and market-report pages are the assets a national portal cannot replicate at your level of specificity, and they're what convinces both Google and an AI engine that your site actually understands a given market rather than just listing inventory in it.

The structure that compounds best is a hub-and-spoke model: a city or metro hub page that frames the market at a high level, linking down to individual neighborhood guides and a market-report section, each of which links back up to the hub and across to relevant active listings. This mirrors how topic clusters and pillar pages work in any vertical, but in real estate the "cluster" is geographic instead of purely topical.

Link equity flows from the city hub through neighborhood and market-report content into a curated, indexed set of listings, kept structurally separate from the noindexed bulk IDX feed.

A neighborhood page earns its ranking with specifics: price bands by property type, inventory trends, school catchment notes, commute times to the nearest employment centers, walkability, and the kind of tradeoff commentary only someone who actually works the area can write. Twenty pages with the town name swapped out are not a neighborhood content strategy. Five or ten pages written from real transaction experience, updated on a real cadence, are.

Market-report pages need the same treatment plus a freshness routine. Pull median price, days on market, and active inventory from your MLS on a monthly or quarterly cycle, update the dateModified, and keep last quarter's numbers visible for context rather than deleting them. A market page that hasn't moved in eight months reads as abandoned to both a buyer and a crawler.

Local and Agent E-E-A-T: Proving You're the Real Deal

Because real estate is YMYL content, Google and AI engines are looking for concrete, verifiable signals that the person or brokerage behind a page actually knows what they're talking about, not just claims to. Generic bios and stock photography actively work against you here.

A bio page that clears the bar includes a license number and issuing state, years actively licensed, NAR or state designations (CRS, ABR, GRI, and similar), specific neighborhoods the agent has closed deals in, an approximate or exact transaction count, and real photography from actual closings rather than a stock handshake. Pair that with genuine client testimonials tied to real transactions and any media mentions or speaking history, and you've given both a search engine and a prospective buyer something to evaluate.

Google Business Profile deserves outsized attention given how AI engines are sourcing agent recommendations right now. Remember the citation split from the intro: 57.6% of ChatGPT's real estate agent citations came from Google Business Profile listings, more than from agent websites. A complete, actively managed profile, correct categories, service area, regularly answered reviews, recent photos, and posts, is not optional local SEO hygiene anymore. It's close to half of your AI visibility. Our guide to optimizing a Google Business Profile covers the specific fields that move the needle for service-area businesses like real estate teams.

Every piece of content on the site, blog posts, neighborhood guides, market reports, should carry a visible author byline linked to that agent's credentialed bio page. This is the same author-entity pattern covered in building E-E-A-T signals that Google and AI engines actually trust, applied to a vertical where the trust bar is unusually high and the payoff for clearing it is unusually concentrated in a handful of citation sources.

Expired, Sold, and Off-Market Listings: Don't Let Dead Pages Bleed Rankings

Inventory churns constantly, and what an agent does with a URL after a listing expires, sells, or gets withdrawn is one of the most consistently mishandled parts of real estate SEO. The two common mistakes are opposite failures: letting every dead listing 404 with no plan, or blanket-redirecting everything to the homepage.

Google's own guidance is straightforward that a 404 is the technically correct response for content that's genuinely gone, and it is not a penalty to fear. The mistake isn't serving a 404. It's redirecting a listing to an irrelevant destination instead. Search engine guidance is explicit that 301 redirecting expired content to an unrelated page, the homepage or a top-level category being the classic offender, is likely to be treated as a soft 404, meaning Google discounts it and no ranking value passes through at all. The right redirect target is the closest matching active listing or the relevant neighborhood hub, never the homepage by default, and you should avoid chaining a redirect from one expired listing to another expired listing, which wastes crawl budget and can break the chain entirely.

The table below is a practical decision framework for what to do at each stage of a listing's life:

A "sold" page is worth more alive than dead. Instead of pulling it down the moment a deal closes, convert it into a brief case study: what it sold for relative to list price, how long it was on market, and a line of context from the agent. That page becomes evergreen proof of results, which is exactly the kind of experience signal E-E-A-T evaluation is looking for, and it avoids the churn of constantly generating and killing URLs.

RealEstateListing Schema and the Structured Data That Actually Pays Off

Structured data on a real estate site has one job: make the relationship between the property, the offer, and the agent or brokerage explicit enough that a machine, whether that's Google's indexing systems or an AI model pulling in a citation, doesn't have to guess. RealEstateListing schema, paired with Offer, RealEstateAgent, and Organization markup for the brokerage, is the current best-practice stack for listing pages, and Person schema tied to a credentialed bio page strengthens the agent-entity connection AI engines increasingly rely on.

The mechanical detail that matters most: generate your structured data from the same system that controls the visible listing content, not from a separate, manually maintained template. Status, price, and availability change constantly on real estate inventory, and a mismatch between what the schema says and what the page actually shows is a well-documented way to lose trust with both crawlers and users. If a listing goes pending, the schema needs to reflect that the moment the page does.

It's worth being precise about what schema will and won't do for you. Adding structured data does not, by itself, produce a Google rich result or a ranking boost for a real estate listing. Google removed HowTo rich results from Search in 2023 and fully removed FAQ rich results from the visual SERP on May 7, 2026, and neither type ever applied to listing content directly. But both FAQPage and Article or BlogPosting schema remain valid, supported schema.org types, and they continue to help AI engines parse and extract content for citation even without a corresponding visual SERP feature. For a market-report or neighborhood guide with a genuine FAQ section, keep the FAQPage markup for that reason alone. Our 2026 schema markup guide breaks down which structured data types are still worth the implementation effort and which ones are legacy holdovers.

Long-Tail Location Intent and How AI Search Is Rewriting Home Search

The keywords that actually convert on a real estate site rarely look like "homes for sale in [city]." That head term is a portal's fight to lose. The volume that agents can realistically win sits in the long tail: "three bedroom homes near [elementary school] under $500k," "condos with parking in [neighborhood] pet friendly," "new construction near [employment center] with a short commute." These queries map almost perfectly to the neighborhood and market-report hubs described earlier, and they're exactly the kind of specific, answerable question an AI engine likes to surface a direct source for. Our keyword research guide for AI and traditional search covers how to build clusters around persona, price band, and feature rather than generic head terms.

The AI layer is moving fast enough that it can't be ignored. Between late 2025 and March 2026, Zillow, Redfin, and Realtor.com all shipped dedicated search apps inside ChatGPT, explicitly targeting the "pre-search" phase where a buyer is still asking exploratory questions before committing to a portal or an agent. That's the same phase your neighborhood and market content is built to answer. And the stakes for showing up are high: the same 2026 citation analysis found that 75% of buyers interview only one agent before choosing, and that ChatGPT typically names a median of just six agents per city when asked for a recommendation. There usually isn't a second chance to make that shortlist.

Google Business Profile and the agent's own website supplied 98.1% of ChatGPT's real estate agent citations in a 2026 analysis; the major listing portals supplied none.

Here's how to sequence the work, roughly in priority order:

1. Audit and noindex the bulk IDX feed so crawl budget and link equity stop leaking into thousands of duplicate pages.
2. Stand up the city hub and 5 to 10 flagship neighborhood pages, written from real transaction experience, not templated copy.
3. Rebuild agent bio pages with license details, designations, transaction history, and a visible author byline on every piece of content.
4. Claim, complete, and actively maintain Google Business Profile, given its outsized share of AI citations.
5. Implement RealEstateListing, RealEstateAgent, and Organization schema sourced from the same system that drives the visible page content.
6. Set the expired-listing decision rules from the table above so no URL is left to rot or blanket-redirected to the homepage.
7. Build out long-tail, persona-based content around the neighborhood and market hubs, and monitor which of it starts showing up in AI answer citations, not just classic rankings.

Guru's audit tooling can surface the IDX noindex gaps, schema mismatches, and citation-readiness issues across a real estate site's full inventory automatically, which matters when the page count runs into the thousands and manual review isn't realistic.

Frequently Asked Questions

Should IDX/MLS listing pages be indexed by Google at all?

Most of them shouldn't be. The bulk of an IDX feed is duplicated across hundreds of other agent sites, so Google typically indexes one canonical version and ignores the rest. Noindex the bulk of the feed, exclude it from your XML sitemap, and reserve indexation for a curated set of listings where you've added genuine, agent-written context.

How many neighborhood pages does a real estate site actually need?

Fewer than most agents think. Five to ten deeply specific, experience-based neighborhood guides will outperform fifty thin pages with the town name swapped out. Depth, price-band data, school and commute detail, and real market commentary matter more than raw page count.

What should happen to a listing page after the home sells?

Don't delete it and don't redirect it to the homepage. Convert it into a short case study showing the sale price relative to list price, days on market, and agent commentary. That preserves the page's E-E-A-T value as proof of results instead of creating another URL to manage.

Does RealEstateListing schema improve rankings on its own?

No. Structured data does not produce a ranking boost or a guaranteed rich result by itself. What it does is make the property, offer, and agent relationship explicit for both Google's systems and AI models pulling citations, which matters more for extraction and trust than for a direct rankings lift.

How is real estate SEO different from local SEO for other service businesses?

Real estate carries YMYL scrutiny similar to health and finance content, competes directly against portals with far higher domain authority for head terms, and deals with constantly expiring, high-volume inventory pages that most local service businesses never have to manage. The winning strategy leans on long-tail, hyperlocal content instead of head-term competition.

Should agents optimize for AI Overviews and ChatGPT, or stick to traditional SEO?

Both, using the same pages. Real estate triggers a Google AI Overview in only about 4.48% of searches, the lowest of any tracked industry, so classic organic and the map pack still carry most of the demand. But a growing share of buyers are asking ChatGPT and similar tools directly, and agent websites and Google Business Profiles are what those tools cite, not the portals.

What's the fastest way for a new agent with limited transaction history to build E-E-A-T?

Lean on what's verifiable now: license number, designations in progress, specific market knowledge demonstrated through detailed neighborhood content, and any mentorship or team affiliation with a more established brokerage. Every closed transaction should immediately become a case-study page to compound that history over time.

How often should market-report pages be updated?

Monthly is ideal for active markets, quarterly at minimum. Pull fresh median price, days-on-market, and inventory figures from the MLS, update the page's dateModified, and keep prior periods visible for trend context rather than overwriting them.

Sources

• The 2026 Real Estate AI Search Report: ChatGPT Names Just 6 Agents Per City, Omni Eclipse
• 82% of Americans Use AI for Housing Market Information, Realtor.com Survey Finds, PR Newswire
• Real Estate Industry: 2026 AEO / GEO Benchmarks, Conductor
• Realtor.com the Latest Portal to Launch Search App in ChatGPT, Real Estate News
• The Essential Guide to Managing Expired Content, Search Engine Journal
• RealEstateListing, Schema.org