Real Estate Marketing 2026: When Portals Lock Down the AI SERP
Summarize this article with AI
Avant d’entrer dans le détail, un chiffre qui résume tout : sur les 37 requêtes transactionnelles que nous avons auditées, 31 sont verrouillées en position 1 par un portail. Voici la répartition brute.
Le mur de verre en un coup d’œil
Sur les 37 requêtes transactionnelles auditées, un portail occupe la position 1 dans 84% des cas.
The glass wall: when Zillow, Realtor.com, and Redfin hold 31 of 37 positions
A real estate developer calls me on a Tuesday morning. He has 14 projects. 945 pages live. 9,500 organic sessions per month.
The glass wall.
Across the 37 queries we audited (« buy apartment + city », « price per m² + neighborhood », « new development + municipality »), 31 show a portal in position 1. Zillow. Realtor.com. Redfin. According to Neil Patel, in his 2026 real estate marketing guide, these three players lead almost every high-volume query. And AI Overviews compress what’s left of informational clicks.
I see the same thing with my clients. A 23% drop in non-brand organic clicks over 6 months. Your content holds up. Google has picked its side: structured data, verified by millions of reviews.
The diagnosis comes quickly. Your keywords hold up. What’s blocking you is entity positioning. Google knows Zillow as a global brand, with millions of pages, fresh data, and authority signals impossible to match short term.
But there’s a crack.
These portals cover wide. They stay shallow. They miss the street, the neighborhood, the developer, the school, the real commute time at 8:12. They omit signed purchase agreements from your agency. That’s where you win.
So we stop the frontal attack. We build an architecture where every page answers a question the portal doesn’t handle.
945 poorly structured pages look like a catalog. Well organized into cocoons, they become a local encyclopedia.
Here, depth beats volume.
What AI Overview changed: the query belongs to the site that gets cited, not the one that publishes
You type « buy a house in Bordeaux for 200,000 euros ». Google displays an AI panel with three properties. Those properties come from Zillow, Realtor.com, or Redfin. Your results page no longer exists in the user journey.
The compression is real.
Neil Patel says it bluntly: AI Overviews are compressing the few informational clicks left. Before, a user hesitated between your listing and a portal’s. Now, the AI decides for them. It extracts and summarizes. Then it compares. And it cites the sources with the strongest trust signals.
The DOSE framework, which I learned with Guillaume Attias at BMO Academy, starts from a simple idea. The query belongs to the one the AI cites, not the one who publishes. So the work changes.
Detect the real need behind the query. Someone typing « what budget for a 2-bedroom apartment near the tram in Bordeaux in 2026 » is looking for something different from someone typing « apartment prices Bordeaux ». Organize pages by intent. Structure each cocoon so Google understands who you are. Then evaluate, iterate.
A developer client applied this principle across 37 cocoons. One cocoon per project. Each cocoon contains 7 to 12 pages: the neighborhood, price per m², schools, commute times, resident reviews, floor plans, options. These pages already existed. But isolated, competing with each other. Today, they answer each other. They interlink.
Result: 37,000 organic sessions per month. That’s +290% compared to the initial 9,500. Without advertising. Without producing a single additional article.
Content was not the problem. Organization was.
AI Overview focuses on the clarity of internal linking and the coherence of entities, not how much you publish.
To be cited, Google must understand your role in a territory. Publishing 600 articles falls short.
Structured data makes the difference. Schema.org for properties, reviews, addresses, prices. Every tag sends a clear signal. Portals already do it. You can do it too, with a local precision they lack.
La méthode tient en cinq étapes. On arrête la production, on range, on structure. C’est exactement le chemin que nous avons fait parcourir à ce développeur de 945 pages.
Du chantier de 945 pages aux 37 cocoons
Le pipeline de restructuration sémantique appliqué au client développeur.
The problem is not the content. It’s the architecture.
945 pages. 14 projects. 23 catégories. No hierarchy. All pages targeted the same keywords: « new development Bordeaux », « new apartment Bordeaux », « buy apartment Bordeaux ». The site was cannibalizing itself.
The result? No page in position 1. Even the best listings ended up on page 2, behind the portals.
Classic diagnosis.
We stopped production. We restructured. We invested the next $8,000 in the right place: internal linking, tags, structured data, neighborhood content.
The DOSE framework served as our compass. Detecting search intent by neighborhood. Organizing pages into cocoons: one pillar page per project, satellite pages per question. Internal links with descriptive anchors, not « click here ». Évaluation every 6 weeks.
Internal linking became the engine. A satellite page strongly linked to its pillar page sends a relevance signal that Google interprets as specialization. Not an isolated page.
Result after 6 months: 14 transactional queries in position 1 to 3. 47 qualified file requests per month, up from 9 before. The cost per file acquisition divided by 4.
Architecture.
Nothing glamorous. It’s mechanical. Portals don’t do that for you. They capture demand. You capture trust.
We stopped content production. We rebuilt the architecture. 6 months. +290%. No ads.
I’m not selling you the method. I’m showing you the pages.
And the pages speak for themselves.
What matters is the site hierarchy. The homepage points to the pillars. The pillars point to the satellites. The satellites link back to their pillar. Each page has a role. None steps on another.
Create assets portals can’t duplicate: local data, 3D tours, price-per-m² history
Zillow has the listings. You have the street.
The portal aggregates listings. It doesn’t know the bakery open at 7am, the school with a 94% pass rate on the brevet exam, the real commute time to the train station at 8:12, the price per m² signed at the notary.
These are the data points that make the difference in 2026.
According to Neil Patel, the content that performs is neighborhood studies, immersive videos, and updated local data. A site that publishes the real price per m² by neighborhood, updated every quarter from signed purchase agreements, becomes a primary source. Google loves primary sources.
A Bordeaux agency I track launched 47 neighborhood pages. Each page contains: actual price per m², 3-year evolution, schools, shops, transit, and a 90-second video shot on foot. Within 11 weeks, 23 pages appeared in local AI Overviews. No paid ads.
47 pages. Not 470. Not 4,700.
Depth beats volume.
Why? Because Google doesn’t need another page listing apartments. It needs a page that answers « is this neighborhood right for me? ». And that’s something only a local player can do.
Leading portals cannot duplicate a video tour shot in your neighborhood with your clients. Interview the baker? They won’t do it. Signed purchase agreements? They don’t publish them.
Nothing theoretical here. It’s field observation.
Every unique asset strengthens your entity. It becomes a citable source. Journalists, local blogs, even portals themselves can reference it. Indirect traffic starts there.
Getting cited by portals while staying independent: press relations, entities, co-citations
You don’t have the budget to beat Zillow on its turf. But you can become the source Zillow cites.
Portals need data. Journalists need angles. A local market study is currency.
Marie Haynes and Lily Ray have repeated it for years: the future of SEO lies in understanding entities, not keywords. Google must understand who you are, what you do, and why you’re an authority in your area.
In practice:
- Publish a quarterly price-per-m² barometer by neighborhood. Send it to local journalists. They need data. You become the source.
- Create a « Who we are » page with your entity data: legal name, address, SIREN number, social profiles, sameAs.
- Get cited by local press, real estate boards, local governments. Each co-citation strengthens your entity.
- Respond to portal requests when they look for local data. If Realtor.com cites you as a source, you earn a link from an 85+ authority domain.
I observe a snowball effect. A developer who publishes exclusive data attracts citations. Those citations strengthen its entity page. That page, in turn, improves the ranking of its project pages. And no new content is needed.
Traffic doesn’t always come from the direct click. It comes from the citation. The user sees your name in the AI Overview, then types your brand into Google. That’s brand traffic, the kind portals don’t capture.
Organic search in 2026 is entity building. Not stuffing.
Measure citations. Use Google Search Console to watch impressions climb for your brand. The stronger your entity, the more queries like « your brand + reviews » or « your brand + project » trigger rich results.
The adjacent e-commerce lever: why agencies, proptechs, and service providers must also build cocoons
Do you sell rental management software? Virtual tour solutions? Home staging? Same fight.
AI Overviews don’t spare the niches living next to real estate. Type « rental management software reviews »: the AI panel cites three players. If your site looks like a feature catalog, you stay invisible.
Same principle. Build semantic cocoons around each customer problem. Forget « our solution ». Your pages talk about « how to reduce rental vacancy by 23% », « 2026-compliant lease template », « maintenance checklist for landlords ».
A proptech client, 200 pages, restructured into 18 cocoons. Within 3 months, 11 pages reached position 1 on high-intent long-tail queries. Nothing magical on the content side. I just linked the pages together.
This also applies to commercial real estate. Agencies, developers, broker networks, coliving platforms face the same portals, the same aggregators, the same AI Overviews.
The line between e-commerce and real estate is blurring. A real estate agency becomes a local content publisher. A proptech becomes a practical guide publisher. A service provider becomes a data source.
Put away the content strategy. Here, we’re talking brand architecture.
You will do it. The real question is speed: structuring your pages before the AI cites your competitors.
I’m writing from Southeast Asia, where real estate is booming. Local portals follow the same mechanics, with the same solutions: cocoons, exclusive data, entities. Countries change. The method stays the same.
Measuring what still matters: conversions, not clicks
37,000 organic sessions. That’s a number.
What matters is what those sessions produce. A click doesn’t sign a purchase agreement. Neither does a call. A qualified file request does.
With AI Overviews, the overall click-through rate on organic results drops. Neil Patel observes it: the share of informational clicks to non-portal sites is compressing. Measuring clicks is no longer enough. You need to measure assisted conversions.
A phone call. A completed form. A booked site visit. A notary appointment. Those are your metrics.
For the developer I worked with, the KPI that mattered was not the session. It was the number of accepted financing files per month. Before the restructure: 9 per month. After: 47. The cost per file acquisition went from €1,100 to €260.
47 files. Not 47,000 visits.
Conversion becomes the only language Google indirectly understands. If your page generates calls, bookings, requests, Google interprets it as a quality signal. And keeps it ranking.
Set up call tracking, GA4 events, form tracking. Not for show. To know which pages of your cocoon generate files.
Portals can’t measure what you measure. They measure clicks. You measure signatures.
The battle is played here.
Is your site structured to be the irreplaceable source?
Real estate SEO audit: your structure under the microscope
For 45 minutes, I audit your internal linking and cocoons live. You leave with 3 priority actions to stand out against portals and in AI Overviews.
Book a strategic call — 45 minFrequently Asked Questions
Why do Zillow, Realtor.com, and Redfin dominate Google results in 2026?
Because they have millions of pages, abundant structured data, verified reviews, and strong entity signals. Google treats them as trusted real estate brands. Independent sites rank far behind.
Are AI Overviews killing real estate SEO?
No. They change the game. AI Overviews cut informational clicks to non-portal sites. In exchange, they highlight primary sources, unique data, and well-understood entities. SEO becomes a matter of internal linking and exclusive data.
What is a semantic cocoon in real estate?
It’s a group of pages linked together on the same theme, for example a real estate project or a neighborhood. The pillar page covers the topic in depth. Around it, satellite pages answer specific questions. And the links between those pages help Google better understand the subject.
How long does it take to beat a portal on a transactional query?
It depends on the competition. On my client sites, when I restructure into cocoons and publish unique local data, results come in 6 to 11 weeks on long-tail queries. On very high-volume generic queries, it takes more like 6 to 9 months.
Can local real estate agencies compete with portals without an advertising budget?
Yes, on one condition: don’t copy the portals. When an agency publishes what Zillow doesn’t have — signed price-per-m² data, neighborhood video tours, interviews with local shop owners — and structures its site into cocoons, it gains visibility and attracts qualified files independent of ad auctions.
