Why Portals Outrank You for Your Own Project Name

By Tessar Napitupulu, Founder & CEO, PT Arfadia Digital Indonesia · July 2026 · 9 min read

A developer searches their own project name on Google and a listing portal answers first, not the official site. This is the single most common complaint developers bring to an SEO review, and it happens for a structural reason: portals such as Rumah123, 99.co, and Lamudi carry decades of accumulated domain authority and index millions of pages, so they win almost every generic query by default. The fix is not to compete harder on keywords. It is to build entity ownership, the layer of authority a portal listing structurally cannot replicate.

Why does this keep happening to serious, well-funded projects?

It is tempting to assume this only happens to developers who neglected SEO. It does not. Rumah123 alone indexes more than 1.8 million active listings and, together with 99.co, now sits under a single ownership structure following a 2026 joint venture between REA Group and 99 Group. As of February 2026, Rumah123 drew over 2.28 million monthly visits, more than triple 99.co's 606,680 and nearly six times Lamudi's 383,760 (Semrush, February 2026). That scale is not something an individual project page needs to out-rank on generic terms like "rumah dijual [kota]." It does need to win the one query type a portal cannot own by default: the project's own name.

Here is what that actually looks like in a results page.

Illustrative SERP pattern based on the branded-query complaint documented across independent research into this problem. Individual rankings vary by project and market.

What is actually winnable if the broad keyword is not?

Two things a portal structurally cannot take from a developer: the project's own name, and the neighbourhood the project sits in. Both are entity-authority problems, not keyword-density problems, and both respond to the same underlying fix.

Google Business Profile completeness is the fastest lever. A marketing gallery that has not sold a single unit yet still needs a fully optimised profile, because for navigational queries that combine a brand with a location, the Map Pack frequently appears above every organic result, portal listings included. Press coverage that carries the project's exact name, published across genuinely independent outlets rather than a single paid placement, builds the kind of third-party corroboration that both Google's ranking systems and AI answer engines weigh heavily. And the on-page project page itself needs enough original substance, generally upward of 2,000 words covering construction status, legal documentation, and neighbourhood context, that a thin, syndicated portal listing has no equivalent to offer.

Does this only matter for Google, or does it carry over to AI search too?

It carries over, and the connection is closer than most teams assume. An analysis of 1.9 million AI Overview citations by Ahrefs in July 2025 found that 76.1% of the URLs cited already ranked in the top 10 of organic results for the same query. There is no shortcut into AI visibility that bypasses ranking well the conventional way first. Sites that do get cited in AI Overviews also see a meaningful organic click-through advantage, Seer Interactive's analysis of 25.1 million impressions in September 2025 found roughly 35% higher CTR for cited sites over uncited competitors on the same query, while sites that never get cited saw click-through fall from 1.76% to 0.61%, a 65% decline. Entity ownership work done for Google ranking is not wasted effort if AI citation becomes more important next year. It is close to the same work.

Is real estate even a category AI search cares about yet?

Less than most categories, which is worth knowing before over-investing in AI-specific tactics at the expense of the fundamentals above. Real estate triggers a Google AI Overview on only around 5% of searches, per research compiled by Perplexity, and separately, FlyDragon's 2026 dataset put the figure at roughly 4.5%, both far below the roughly 48% trigger rate across all search categories combined. That makes real estate the lowest AI Overview vertical of any major consumer category tracked. Entity ownership and neighbourhood content remain the correct long-term investment regardless, both because they are the same asset AI citation eventually rewards and because the trigger rate for this category will not stay this low indefinitely.

Is Rumah123 really the only portal worth worrying about, or does the competitive picture matter too?

Rumah123 is the clear leader, but the portal landscape consolidated further in 2026 in a way worth understanding before assuming a single-competitor mental model. Rumah123 and 99.co both now sit under a single ownership structure following a 2026 joint venture between REA Group, the Australian parent of Rumah123, and 99 Group, which also owns iProperty.com.sg in Singapore. That means the two platforms that together account for well over half the total monthly visits among Indonesia's major property portals are now commercially aligned rather than genuinely independent competitors, which has implications for anyone assuming that outranking one automatically creates room against the other.

Below the top two, the picture is more fragmented and, in places, more genuinely winnable. Lamudi drew roughly 383,760 monthly visits as of February 2026, itself indexing a comparatively modest 49,300 pages, an order of magnitude smaller footprint than Rumah123's 1.8 million-plus listings. Pinhome, a newer entrant focused more on end-to-end transaction facilitation than pure listing aggregation, reached around 590,000 monthly organic visits in early 2026 with a domain rating near 65, evidence that a well-executed, differentiated model can still build meaningful organic presence even in a market with two dominant, now-aligned incumbents. The lesson for a developer or agent site is not that competing head-on with any of these is realistic on generic terms, it is that the competitive landscape below the top two has real texture worth understanding rather than treating "the portals" as one undifferentiated block.

Is there a way to actually monitor AI-driven visibility for this kind of branded-query problem, not just rely on manual searches?

Yes, and the tooling has improved recently. Google introduced dedicated Search Console reporting for AI-driven visibility in June 2026, giving site owners a more direct way to see how their pages perform specifically within AI Overview contexts rather than inferring it indirectly from overall organic performance. For a project actively working through the branded-query problem described here, this is a genuinely useful addition to the manual incognito-search check and the paid-ad-monitoring practice covered earlier, since it surfaces AI-specific visibility data in the same dashboard many teams already check weekly for conventional search performance, rather than requiring a separate manual process to approximate the same insight.

What does the technical side of "entity ownership" actually look like in code?

Entity ownership is not just a content strategy, it has a technical layer that most developer websites skip entirely. The core mechanism is a schema markup pattern called sameAs: an array of URLs on your Organization or LocalBusiness schema that explicitly links your website entity to your Google Business Profile, your official social media accounts, and any verified press or directory listings. Without it, a search engine or an AI answer engine has to infer that your website, your GBP listing, and your Instagram account all refer to the same real-world entity. With it, that inference becomes an explicit, machine-readable fact.

The second technical layer is consistency of NAP, meaning name, address, and phone number, across every single place your project appears online: your own site, your Google Business Profile, property directories, and any press mentions that include contact details. A project listed as "Grand Wisata Bekasi" on one platform and "Grand Wisata Residence Bekasi" on another is, from a machine's perspective, potentially two different entities rather than one strongly corroborated one. This sounds like a minor detail, and in isolation it is, but at scale across dozens of directory listings it is one of the most common reasons entity authority stays diffuse instead of consolidating.

The third layer is making sure the project page itself carries RealEstateListing schema with complete, accurate structured data: price, address, geo-coordinates, and unit specifications. This does not directly fix the branded-query ranking problem on its own, but it is the foundation that everything else builds on, and its absence is one of the most common technical gaps found in a real estate SEO audit.

How do you know if any of this is actually working?

The most direct signal is simple: search your own project name in an incognito browser window, without location personalisation skewing the result, and track your official page's position over time. Position 6 moving to position 3 moving to position 1 over a few weeks is the clearest evidence the entity-ownership work is compounding. Google Search Console's performance report, filtered to your branded query variants, gives a more precise version of the same signal, including impressions and click-through rate rather than just rank position.

A second, less obvious signal worth tracking is whether portal-run paid ads on your own project name start disappearing or dropping in position. Portals frequently bid on developer project names precisely because a developer's organic presence is weak enough that a paid ad can win the click cheaply. As your organic entity ownership strengthens, that arbitrage becomes less attractive for the portal to keep running, and a decline in competing paid ads on your own brand terms is itself indirect evidence the fix is landing.

How long does this actually take to fix?

Meaningful movement on branded queries is usually the fastest win in SEO, often visible within 4 to 8 weeks once Google Business Profile is complete and the project page reaches sufficient depth, because branded, low-competition queries respond faster than broad category terms. Broader neighbourhood and category ranking is a longer game, typically 3 to 5 months in moderately competitive markets and 6 to 12 months in highly competitive ones. Treat the branded-query fix as the first sprint, not the whole project.

Frequently Asked Questions