When I built REN.PH, I indexed over 60,000 pages. Broker profiles, provincial zonal value pages, city and municipality pages, barangay pages. All verified, all traceable to government sources, all doing the job I designed them to do.
I also had developer and project profile pages. Hundreds of them. Pages for Vista Land, SMDC, Ayala Land Corp, DMCI, and their respective projects across the country.
I deindexed all of them.
Not because they were broken. Not because someone complained. Because they didn't meet the standard I set for everything else on the platform.
Why I Built Them in the First Place
I didn't build these project pages as a random SEO play. I started from one of the most important datasets in Philippine real estate: DHSUD's License to Sell records.
Every legally sellable real estate project in the Philippines must have a License to Sell from the Department of Human Settlements and Urban Development. No LTS, no legal right to sell. It's the clearest compliance signal in Philippine real estate, and exactly the kind of verified government data REN.PH is built on.
The plan was straightforward: pull the LTS data from DHSUD, build project pages around legally compliant projects, then group them under their respective developer pages. Start from verified government records, build outward. The same approach that works for broker verification (PRC board exam data) and zonal value pages (BIR schedules).
On paper, it made perfect sense.
What I Found When I Started Building
The LTS data itself was solid. I could verify which projects had active licenses. That part worked.
But a project page needs more than a license number to be useful. As I started researching the supplementary data (pricing, unit models, amenities, photos, site details), I hit a wall. I realized that despite the verified LTS data, most pages weren't going to provide meaningful insight for users.
The content was thin. Most pages ended up template-generated with minimal unique information beyond the LTS status. A project page that tells you the developer name, a location, and that it has a License to Sell, but can't show you pricing, photos, or unit details, isn't helping anyone make a buying decision. It's just taking up space in an index.
The visuals were missing. This is real estate. People need to see what they're buying. A project page without renders, site photos, or even a proper map is like a restaurant listing with no photos of the food. Worse, actually, because a wrong photo in real estate is a liability, not just a disappointment. Most projects didn't have publicly available visual assets I could verify and use.
The pricing data was unreliable. Developer pricing in the Philippines changes constantly. Quarterly revisions, broker-specific rates, promos that appear and disappear. Unlike LTS status, which is a binary government record (you have it or you don't), pricing is a moving target with no single source of truth. I couldn't guarantee that the prices on those pages were current. Putting an unverified price on a page that sits next to government-verified data is how you erode trust across an entire platform.
A verified compliance flag is necessary for trust, but insufficient for usefulness.
The one verified data point I did have, the LTS status, wasn't enough to carry a page on its own. And everything I needed to add around it couldn't meet the same verification standard.
On the rest of REN.PH, every data point traces back to a government source. Broker verification comes from PRC board exam records. Zonal values come from BIR schedules. Provincial and barangay data comes from official geographic classifications.
The project pages had one link in that chain and gaps everywhere else. Pages with gaps sitting next to pages with full provenance don't just look inconsistent. They undermine the pages that earned their place.
Why Deindex Instead of Just Improving Them?
I chose to deindex because the timeline to improve them properly was months, and keeping them live would have been actively harming the trust I worked hard to build across REN.PH.
Search engines evaluate your site as a whole. If you have 60,000 strong pages and 500 thin ones, those 500 drag down the quality signal of the entire domain. Google's own documentation on helpful content makes this clear: site-wide quality signals matter. Your best pages pay the tax for your worst ones.
For AI citation, the stakes are even higher. When an LLM retrieves data from your site, it doesn't add a disclaimer that says “this might be from one of their weaker pages.” It presents whatever it finds as your answer. If ChatGPT pulls an outdated project price from REN.PH and presents it to a prospective buyer, that buyer makes a decision based on wrong information. The trust damage doesn't land on that one page. It lands on the entire domain.
I already have proof that this matters. ChatGPT currently cites REN.PH for broker verification queries. It does so because the verification pages are explicit about what they confirm (PRC board exam passing) and what they don't (current license status, CPD compliance, DHSUD accreditation). That honesty, scoping the claim precisely, is what earned the citation.
Keeping pages indexed that couldn't meet that same standard was a risk I didn't need to take.
The Technical Approach
I didn't delete the pages. I applied noindex, follow meta tags.
The pages still exist. The URLs work. Internal links still pass value through them. The site structure is preserved. But search engines won't show them in results, and they're far less likely to be retrieved or cited by AI systems.
When I'm ready to rebuild these pages properly, I don't have to start from zero. The URL structure, the data models, the relationships between developers and projects and locations are all intact. I just need to fill them with content that meets the standard.
Think of it as pulling a product from the shelf, not shutting down the factory.
What These Pages Need Before They Come Back
Before I re-index any developer or project page, I personally ensure it clears a strict checklist:
Accurate, current project information. Name, developer, location, status. All verified against the developer's latest published data.
At least one unit model with current pricing. Not a number I scraped six months ago. A price I can stand behind today.
Working calculator pre-loaded with real data. So a visitor can estimate monthly payments, transfer costs, and total acquisition cost without leaving the page.
Financing information. Pag-IBIG eligibility, bank financing options, in-house terms if available.
Photos or renders. Actual visuals of the project. Not stock images, not placeholder graphics.
Contact path. A clear way to connect with a verified broker who handles that project.
If a page can't check those boxes, it stays noindexed. The bar is the bar.
The approach from here is demand-driven. I look at what people are actually searching for, use Search Console data and broker inquiry patterns to identify which developers and projects to build first, then do the research properly for each one. Each page earns its way back into the index by meeting the standard.
What This Means for the Bigger Picture
There's a version of this decision that looks like retreat. Fewer indexed pages. Less search surface area. A temporarily smaller footprint.
But this was never about having the most pages. It's about having pages that tell the truth. Every page in your index is a claim: that the information on it is accurate enough to be cited, trusted, and acted on. If you can't back that claim, the page shouldn't be making it.
The 60,000+ pages that remain indexed on REN.PH earned their place. They have verified data, clear provenance, and explicit scope. Developer and project pages will earn their place too, when they're ready.
I'd rather leave a gap in my index than risk putting a lie in it.
Aaron Zara is a builder and operator based in the Philippines. Founder of GodMode.ph and builder of Ren.ph, a 60,000-page verified Philippine real estate platform built and operated by one engineer.