Programmatic pages are justified only when the answer materially changes

A practical way to decide when a scalable property page deserves its own searchable URL and when it is only a filter state with a different label.

2026-09-10

This article reflects the named expert’s practical perspective. See NovAsia’s editorial policy for how material is prepared and reviewed.

A large property database makes page creation feel almost free. A city can be combined with a district, a budget, a bedroom count, a completion status and a property type. A few rules can turn those combinations into thousands of URLs before anyone has asked whether thousands of separate pages are useful.

That is the point where I stop thinking about keywords and ask a product question: does this URL give the buyer a materially different answer?

Google's spam policies do not define scale itself as the problem. The concern is scaled content created primarily to manipulate search rankings while adding little or no value for users. That distinction matters for property sites because programmatic publishing can be genuinely useful. A catalogue often has structured data that would be impossible to maintain manually. The risk is using the existence of that data as proof that every possible combination deserves to become a search landing page.

A different query string is not yet a different decision

Suppose a site can generate pages for “Pattaya condos”, “Pattaya condos near the beach” and “Pattaya condos near the beach under THB 6 million”. Those phrases are progressively narrower, but the pages should not be judged by wording alone.

The second page earns its place if “near the beach” changes what must be explained. The answer may need an actual access route, a distinction between sea view and beach access, neighbourhood differences, and projects that meet the condition in meaning rather than only in straight-line distance. The third page may deserve a separate role if the budget substantially changes the available areas, project types and compromises the buyer needs to understand.

But if every page uses the same introduction, the same projects, the same advice and the same conclusion, with only a number or district name swapped, the site has created separate URLs without creating separate answers.

Data variation is strongest when it changes the page's reasoning

A scalable page does not need a hand-written essay to be useful. In many catalogues, data can do more work than prose. A district page may have a distinct set of projects, a map bounded to that district, relevant comparisons, local transport or lifestyle anchors, and factual constraints that are actually specific to the location.

The important part is that the page-specific information affects what the buyer can conclude.

A weak template often has unique values but generic reasoning. It might show a different average price, a different project count and a different H1 while every explanatory paragraph could be copied to another city unchanged. That is better than a pure keyword swap, but it still leaves the user with an oddly thin answer.

A stronger template lets the underlying data change the explanation. If the selection is dominated by completed resale units, say so and explain what that changes for the buyer. If a district page has few relevant projects, the answer should not pretend the market is broad. If the page is built around a budget that excludes most new developments, that limitation belongs in the useful part of the page, not in a hidden filter state.

Some combinations are filters, not destinations

Faceted navigation is valuable because people need to refine a catalogue. That does not mean every filter result should become an indexable landing page.

A filter can be temporary and personal: two bedrooms, pet-friendly, high floor, completion next year, a narrow budget band. The same buyer may change three of those choices in a minute. Turning every combination into a permanent search page can create a huge set of near-duplicates whose only purpose is to preserve one momentary configuration.

I prefer to separate navigation from editorial architecture. Filters help an individual explore. Search landing pages should represent recurring, recognisable tasks that deserve a stable explanation and a stable place in the site hierarchy.

That distinction also improves maintenance. If a page has a defined role, the team knows when its content is stale and which data fields matter. If it exists only because a generator can produce it, nobody owns its usefulness after launch.

Architecture should be designed before the generator

The safest way to scale is to define page types first.

For a property site, those types might include country hubs, city catalogues, district pages, project pages, ownership guides, comparison pages and a limited set of high-value budget or lifestyle pages. Each type should have a clear user question, evidence requirement and relationship to neighbouring pages.

This prevents two common failures. The first is fragmentation: one useful answer is divided across dozens of thin pages. The second is collision: several templates accidentally target the same task and then compete for the same internal links and search demand.

Internal linking is a useful reality check here. If the team cannot explain where a programmatic page belongs in the user's journey, the search engine is being asked to infer a purpose the product itself does not express. A page that has no natural parent, no sensible next step and no reason to be linked except from an automated index may be a database state rather than a destination.

Scale needs an editorial stopping rule

Before launching a large family of pages, I would define the conditions under which a combination is not published as a searchable URL.

One rule might be that the page must have a distinct inventory or dataset. Another might require a different decision framework or a minimum set of page-specific fields. A page could be withheld when it would simply reproduce a neighbouring page with fewer results. These are product rules, not universal SEO thresholds.

The purpose of a stopping rule is to prevent the generator from making editorial decisions implicitly. Without one, every new database dimension creates pressure to multiply the indexable site.

After launch, I would not treat the number of indexed pages as the success metric. I look for whether each page family attracts the type of query it was designed to answer, whether neighbouring pages overlap heavily, whether users have a meaningful next step, and whether thin combinations accumulate without engagement or distinct search demand.

The page should survive one simple test

Imagine the page receives no search traffic tomorrow. Would it still be worth linking to from the site because it helps a buyer understand a meaningful choice?

If yes, the programmatic URL probably has a product reason to exist. If the only argument is that someone might type the exact phrase into a search engine, the justification is much weaker.

Programmatic publishing is most powerful when it scales genuine differences: different inventories, different constraints, different comparisons and different decisions. It becomes fragile when it scales only labels. The generator should multiply useful answers, not multiply the appearance of coverage.

Sources

• Google Search Central — “Spam policies for Google web search”, scaled content abuse section: defines the risk of producing many pages primarily to manipulate rankings while providing little or no user value. Accessed 2026-10-06.
• Google Search Central — “Creating helpful, reliable, people-first content”: frames useful content around an intended audience, a clear site purpose and a satisfying answer to the user's goal. Accessed 2026-10-06.