Discussion
  • Technical SEO

Property portals: keep sold and leased listings live as price-history pages, or remove them? Which helped area-page rankings?

Published 1 reply

Discussion starter: a programmatic SEO question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.

Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.

1 reply

  • admin
    Admin
    Original poster

    Short answer: keep the sale and lease history, but not necessarily as the old listing URLs. The model I would build towards is one persistent page per property that absorbs the history of every listing at that address, with ended listing URLs redirected into it. Records that carry no useful data and attract no searches, links or leads can return a 404 or 410. What I would avoid is keeping millions of “this property has sold” listing pages indexable by default.

    On the area-page question, public evidence is thin. No controlled public test that I can point to shows that keeping sold listing URLs live lifted suburb or area rankings. The plausible route from sold activity to a stronger area page is the area page itself carrying sold data (recent sales, prices, time on market) and linking to the properties, and you can publish that whatever you decide about the individual listing URLs. So I would make this call on your URL model, crawl logs and lead data rather than on how many URLs stay indexed.

    The real choice: listing URLs or property URLs

    A listing is an event: a campaign to sell or lease a home, with a start and an end. A property is what persists: the address, the land, the building and its history. Portals built around listing IDs create a new URL for every campaign, so a house that has sold twice and been leased four times can leave six overlapping URLs behind, each showing a slice of the same history. “Keep sold listings as price-history pages” on that model means keeping all six.

    A listing page is like a classified ad, and a property page is closer to an entry in a land title register. The ad is useful while the campaign runs and then goes stale. The register entry becomes more useful each time a sale is recorded against it. Someone searching what 12 Smith Street sold for wants the register version: one place with the full history. The analogy has a limit. Your property page is not an official record, and the data behind it (sale prices, photos, floor plans) comes with licence and privacy conditions covered further down.

    If you already run persistent property pages, most of the decision is made: ended listings redirect into the property page, and each property page earns its place in the index on its own merits. If you don’t, building that layer is a bigger and more valuable project than fine-tuning status codes on listing URLs. The hard part is address matching: units, subdivisions, renamed streets and knock-down rebuilds all break a naive one-address-one-page model.

    What each option does to crawling, indexing and area pages

    Large portals are the audience Google’s crawl budget guide was written for: sites with more than a million unique pages whose content changes moderately often (about weekly), or 10,000+ pages changing daily. The guide asks you to return a 404 or 410 for permanently removed pages, describes a 404 as “a strong signal not to crawl that URL again”, tells you to eliminate soft 404s because those pages keep being crawled, and recommends consolidating duplicate content so crawling focuses on unique content rather than unique URLs. Google’s HTTP status code documentation adds that all 4xx errors except 429 are treated the same, and that Search Console shows a soft 404 when a page’s content suggests an error, an empty page or an error message.

    Option What Google receives Effect on area pages Main risk Evidence to inspect
    Keep each ended listing live and indexable A 200 page per campaign, overlapping with any other campaigns at the same address Adds upward internal links; adds no new information to the area page Many thin, overlapping URLs; pages reduced to a “no longer available” message can be reported as Soft 404 Page indexing statuses for ended URLs; their share of Googlebot requests
    Keep live with noindex A page Google still has to fetch to see the noindex Little; links from pages outside the index are weak support Saves index space more than crawl requests “Excluded by ‘noindex’ tag” counts; log hits on those URLs
    Return 404 or 410 Content treated as gone Removes whatever internal links those pages sent upwards Loses address-query traffic and any external links to those URLs Clicks, queries and referring domains for the URLs before removal
    Redirect to the area page A redirect to a page that doesn’t describe the home searched for Visitors land on a list, not the record they wanted Not a like-for-like replacement, so don’t expect it to behave like one “Page with redirect” counts; engagement of redirected visits
    Redirect into a persistent property page One URL per address that accumulates history Property pages link up to their area; area pages link down to recent sales Needs address matching and data quality work Match rate of listings to properties; clicks and leads on property pages

    One option is missing from the table on purpose: a canonical from the ended listing to the property page. It is a weaker version of the redirect, because the two pages don’t look alike, and a canonical between dissimilar pages is easy for a search engine to override.

    How sold data can help area pages, and where it can’t

    An area page serves two different jobs. “Houses for sale in [suburb]” is an inventory query: the page earns its ranking with current listings, useful filters and freshness. “[Suburb] house prices”, “sold [suburb]” and “median price [suburb]” are data queries: the page earns its ranking with recent sales, prices and trends. Sold data mainly helps the second group. If you report area-page rankings as one number, a real gain on price queries can hide inside a flat result on inventory queries, so split the two before judging anything.

    There are three routes from sold data to area-page performance, and they don’t carry equal weight:

    1. Data on the area page. A recent-sales module with prices, dates and links to property pages, plus median price and time on market. This is the lever I would expect to matter most for price queries, and it doesn’t require individual sold URLs to be indexable at all.
    2. Internal links. Property pages link up to their area page through breadcrumbs, and area pages link down to recent sales. The upward links are real, but each comes from a page with little demand, so I would treat their contribution as small per page and unproven in aggregate.
    3. Crawl allocation. The case for removal is that crawling spent on dead listings is crawling not spent on new listings and area pages. That only holds if crawling is actually constrained. Google’s guide says crawl demand is driven partly by popularity (more popular URLs tend to be crawled more often) and staleness (recrawling often enough to pick up changes). Old sold listings attract little traffic and don’t change, so Googlebot may already visit them infrequently. Your logs answer this, not your URL count.

    Treat leased records differently from sold ones

    Rentals turn over faster than sales, so one property can collect several leased URLs between sales, and each leased record describes a home someone now lives in. Interior photos of an occupied home are a security and privacy problem, whatever their SEO value. Listing photos are also commissioned to market a campaign, so check that your licence covers showing them after it ends, and keep a fast process for owner removal requests. My default would be to keep the factual record (dates, the price where you’re permitted to show it, bedrooms, land size) and strip interior photos from ended listings unless the licence and the owner clearly allow them.

    I would also expect less search demand for past rents than for past sale prices, but that is an assumption to check in your own query data. If leased records earn nothing, fold them into the property page’s history with a redirect, and don’t index a property page whose only content is lease records nobody searches for.

    What I would check before choosing

    1. Search demand for ended listings. Filter Search Console performance data by URL pattern and pull 16 months of clicks, impressions and queries for sold and leased URLs, through the API at this scale, since the interface caps the rows it shows. Look for address queries, “sold” queries and agent-name queries. These are the URLs a blanket 404 would throw away.
    2. How Google already treats them. In the Page indexing report, check how many ended-listing URLs sit in “Crawled - currently not indexed”, “Discovered - currently not indexed” or “Soft 404”. Large counts there mean Google has already declined to index many of them, which weakens the case for keeping them as indexable pages.
    3. Crawl share and freshness. From server logs: the share of Googlebot requests by listing status, how often ended URLs are recrawled, and the median time from a new listing going live to Googlebot’s first request. The Crawl stats report groups requests by response, file type, purpose and Googlebot type, not by listing status, so you need logs for this.
    4. External links. Referring domains pointing at ended listing URLs, such as agency sites or local news stories about notable sales. Those URLs deserve a redirect into the property page, not a 404.
    5. Leads. Appraisal requests, “track this property” sign-ups, agent enquiries and finance referrals that start on sold pages. Sold data attracts homeowners checking what nearby homes sold for, and some of them are future sellers, which is the audience agents pay portals to reach. Removing pages that feed that funnel can cost more than any SEO gain.
    6. Area-page content. Whether area pages show recent sales today. If they don’t, adding that module is the cheapest test of the area-page question.

    Put together, those checks give a default rule for each ended listing:

    A listing ends (sold or leased)
    ↓
    Can it be matched to a persistent property page? If yes, redirect the listing URL into it.
    ↓
    No property layer yet: does the listing URL earn clicks, external links or leads? If yes, keep it live with the full record until the property layer exists.
    ↓
    No record worth showing, no clicks, no links, no leads: return 404 or 410
    ↓
    Property page: index it when it has recorded history, search demand or links; otherwise keep it out of the index

    Hypothetical example:

    A portal has 250,000 active listings and 6 million ended sale and lease listing URLs covering 2.2 million properties. Logs show ended listings take 40% of Googlebot requests, and new listings wait a median of two days for a first crawl. Search Console shows 180,000 ended URLs earned clicks in the past year, mostly on address queries, while 2.5 million sit in “Crawled - currently not indexed”. The 180,000 earning URLs are the reason not to delete everything; the crawl share, the two-day lag and the 2.5 million unindexed URLs are the reason not to keep everything. The flow above resolves both.

    A test that separates the two levers

    Portals have an advantage most sites lack: hundreds of comparable area pages, which makes a geographic split test realistic. Because two different levers are in play, I would use three arms rather than two:

    Arm Change What it isolates
    A. Control No change Baseline, including seasonality and any core update
    B. Data module Recent-sales module and links to property pages on area pages; ended listing URLs untouched The effect of sold data on the area page
    C. Data module plus consolidation Arm B, plus ended listing URLs redirected into property pages and empty records returning 404 The extra effect of cleaning up listing URLs

    Match areas across arms by listing volume, price band, current area-page traffic and seasonal pattern. The primary metric is area-page clicks, split into inventory queries and price queries with a query filter (for example, queries containing “sold”, “price”, “median” or “value”). Secondary metrics: time to first crawl and indexing for new listings in treated areas, property-page clicks, Soft 404 and redirect counts in the Page indexing report, and leads by source.

    Set the duration from your logs rather than a default. Redirects and 404s only take effect once Googlebot re-fetches the URLs, and old listings may be recrawled only every few weeks or months. If ended listings are recrawled about every 60 days, a test that stops at week four has measured very little in arm C. Start the clock when most treated URLs have been recrawled, then run long enough to cover several weekly cycles of listing volume. Avoid launching just before your market’s seasonal peak in listings, and if a core update lands mid-test, the control arm is what keeps the result readable.

    What would change the decision:

    • B beats control on price queries and C adds little: the data did the work. Consolidate listing URLs for crawl hygiene and user experience, but don’t sell the clean-up internally as an area-page ranking project.
    • C beats B: the clean-up added something, probably through crawl allocation or less duplication. Roll it out and keep watching leads.
    • Property-page clicks or leads fall in C: some ended listings were earning. Loosen the removal rule and redirect more of them into property pages instead of returning 404.
    • Nothing moves: sold data isn’t what holds your area pages back. Look at inventory depth, filters and internal linking.

    Where I would start

    If this were my portal, I would stop treating “sold” as a status on a listing and start treating it as an event on a property. The first step is the logs and lead data, because they show whether removal saves anything and what it would cost. The second is the property layer, because it lets you keep the history without keeping the clutter. Only then would I run the three-arm test, and I would not claim an area-page ranking gain from either lever until it has been separated from the other.

    Need help with this on your own site?

    Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.

    See SEO services

Add a reply

Replies are open to approved members. Log in or apply to join.