Google fully launched its Site Diversity Change on 6 June 2019 to reduce result pages crowded by one website. Google said users would usually see no more than two listings from the same site among the top results, while keeping exceptions when additional results were especially relevant. “Usually” is the essential word: Google did not announce a universal two-URL cap or a site penalty.
June 2019
What changed
Google described this as a change to result composition rather than a new way to rank pages. Pages that already ranked highly could remain relevant, but Google might display fewer additional URLs from the same site so users could see more sources. The launch began roughly two days before the announcement and was separate from the June 2019 Core Update running at the same time.
For diversity purposes, Google said it would generally treat a root domain and its subdomains as one site. It also reserved the right to treat a subdomain separately when that was useful. Google did not publish a formula for those exceptions, a report in Search Console or a setting that lets site owners opt out.
Site diversity is only one possible reason several URLs do not appear together. Canonicalization selects a representative URL from duplicate or very similar pages; deduplication avoids unhelpful repetition; ranking determines relevance and usefulness; sitelinks can cluster shortcuts beneath one result. Keyword cannibalization is an SEO diagnosis for overlapping pages, not the name of a Google penalty.
June 2019 timeline: what Google actually announced
The dates matter because the Site Diversity Change overlapped the June 2019 Core Update. A traffic change in that week cannot be attributed from timing alone.
| Date | Confirmed event | Responsible interpretation |
|---|---|---|
| Around 4 June 2019 | Google said the diversity launch had started roughly two days before it became fully live. | Use this as an approximate rollout window, not a precise page-level change timestamp. |
| 6 June 2019 | Google Search Liaison announced that the change was fully live. | The stated aim was more variety in top web results. |
| Same announcement | Google said users would usually not see more than two listings from one site, with relevance-based exceptions. | Two describes a tendency, not an enforceable ceiling. |
| Same announcement | Root domains and subdomains would generally count as one site, with exceptions. | Moving content to a subdomain is not a dependable way to gain another listing. |
| Same week | Google explicitly separated Site Diversity from the June 2019 Core Update. | Segment the evidence before assigning a cause to a decline. |
| After launch | Google provided no site-owner score, manual control or Search Console report for diversity. | A single SERP screenshot cannot prove the system caused a missing URL. |
Four evidence levels for a same-site SERP change
Label the evidence before deciding what to change. This prevents a historical announcement from becoming an imaginary diagnosis tool.
| Evidence level | What it can support | What it cannot prove |
|---|---|---|
| Google-confirmed | The 2019 launch, the usual two-listing tendency, relevance exceptions and general root-domain/subdomain grouping. | The private rules, weights or exact current treatment of one query. |
| First-party data | Which queries and canonical URLs lost impressions, clicks or average position in Search Console. | That Site Diversity was the causal system. |
| Repeatable SERP observation | Whether several clean, comparable searches show fewer same-site web results. | A universal rule; results vary by query, location, device, time and format. |
| Third-party tool or isolated screenshot | A lead worth checking against first-party data. | Complete query coverage, canonical attribution or a Google classification. |
What it means for SEO
Build one indexable page for each genuinely distinct task—not one page for every keyword variation and not several near-duplicates designed to occupy the same results page. Separate pages when the audience, intent, deliverable, location, product or expected answer format is materially different.
When two useful pages serve adjacent intents, make their roles explicit through titles, headings, internal links and content. When pages answer essentially the same task, consolidation often gives users and search systems a clearer preferred destination.
Do not judge success by how many blue links one domain owns. Measure whether the right landing page earns qualified impressions, clicks and conversions across the topic. A useful support, comparison or documentation page can still create value even when it does not appear beside the main page for a head term.
Misconception vs responsible interpretation
| Misconception | Responsible interpretation |
|---|---|
| Google enforces exactly two results per domain. | Google said “usually” no more than two and explicitly allowed more when especially relevant. |
| The Site Diversity Change demoted large websites. | Google described a presentation change that showed fewer additional same-site pages, not a size-based penalty. |
| A subdomain guarantees another top result. | Google generally groups subdomains with the root domain for diversity and can make query-dependent exceptions. |
| Cannibalization and site diversity are the same. | Cannibalization is a site-architecture diagnosis; diversity concerns the mix of sources displayed in results. |
| Canonical tags can force both pages to rank. | Canonicalization indicates the representative version of duplicate or very similar content; it does not request multiple listings. |
| More pages create more SERP shelf space. | Overlapping URLs can fragment maintenance, links, internal signals and reporting without creating another useful result. |
| A missing second result proves an algorithm update. | Intent, competitors, demand, indexing, canonical selection and SERP features must also be checked. |
| Sitelinks violate the two-result tendency. | Sitelinks are automated shortcuts clustered under one text result and are a different presentation pattern. |
Diagnose fewer same-site listings without guessing
Start with the observable symptom, then rule out more testable explanations. Search Console normally assigns performance data to Google’s selected canonical URL, so inspect canonicalization before assuming a page disappeared.
| Check | Evidence to collect | Decision it supports |
|---|---|---|
| Scope | Affected query, URL, country, device, search type and first sustained date. | One-query composition change or a broader visibility loss. |
| SERP format | Comparable logged-out searches and records of ads, local packs, video, shopping, featured snippets and sitelinks. | Fewer web listings or simply less available screen space. |
| URL performance | Page-filtered clicks, impressions and position before and after the change. | A URL stopped being shown or still ranks for other queries. |
| Canonical/indexing | URL Inspection, declared canonical, Google-selected canonical, indexability and redirects. | Duplicate consolidation, technical exclusion or an eligible independent page. |
| Query overlap | A page–query matrix for the competing URLs. | Useful specialization, unstable substitution or redundant targeting. |
| Intent and competitors | Current result types, winning page purpose, freshness, locality and evidence. | The market’s expected answer changed or another page became more useful. |
| Release history | Deployments, migrations, template changes, internal-link edits and Google’s published update dates. | A plausible technical or system-change hypothesis. |
| Business impact | Conversions, assisted journeys and qualified leads by landing page. | Whether the loss needs intervention or is merely less SERP duplication. |
Practical response
- Export at least 8–16 weeks of Search Console data around the sustained change and keep the country, device and search type comparable.
- Identify the exact canonical URLs Google credits with the affected queries; remember that anonymized queries and table truncation make the visible rows incomplete.
- Create a page–query matrix showing the primary task, intended audience, conversion goal and overlapping queries for each URL.
- Keep separate pages only when each one completes a materially different task or needs a distinct result format.
- Merge true overlap into the strongest destination and use an appropriate permanent redirect when retiring a duplicate URL.
- Align redirects, self-referential canonicals, sitemap URLs, hreflang and internal links around the preferred URL.
- Do not block duplicates in robots.txt to solve canonicalization; Google cannot read a blocked page’s canonical signal.
- Rewrite titles and headings so each surviving page communicates a distinct purpose instead of swapping synonyms.
- Add comparison tables, examples, specifications, local details or first-hand evidence only where they serve the page’s task.
- Build a logical information architecture with concise, relevant anchors so users can reach deeper pages even when only one main result appears.
- Track domain-level topic visibility and URL-level performance separately; include qualified conversions, not only average position.
- Recheck after recrawling and enough demand has accumulated; avoid repeated rewrites based on one daily SERP snapshot.
Page-overlap audit: keep, differentiate or consolidate?
Use this decision table before changing URLs. The goal is a clearer information system, not an artificial promise of extra listings.
| Situation | Likely decision | Implementation check |
|---|---|---|
| Same task, same audience, near-identical answer | Consolidate. | Choose the strongest URL, merge unique value, redirect retired URLs and update internal links. |
| Same topic, different task or format | Differentiate. | State the role clearly and link between the guide, tool, comparison, product or support page. |
| Location pages with genuinely different operations | Keep when useful. | Provide real local details, availability, proof and ownership—not city-name substitutions. |
| Language or regional alternates | Keep valid alternates. | Use self-canonical URLs in the same language plus reciprocal hreflang where appropriate. |
| Parameters, print views or tracking variants | Canonicalize or control generation. | Keep one indexable preferred URL and avoid conflicting signals. |
| Subdomain created only to gain another result | Reconsider architecture. | Choose root, subfolder or subdomain for users and operations, not diversity evasion. |
| Two pages alternate for the same queries | Investigate intent and overlap. | Compare query clusters, links, canonicals, content roles and conversions before merging. |
| Brand query shows several useful pages | Often leave intact. | Ensure the home, contact, login, product and support paths are accurate and easy to navigate. |
What this does not mean
- Google did not announce an absolute two-results-per-domain rule.
- Google did not publish a Site Diversity score, report, manual action, recovery request or opt-out setting.
- The original announcement did not disclose the private exception logic or a permanent list of affected result features.
- A root domain and subdomain are generally grouped, but Google can treat them separately when relevant.
- Canonicalization, deduplication, ranking, sitelinks and site diversity are related result-handling concepts, not interchangeable names.
- Search Console usually attributes data to Google’s canonical URL and omits some queries, so visible rows are not a complete count of every listing.
- Average position is aggregated and does not show every observed ranking configuration.
- A decline during 4–6 June 2019 could overlap the separate June Core Update or unrelated site changes.
- Creating subdomains, doorway pages or near-duplicate location pages is not a responsible workaround.
- No external rank tracker can prove which internal Google system removed a particular URL from one SERP.
Frequently asked questions
Does Google allow only two results from one website?
No hard cap was announced. Google said users would usually see no more than two top listings from one site, but more can appear when especially relevant.
Is the Site Diversity Change a penalty?
No. Google described it as a change to result composition, separate from a penalty or manual action.
Does the rule include subdomains?
Generally yes. Google said root domains and subdomains would normally count as one site, while allowing relevance-based exceptions.
Can a brand query show more than two results?
Yes. Navigational or highly specific queries may justify several useful results from the same source.
Are sitelinks counted as separate results?
Sitelinks are automated shortcuts clustered beneath a text result. Treat them as a different search presentation rather than proof of several independent rankings.
Is this the same as keyword cannibalization?
No. Cannibalization is a practical label for overlapping pages competing or substituting for one another; diversity controls how varied the displayed sources can be.
Should I merge every page that targets a related keyword?
No. Merge pages that serve essentially the same task. Keep pages that provide distinct intent, audience, format or operational value.
Will moving a page to a subdomain create another listing?
There is no guarantee, and Google says subdomains are generally grouped with the root for diversity. Choose architecture for users and maintenance.
How do I prove Site Diversity caused a loss?
You usually cannot prove it from one site’s data. You can build a reasonable hypothesis only after checking query scope, canonical URLs, indexing, SERP composition, competitors and release history.
What should I measure instead of the number of listings?
Measure topic-level impressions and clicks, the performance of the intended canonical landing page, qualified conversions and whether users can reach the next useful page.
Official sources and primary references
- Google Search Liaison announcement archived by Search Engine Journal
- Google Search: What is canonicalization?
- Google Search: How to specify a canonical URL
- Google Search ranking systems guide: Deduplication systems
- Google Search: Sitelinks
- Google Search: Debugging drops in Search traffic
- Search Console: Dimensions and data groupings
- Search Console: Impressions, position and clicks
- Google Search crawling and indexing FAQ: Subfolders or subdomains
Continue the Google Search systems series
Review visibility, content, links and technical health without chasing an invented algorithm score.
How Google Ranking Evolved: From PageRank to Modern Search SystemsSeptember 2, 2026
Google AI Content and SEO: What Is Allowed, What Is Spam, and How to Publish SafelySeptember 2, 2026
Google Florida Update (2003): What We Know, What Remains TheorySeptember 2, 2026