Pillar Pages In 2026: When To Build, When To Retire, And When To Split

A content team built 14 pillar pages between 2019 and 2022. The pages anchored the site's topical authority and drove substantial organic traffic for years. By mid-2026, three of the original 14 still drive meaningful traffic. Five drive almost no traffic but still rank for their target keywords. Six have lost rankings and traffic entirely. The team's instinct is to refresh all 14. The right answer is more nuanced: refresh some, retire others, split a couple, and consider building new pillars for topics that have emerged since 2022.

This pattern is common in 2026 for content teams with substantial historical content investments. The pillar-and-spoke pattern still works fundamentally, but the specific pillars that worked in 2019 may not be the right structure for 2026. The topic landscape has shifted. Some topics have fragmented. Others have consolidated. AI engines have introduced new retrieval patterns that reward different content structures.

This piece unpacks the decision framework for when to build, retire, and split pillar pages in 2026. The work is strategic rather than tactical; the choices made here shape the site's topical authority for years.

The Pillar Page Pattern Still Works, But Needs Maintenance

The pillar-and-spoke pattern emerged in the late 2010s as a response to Google's increasingly topic-based ranking signals. A comprehensive pillar page anchors a topic; spoke pages cover specific subtopics; internal links connect pillar to spokes establishing the topical cluster.

The pattern remains effective in 2026 for several reasons. Topic-based ranking signals have strengthened further with BERT, MUM, and AI engine semantic retrieval. Content clusters provide explicit topical authority signal that flat content sites lack. Pillar pages serve as natural reference content that earns backlinks and citations.

The maintenance requirement has also increased. Topics evolve faster than they did in 2019. New subtopics emerge; old ones fade. The pillar page that comprehensively covered a topic in 2020 may have outdated framings, missing recent developments, or coverage gaps by 2024 and beyond. Without maintenance, pillars decay.

The brands that succeed with pillar pages in 2026 treat them as living documents requiring quarterly or biannual refresh. The brands that built pillars and walked away see the decay pattern: ranking drift, traffic decline, citation rate decline.

The work involves not just updating individual pillars but evaluating whether the pillar structure still matches the current topic landscape. New pillars may be needed for topics that did not exist three years ago. Old pillars may need retirement because the topics they covered have changed substantially.

Building topical authority is the broader discipline; pillar maintenance is one operational layer.

When To Build A New Pillar Page

New pillar pages make sense under specific conditions.

The topic has emerged as a distinct subject area. Topics like generative engine optimization, AI search, or query fan-out did not exist as established content areas in 2020. Sites covering these topics now benefit from pillar pages that anchor the topic on their domain.

The topic has sufficient depth for substantive coverage. A pillar page needs 3,000 to 5,000 words of substantive content with multiple subtopics that warrant spoke pages. Topics that fit in 1,500 words do not warrant pillar treatment; they fit in regular long-form posts.

The site has expertise or content commitment to support the pillar. Building a pillar without the underlying content commitment leads to stale pillars within a year. The site should have spokes already published or planned that cover the cluster's subtopics.

The topic has business or audience relevance. Pillars on topics outside the site's core focus typically do not produce ROI commensurate with the effort. The pillar should serve a specific buyer audience or business goal.

There is an underserved opportunity. If the topic is already dominated by exhaustive coverage from established sources (Wikipedia, major industry publications), entering with a new pillar requires differentiation. The new pillar should have a clear positioning beyond what existing coverage provides.

For brands considering new pillars, the audit question is which topics meet all five conditions. Each potential pillar should be evaluated against the criteria before committing to the 6 to 12 month investment a strong pillar requires.

For most established sites, the number of high-priority new pillar opportunities in any given year is small (typically 2 to 4). The rest of the content effort should go to expanding existing pillars with new spokes, refreshing existing pillars, or producing tactical content that does not warrant pillar treatment.

When To Retire An Existing Pillar Page

Retiring pillar pages is harder emotionally than building new ones. The investment in the original pillar feels lost if retired.

The criteria for retirement include:

Traffic has dropped below the pillar's value threshold. A pillar that earned 5,000 monthly visits but now earns 200 has become a maintenance burden without proportional return. The threshold depends on the site's overall traffic baseline; the principle is that the pillar should still be earning meaningful traffic.

The topic has become commoditized or obsolete. A pillar on "what is COVID-19" was high-value in 2020 but has limited current relevance. Pillars on technology categories that have been superseded or terminology that has changed face similar retirement pressure.

Maintenance burden exceeds value. Some pillars require frequent updates because the underlying topic changes rapidly. If the maintenance cost outweighs the traffic value, the pillar should be retired or scoped down.

The site's strategic focus has shifted away from the topic. Brands occasionally pivot or refocus, leaving prior topic clusters off-strategy. Retiring those pillars frees resources for current strategy.

The pillar is structurally weak and beyond rehabilitation. Some pillars never quite worked: weak content, poor internal linking, mismatched architecture. Rather than trying to fix what fundamentally does not work, retirement is sometimes the right call.

Retirement does not have to mean deletion. Several approaches preserve value: redirect the URL to a successor pillar that covers the topic better, repurpose the content into a spoke under a different pillar, archive the page with a noindex tag, or convert the page into a historical reference with a note about its dated status.

The retirement decision should be deliberate. Pages occasionally still serve niche traffic worth preserving; mass retirement can hurt the site's content footprint. The right answer is selective retirement of pillars meeting multiple criteria.

When To Split A Sprawling Pillar Page

Some pillars grow over time into sprawling pages covering what have become multiple distinct topics. Splitting is sometimes the right answer.

The criteria for splitting include:

The pillar exceeds 8,000 to 10,000 words and the subtopic sections could each support their own pillar. Sprawl beyond this length usually indicates the topic has fragmented into multiple distinct subjects.

User search intent for different parts of the pillar diverges. If users searching for one section of the pillar have fundamentally different intent than users searching for another section, the unified pillar serves neither well.

The internal section ranking signals diverge. Some sections of the pillar drive substantial traffic; others underperform despite covering legitimate topics. The pattern often suggests the underperforming sections would do better as standalone pillars.

The topic has fragmented into recognized distinct subject areas. Industry vocabulary shifts; what was once one topic becomes two. A pillar covering both falls between the new positionings.

The pillar covers loosely related rather than tightly clustered subtopics. Pillars that grew organically by accumulating loosely related sections may have become hub pages without strong topical coherence.

The split is a substantial editorial undertaking. Each new pillar inherits some content from the original but requires substantive rewriting to function as a standalone pillar. The internal linking structure has to be updated to point to the new pillars. Redirects may be needed for the original URLs.

The benefits of successful splits include: each new pillar can rank for its specific topic without conflict, internal linking becomes cleaner with each pillar having focused spokes, maintenance becomes more manageable with smaller focused pillars, and AI engine retrieval improves as each pillar provides clearer topical signal.

For brands considering splits, the audit involves identifying pillars that have grown unwieldy and mapping their internal structure. Pillars with clearly distinguishable subtopic sections are candidates; pillars with tightly integrated content covering one core topic are not.

The Spoke Architecture Each Pillar Needs

A pillar page without spokes is just a long article. The cluster structure requires the spoke pages that link to and from the pillar.

The spoke architecture for a healthy pillar typically includes:

10 to 30 substantive spoke pages covering specific subtopics within the pillar's scope. Fewer than 10 spokes produces a weak cluster; more than 30 starts to dilute focus.

Each spoke is 1,500 to 3,000 words on its specific subtopic. Spokes shorter than this lack substance; spokes longer typically should be split into their own subtopic clusters.

Spoke pages link to the pillar with descriptive anchor text. The internal linking establishes the pillar as the canonical reference for the topic.

The pillar links to relevant spokes from within the pillar content. The bidirectional linking strengthens the cluster signal.

Spokes link to other spokes where topical connections exist. The internal interlinking among spokes builds the cluster's structural integrity.

External backlinks point to both the pillar and individual spokes. The link profile distribution shows the cluster is actively referenced, not just the pillar.

For pillars without strong spoke architecture, the path forward is building spokes systematically. A pillar with five spokes can grow to fifteen or twenty over a year with consistent editorial calendar focus.

Internal linking strategy discusses the broader linking work; the pillar-and-spoke pattern is one specific application.

Measurement Criteria For Pillar Page Health

Pillar page health is measurable across several dimensions.

• Organic traffic - The most basic metric. A healthy pillar earns substantial monthly visits, typically 1,000+ for most B2B contexts and higher for consumer topics. The trend matters more than the absolute number; declining traffic is the warning signal.
• Ranking position on target keywords - The pillar should rank in the top 5 for its primary keyword and the top 10 for several secondary keywords. Drops out of the top 10 indicate weakening signal.
• AI citation rate - AI engines should cite the pillar for relevant category queries. The citation rate over time tracks whether AI engines treat the pillar as authoritative.
• Backlink growth - Healthy pillars accumulate backlinks gradually. Stalled or declining backlink growth indicates the pillar is no longer being referenced by other sites.
• Internal link density from spokes - The number of spoke pages linking to the pillar should grow over time as the cluster expands. Stagnant internal linking indicates an inactive cluster.
• Time on page and engagement - Substantive pillars produce longer time-on-page than thin pages. Engagement decline indicates content quality issues or user intent mismatch.
• Conversion attribution - For brands with conversion tracking, the pillar's role in user journeys (assisted conversions, last-click conversions, multi-touch attribution) measures its business value beyond just traffic.

The dashboard for pillar health should track these metrics for each pillar on a quarterly basis. Patterns across pillars surface the strategic decisions (which pillars are healthy, which need refresh, which need retirement or splitting).

Six Pillar Page Mistakes That Cost Traffic

Six recurring mistakes consistently produce pillar page decline.

1. Building once and walking away. Pillars require ongoing maintenance. Sites that built pillars in 2020 and never refreshed them see decline as topics evolve.
2. Thin spoke architecture. A pillar without 10+ substantive spoke pages is incomplete. The cluster signal requires the spokes.
3. Inconsistent internal linking. Pillars that should be linked from spokes but are not lose the cluster authority signal. Audit and fix the internal linking systematically.
4. Stale content with no update flag. Pillars showing 2020 publish dates and 2020 information signal abandonment even when the underlying content is partially valid. Update and re-date.
5. Over-broad pillar scope. Pillars covering loosely related subtopics produce thin signal across the breadth. Split or focus.
6. No measurement framework. Pillars without metrics tracking decay invisibly. Build the dashboard early.

Frequently Asked Questions

How long should a pillar page be?

3,000 to 5,000 words for most topics. Substantively long enough to comprehensively address the topic without being unwieldy. Pillars exceeding 8,000 to 10,000 words often should be split. Pillars below 2,000 words are typically too thin to function as pillars.

How often should pillars be refreshed?

Quarterly for fast-moving topics (AI, technology, ranking algorithms); biannually for moderate-velocity topics (marketing tactics, business strategies); annually for slower-moving topics (foundational concepts, historical analysis). The refresh should add new sections, update outdated references, refresh statistics and examples, and review internal linking.

Should I publish pillar pages all at once or gradually?

Gradually. A pillar without spokes underperforms; building the entire cluster (pillar plus 15 to 25 spokes) over 3 to 6 months produces stronger overall outcomes than publishing pillar and spokes in the same week.

Can a pillar page rank for multiple primary keywords?

Yes, with skill. The strongest pillars rank for the head term plus several related variations. The structure that supports this is comprehensive coverage with sections that each target specific subtopic queries.

How do I handle pillar redirects when retiring or splitting?

301 redirect the retired pillar to its replacement or to the most-related successor page. Avoid redirecting many pages to a single page; some redirects should go to specific spokes that cover specific aspects. The redirect map should preserve user intent.

Are pillar pages necessary for ecommerce sites?

Different application. Ecommerce uses category and collection pages as "pillars" with product pages as "spokes." The pattern is similar but the implementation differs. For ecommerce content marketing alongside product pages, the traditional pillar-and-spoke pattern still applies to the blog or resources section.

Pillar pages remain effective in 2026 with active maintenance and strategic decisions about when to build, retire, and split. The work is more nuanced than the original pillar-and-spoke framework suggested; topics evolve and the structure has to evolve with them.

The brands that maintain pillar architectures well earn compounding topical authority over years. The brands that built pillars and walked away see decline as the topics shift around the stale content.

If your team wants help auditing your pillar page portfolio and making the build/retire/split decisions, that work sits inside our generative engine optimization program. The pillars that anchor topical authority in 2026 are the pillars whose owners have continued to invest in them as the topics matured.