Are Your Old Blogposts Burying You?

The issue isn’t simply that old content is, um, old. It’s that years of publishing can leave a company with hundreds—or thousands—of pages describing its products, categories and expertise in subtly different ways.

Contentful and Benchmarker’s 2026 research surveyed 321 B2B SaaS content, brand and demand leaders and found a median publishing rate of 11–20 blog posts per quarter. Contentful’s report page similarly characterizes the norm as roughly 10–20 posts quarterly.

Content debt used to mean a few stale landing pages and an embarrassing 2018 thought-leadership post confidently predicting the future of work. Now it means an AI system trying to reconcile a decade of shifting product names, competing category definitions, and vocabulary that changed with every reorg, and it doesn’t know which version to believe.

Google says its generative search experiences use techniques including retrieval-augmented generation and “query fan-out,” in which multiple related searches retrieve relevant material that can support an AI-generated response. Its current guidance for generative AI search emphasizes useful, original content, clear organization and solid technical SEO rather than special AEO/GEO tricks.

That makes an overlooked question increasingly important:

What story does your entire content corpus tell?

The 1,500-post problem

Consider the scale. Maybe take a Xanax first, though.

Combine that publishing cadence noted above with typical company age and content maturity, and a reasonable planning estimate looks something like this:

Company size	Typical profile	Estimated blog corpus	Working midpoint
Startup	<100 employees; ~2–5 years publishing	50–250 posts	~150
Mid-sized	100–1,000 employees; ~5–10 years	250–750 posts	~500
Enterprise	1,000+ employees; often 10+ years	750–3,000+ posts	~1,500

Source: Estimates based in part on publishing-volume benchmarks from Contentful/Benchmarker’s 2026 B2B SaaS research. These are directional corpus estimates, not reported industry medians.

For an enterprise marketer, then, the AI-search challenge may not be optimizing the next 20 articles.

It may be reconciling the previous 1,500.

When your website disagrees with itself

Picture a software company that spent a decade publishing. In 2017 it described its category as marketing automation. In 2020 the language shifted to customer experience. In 2023 it became revenue orchestration. Today the company wants AI systems to describe it as an agentic customer engagement platform.

A human reader clocks that as a rebrand. A crawler just sees four different companies who happen to share a domain. Unless old pages get updated, consolidated, or retired, all four versions stay live and readable. Product names change, definitions evolve, statistics expire, acquisitions bolt on a fifth vocabulary, and every writer who ever touched the blog described the same capability slightly differently. Nobody planned this. It just accumulates, the way a garage does.

Google says its generative search systems can understand synonyms and meaning rather than requiring identical wording sitewide, and it explicitly cautions against simplistic AEO/GEO tricks. So the risk isn’t “old page, lower AI ranking.” It’s that a sprawling legacy corpus can hand an AI system weaker, contradictory, or outright expired evidence about what a company actually is, via retrieval-augmented generation pulling from whichever page happens to match the query.

AEO may start with archaeology

The instinctive response to AI search has been creation: publish AI-friendly FAQs, add comparison pages, restructure articles around conversational queries. All fine tactics. They also only address the front door, while the attic stays exactly as cluttered.

The bigger opportunity is a corpus audit. Start with the concepts your company most wants to own, then check how consistently those concepts are defined across years of material. Find outdated product descriptions, competing category definitions, contradictory statistics, obsolete positioning, clusters of near-duplicate posts. Decide what gets refreshed, consolidated, redirected, or quietly buried.

Is it a lot of work? Yes. Is it necessary? Even more yes.

This is what a Content Effectiveness Audit does at the portfolio level, and it feeds directly into the AI Visibility Audit: a company’s own back catalog is one of the strongest signals an engine has for entity definition, and a self-contradicting one tanks that score before a single new post ships.

It’s also more measurable now. Google rolled out generative AI performance reporting in Search Console worldwide in August 2026, giving site owners direct visibility into impressions from AI Overviews and AI Mode.

The next great GEO article you publish is walking into a conversation your company might have been having with the internet for ten years, whether it meant to or not.

Before publishing 20 more posts explaining who you are, it might be worth finding out what the previous 1,500 already told everyone.