What did the stale website scan actually measure? (And what it didn't)

The short answer

We measured websites, not AI answers. We scanned 32,574 B2B domains. Of the 20,496 homepages that showed a visible copyright year, 5,022 still said 2025 or earlier. That tells you how many companies let the most authoritative page they own go quiet. It does not tell you what ChatGPT says about any of them. Those are separate questions, and they'd need a separate study. PitchKitchen founder Greg Rosner ran the scan behind The Truth About Your Website; Demand Gen Report published a Q&A on it on October 12, 2026.

Why a copyright year, of all things?

Because it's the one line on a homepage that a human has to touch. A footer that says 2025 in October 2026 is a small fact with a big tell behind it: nobody has owned the page in a year. Changing the date takes a minute. Understanding what's happening behind it takes a conversation. In our client work that conversation usually reveals a business that has moved a long way and a website that hasn't come with it.

What the scan does and doesn't prove

Does: one in four dated B2B homepages hasn't been touched this year. Firms with 11 to 50 employees were stale more often (27.5%) than firms with 51 to 200 (22.2%). WordPress carried about the same share of stale sites as fresh ones (53.1% vs 55.8%), so the platform isn't the cause.

Doesn't: prove an AI engine repeated an outdated story, or skipped the company. We didn't ask the engines. When Demand Gen Report asked me how long a company has before the engines lock in the old story, my answer was that we haven't established a countdown, and anyone who tells you there is one is selling the countdown.

Why it matters anyway

Your buyer's first move is a conversation, not a search. They open ChatGPT or Claude, describe their problem in plain language, and ask who can help. Your website is what the assistant has to work with. If it explains the business you ran two years ago, that's the business the buyer hears about.

One customer invested $3 million in an intelligent compliance system. The website hadn't changed in a year. The capability was nowhere. Three million dollars of progress, and the buyer couldn't see it. That's the gap the footer points at.

What a CEO should do with this

Two jobs. Get the story clear: who it's for, what changed, why it matters, why now. Then modernize how you publish it, so an authorized person can get an accurate page live without a week of emails. Start with either. Both have to get done. Don't go shopping for a new CMS first; new platform, same vague story, same waiting.

FAQ

Is a stale copyright year bad for SEO or AI search?

By itself, no. It's a symptom. The page behind it is what engines and buyers read. Fix the page, and the date fixes itself.

Our footer updates automatically. Are we fine?

Maybe not. A footer that updates every January while the story above it hasn't moved in years is the worse case, because nothing flags it.

Did you name the companies?

No. No company is named anywhere in the data, and we don't release the domain list. Editors can get the counts by vertical and platform on request.

Where was this published?

Demand Gen Report ran a Q&A with Greg Rosner on the findings on October 12, 2026: PitchKitchen's Greg Rosner on AI Search and Your Website. The full report is at pitchkitchen.com/truth.

Check your own homepage

Run the Brand Signal Score on your homepage. It reads the page the way a buyer and an AI assistant do and tells you in under a minute whether your story is clear, current and worth repeating.