The Undated Web: A Third of "2026" Marketing Articles Carry No Machine-Readable Date

By Nadia Sharpe, AI Content & Organic. Last updated: 2026-08-16

The plan for this audit was a story about deception. The data told a story about omission. We are publishing the second one, because it is the one that is true.

Do publishers relabel old articles with the new year?

Not in this corpus, or not detectably. We harvested the sitemaps of the 54 martech vendors from our other research audits plus 20 independent marketing publications, sampled URLs with a year in the slug, fetched them, and kept the ones whose <title> actually contains "2026." That left 142 articles across 48 sites. For each we read datePublished and dateModified from the page's JSON-LD, falling back to Open Graph article meta tags.

Nine articles wore a year later than their dates. We read all nine. Six were published in December 2025 with titles like "2026 Marketing Trends Every CMO Should Be Watching" (Iterable) and "5 B2B Marketing Predictions for 2026" (B2B Marketing), which is exactly when a predictions piece should be published. Three were news items whose headline happened to contain the year, "YouTube Will Livestream Coachella Through 2026" (Adweek, 2023), and are false hits on our filter. The relabelling story evaporated on contact with the pages.

How many year-titled articles are undated to a machine?

Forty-five of 142, or 32%. These pages tell a human reader "2026" in the headline and tell a crawler nothing about when they were published or last touched: no datePublished, no dateModified, no article:published_time.

That is a different failure from relabelling, and in one way a worse one. A relabelled article at least commits to a date that can be checked. An undated one cannot be checked, cannot be sorted by freshness, and cannot tell a retrieval system whether "2026" means "written this year" or "written whenever." Search engines fall back to guessing from page content, and answer engines increasingly weight recency signals they can actually parse. Our schema audit of the same publishers found dateModified missing on 39% of newest posts; this cut shows the omission is not confined to old content.

Which sites publish undated year-titled articles?

Nineteen of the 48 sites in the sample. Four sites had four undated articles out of four sampled: Pipedrive, Outreach, Hootsuite and Buffer. HubSpot, Instantly, Cognism and Qualified had three each. Braze, Clay, Apollo, Livestorm and The Drum had two. Copper, Zoho CRM, Salesloft, Seamless.ai, Jasper and Demio had one.

The Drum is the only independent publication on the list. Independent titles were undated on 3 of 24 year-titled articles (13%); vendor blogs on 42 of 118 (36%). Publishers, whose platforms are built for dated content, encode dates. Vendors, whose blogs often run on marketing CMSs, frequently do not.

Why does a machine-readable date matter more than the year in the title?

Because the year in the title is a promise and the date in the markup is the receipt. A reader takes "Best CRM Software for 2026" at face value. A system deciding whether to surface that page for a query asked in August 2026 wants to know whether it was updated in 2026 or written in 2024 with the number changed, and only the markup answers.

There is a second-order effect for the vendors on the list. A page with no date is also a page whose freshness cannot be proven when it is fresh. Hootsuite and Buffer both update their social-statistics posts regularly; the undated versions we fetched get no credit for it from any parser. Encoding datePublished and dateModified in the Article schema is the same five-line change our schema audit keeps recommending, and it is one of the freshness signals our guide to showing up in Perplexity and AI search puts near the top, and it is the difference between a freshness claim and a freshness signal.

What did the audit count, and what did it not?

Sitemap URLs with a year in the slug, filtered to titles containing the year, up to four per site for 2026 and two for 2025, fetched once on 2026-08-16. Dates from JSON-LD first, Open Graph meta second. "Undated" means neither source carried a parseable published or modified date; a date rendered only as visible text on the page does not count, because the study is about what machines can read. Every article that failed the "date before title year" test was read by hand, and the classification of all nine is in the data file.

The 2025-titled sample behaves the same way: 66 articles, 37 consistent, 6 older (all December-2024 predictions or older news), 23 undated (35%).

What are the limits of this audit?

The sample is a slice, not a census: four sites' sitemaps returned nothing (Lusha, Clearscope, MarketingProfs, Animalz) and the per-site cap of four means Salesforce's hundreds of year-titled pages are represented by four. Machine-readable dates can exist in formats we did not parse (a time element with datetime, for instance), so "undated" is a floor on absence, not a ceiling. And a site could add schema tomorrow. The finding is about the state of the pages on the day.

FAQ

How many marketing articles with 2026 in the title have no publish date?

45 of the 142 we checked (31.7%) carry no machine-readable published or modified date in JSON-LD or Open Graph meta tags.

Do marketing blogs relabel old articles with the new year?

Not detectably in this sample. Of 9 articles whose on-page date preceded their title year, 6 were legitimate December-2025 predictions pieces and 3 were news items mentioning "2026" in a different sense.

Which sites publish undated year-titled articles most often?

Pipedrive, Outreach, Hootsuite and Buffer each had four undated articles in a sample of four; HubSpot, Instantly, Cognism and Qualified had three each.

Are vendor blogs or independent publications more likely to omit dates?

Vendor blogs: 42 of 118 vendor articles (36%) were undated versus 3 of 24 independent (13%).

Why does an undated article matter if the year is in the title?

The title is a claim; the markup is what search and answer engines read to verify it. An undated page cannot prove it is fresh even when it is, and cannot be sorted or weighted by recency.

Sources

• Google Search Central: Provide a publish date
• Schema.org: dateModified
• Open Graph protocol: article object type

GTMStack publishes original audits of the AI marketing stack. If you want the next one when it lands, the newsletter is the place to get it.