Content and evidence
Refresh the facts before changing the date
Review changing claims, sources, examples and dates with a completed refresh record that distinguishes substantive updates from cosmetic edits.
TL;DR
- Refresh a page when the answer, evidence or buyer decision has changed; a newer year in the headline is not the work.
- Recheck volatile facts such as availability, pricing, versions, research definitions and comparisons.
- Keep a change record, then make the summary, body, tables and date metadata agree.
- A useful historical result can remain on the page if its period and limits stay clear.
A content refresh should leave a reader with a more accurate answer. Changing “2025” to “2026” without checking the underlying claims can make the page more misleading: the presentation promises current guidance while the facts remain old.
Start with the reader's task. A plan change can make a product comparison wrong. A how-to guide may still help years later if its method and limits still hold. Age alone does not tell you which case you have.
Separate currentness from a fresh-looking label
Google's people-first content guidance explicitly questions changing dates without substantial content changes. Microsoft's account of Bing ranking also treats freshness as keeping information current, while recognizing that older content can remain relevant.
These are platform explanations, not controlled tests of how much traffic an update will add. They support a useful editorial distinction: change the facts that affect the answer, then describe the update truthfully.
A review need not lead to a rewrite. Keep a claim that is still sound. Replace a fact that has changed. Add context to an old result that still helps. Remove a claim if you cannot find its source.
Worked example: a software-selection guide
This is a fictional guide for a logistics team choosing a CRM. It says a product includes a named integration on its entry plan, quotes an old price and cites a survey of enterprise buyers as though it describes small firms.
Read the current vendor plan page, setup guide, source survey and published article. In this fictional review, the plan sheet puts the integration on Professional, not Starter. The old price lacks its currency and billing term, and no current quote is available. The survey never covered the small firms that the article claims to help.
| Existing claim | Review finding | Revised copy |
|---|---|---|
| The entry plan includes the dispatch integration. | The current plan sheet places it on Professional. | Choose Professional if the dispatch integration is required. Starter does not include it. |
| The plan costs $20. | No current quote confirms currency, billing term or seat basis. | Request a current quote showing currency, billing term and price per seat. We have removed the old price. |
| Most small teams prefer this feature. | The cited survey sampled enterprise buyers. | The survey describes enterprise buyers; it does not establish which features small teams prefer. |
| Migration takes two weeks. | No delivery evidence supports a fixed duration. | Timing depends on data quality and mapping needs. Share a sample export before requesting an estimate. |
| Updated this year. | The page has no record of what changed. | Update note: corrected plan availability, removed the old price and fixed the survey's scope. |
The revised answer may point the team to a different plan. That changes the buying advice. A new date alone would have left the wrong answer in place.
The subject expert checks the facts. The editor checks each changed fact in the TL;DR, tables, FAQ and linked pages. A reader then uses the guide for the same task. The advice must follow from the facts that are now on the page.
Preserve dates that explain the evidence
Do not silently change a study's year to match the article's update date. The research period describes when the data was collected. The publication date describes when the article first appeared. The last-updated date describes a later change to that article. These fields have different jobs.
Google's date guidance recommends a visible, clearly labeled date and appropriate datePublished or dateModified structured data. Keep those values consistent with the actual work. A template deployment or a corrected comma does not automatically justify marketing the whole article as a new analysis.
For substantial revisions, a short note can say what changed: “Updated the plan comparison and integration requirements; retained the historical survey with its original sample.” It helps a returning reader find the difference without reading an internal edit log.
Schedule reviews around what can go stale
Give each key claim a source and an owner. Recheck prices and stock when vendors announce changes. Stable terms may need fewer checks. A review schedule helps, but it cannot keep a claim true between reviews.
Semanta uses evidence review to connect website claims with the materials that support them. For service pages, this directly supports the Evidence dimension of our audit framework. Review the claims that affect fit or commitment first.
Common questions
Should an old article be deleted?
Only when that is the right outcome for its purpose and value. A supported historical account may deserve an explicit historical label. A duplicate or superseded guide may need consolidation and a redirect.
What if a newer study has a different sample?
Explain the change in population or measurement. Do not present the two values as a trend unless they are comparable.
When product, sales and marketing teams disagree about what is current, bring a priority page and its source materials to discuss a Semanta evidence review.