One of the interesting parts of the API leak was the discussion of the FreshnessTwiddler. We’ve known for a long time that Google prefers fresh content. There are a lot of different types of dates Google can use to determine freshness.
From the API leak, we know of three types of dates Google pays attention to:
- The date specified in the page’s text. This was referred to as the bylineDate.
- The date specified in a URL or page title. This was referred to as the syntacticDate.
- The API also discussed semanticDate, where Google derives a date from the page’s content.
However, there are other types of dates we’ve long known Google uses:
- Article schema allows for providing datePublished and dateModified. datePublished is the date the article was first available on the website. dateModified is the date the page was most recently updated. Typically, dateModified is only updated with significant updates (for example, adding a new image to the page vs. correcting a small typo).
- There are meta tags for article:published_time and article:modified_time. Google’s documents do not say they use these tags. However, Facebook does. Here again, article:published_time is when the article was first available on the website and article:modified_time is for the last update. As with article schema, article:modified_time typically means the last significant update.
- You can also specify a date in the lastmod element in the XML sitemap. Google will only use this tag if they can verify it against the date provided on the page.
- A website’s RSS feed also provides a date. Each item has a pubDate to indicate when the item was published. It is not clear if Google uses this date. However, given how often I see Google crawling RSS feeds in log files, it is possible they are. The RSS specifications do not clarify if pubDate is the publication date or the last modification date. Typically, I’ve seen people use this as the original publish date, not the modification date.
Whatever date fields you use, you want to make sure the dates are communicated consistently.
Recap
| Date Type | Definition |
| datePublished (Article Schema) | The date the article was first available on the website. |
| dateModified (Article Schema) | The date the page was most recently updated, typically for significant updates. |
| article:published_time (Meta, Open Graph) | The date the article was first available on the website. |
| article:modified_time (Meta, Open Graph) | The date the article was last updated, typically for significant updates. |
| lastmod (XML sitemap) | The date the page was last modified. Verified against the date on the page. |
| pubDate (RSS feed) | The publication date of the item, often used as the original publish date. |
| bylineDate | The date specified in the page’s text. |
| syntacticDate | The date specified in a URL or page title. |
| semanticDate | A date derived by Google from the page’s content. |