How we do it
At Cloudflare, we track the following information about each page:
Of all of these values, there is a bit of nuance to our Last reviewed metadata. Last reviewed differs from Last modified because a review is more thorough than an update. A review implies that all contents of the page have been vetted for accuracy.
Because of this extra effort, we only track Last reviewed for content types that are particularly important to the user journey and require an additional level of maintenance. At the moment, those content types are and specific types of .
There are two aspects to our metadata implementation: how we set these values and how we make them available.
In setting values, we automate what we can and manually set the rest:
|Product||Set automatically by using the subfolder.|
|Product Group||Defined at the product level and then automatically applied to each page in the subfolder.|
|Content type||Manually set for each page.|
|Last modified||Automatically pulled from the Git history.|
|Last reviewed||Manually set for some pages.|
|Word count||Automatically generated during build time.|
Making values available
We choose to render all of these values as specific
meta properties for each page.
This decision - combined with the multiple web scrapers that crawl our pages - means that our metadata is available for reporting and search improvements (ranking and faceting).
It also means that our GitHub repo is always the source of truth and we do not have to keep a spreadsheet or mapping updated elsewhere.