Amidst the ongoing chaos of many users’ Google Discover feed, there has been another tweak to the official documentation. Earlier this month, Google updated the documentation for Discover to include a paragraph about types of content that might not appear in feeds. These were listed as job applications, petitions, forms, code repositories, or satirical content “without any context”.
Additional changes were made on 18 April (according to the last updated time stamp), with one of the most intriguing of the lot being a sentence that reads:
“Our ongoing work to improve Discover’s user experience means sites may see changes in their traffic unrelated to the quality or publishing frequency of their content.”.
Because Discover is a bit of a black hole – with publishers not having much control over content that surfaces in the feeds, we have been running feeds on a number of test devices to assess what might influence the appearance of content – long before the current content catastrophe began to unravel.
Since the start of March, several publishers – including some big players in the tech space – have reported dramatic drops in their Discover traffic, which makes the reference to “changes in traffic unrelated to the quality” intriguing. There are some suggestions that these dramatic drops to Google Discover relate to recent Android or Chrome updates and are errors rather than algorithmic, however, beyond the update to the documentation, Google has not publicly commented or acknowledged any known bugs.
What makes the addition of that specific sentence even more interesting is that the first port of call from the copy and paste answers publishers receive from so-called “Product Experts” (non-Google employees who answer questions voluntarily) on official publisher support forums is… erm…well, a link to the Discover documentation.
The full changes to those paragraphs are indicated below. The original paragraph read:
Given the serendipitous nature of Discover, traffic from Discover is less predictable or dependable when compared to Search, and should be considered supplemental to your Search traffic. This means that you might create and optimize content to fulfil specific search needs for search engine traffic, but there is no way to create content that explicitly targets Discover’s interest matching.
And that has now been updated to the following:
We are always improving Discover to serve personalized, helpful content to users. Given the serendipitous nature of Discover, traffic from Discover is less predictable or dependable when compared to Search, and should be considered supplemental to your Search traffic. Our ongoing work to improve Discover’s user experience means sites may see changes in their traffic unrelated to the quality or publishing frequency of their content.
The official Google Discover documentation also removed a reference to there being no way to “explicitly” optimise content for Discover. The line that was removed read: “You might create and optimize content to fulfil specific search needs for search engine traffic, but there is no way to create content that explicitly targets Discover’s interest matching.”
Also notable is moving a paragraph making specific reference to EAT (as in: authoritativeness, and trustworthiness in the context of search) higher up in the documentation to appear above the bullet-point guidance on content appearing in Discover.
Most curious of all is that a paragraph clearly referring to bylines and publisher details being required has been removed entirely and there is now no direct reference to the requirement for bylines at all.
The paragraph that has been removed read as follows:
Providing clear dates, bylines, information about authors, the publication, the publisher, company or network behind it, and contact information to better build trust and transparency with visitors.
While the official Google documentation on EAT does reference authors, the removal of this paragraph is especially interesting considering the extensive appearance of articles without any bylines at all in some of our Discover test feeds over the last few weeks. The removal of this paragraph seems to somewhat contradict the EAT policies.
There have been several reports by users – both on Google’s official support forums as well as social media platforms – about issues with their feeds, too.
The issues range from completely empty feeds to irrelevant content being suggested. Under that context, the addition of a paragraph that references Discover being a “highly personalised feed” that is “automatically refreshed” makes for interesting reading.
The full paragraph reads:
As a highly personalized feed, Discover actively tunes itself to a user’s interests and displays content that aligns with those interests. The content in Discover is automatically refreshed as new content is published, however, Discover is designed to show all types of helpful content from across the web, not just newly published content.
In some of our Discover Feed tests, we have recorded several instances where content would display despite the user having explicitly added it to the “Hidden” personalisation.
Many users have also complained about spoilers in their Google Discover feeds and an option to “Hide Spoilers” recently appeared on one of our test devices. It appears as if this will only apply to scorecards served by Google since a match report for the specific scorecard we’d hidden appeared in our feed later that day.
There are additional suggestions that the Discover feed will come to a Chrome desktop browser in the near future and that perhaps some of the strange behaviour could be attributed to some early tests to allow for this implementation. However, yes, you guessed it: there has been no official comment from the G.
You can see the full comparison for all the most recent changes to the Google Discover documentation with this comparison sheet on the Wayback Machine. The screenshot above is taken from that comparison with the date range 8 April 2021 vs 19 April 2021.