The Google Discover Mistake Killing Good Articles Before They Ever Get a Chance

A lot of articles fail in Google Discover long before anyone gets to argue about “quality signals.” They fail because the preview package is weak.

If the title is muddy, the image is generic, the dates look messy, or the metadata does not line up, the card has already lost a chunk of its competitiveness in the feed. That is one of the clearest practical lessons from the Discover playbook already in this repo.

The short answer

The mistake is treating Discover like normal search. In search, a good query match can compensate for mediocre packaging. In Discover, packaging is part of the ranking contest because the feed has to win the click before the reader has committed to the topic.

That is why weak preview packaging kills good work. The article can be excellent and still underperform because the card looks generic, confusing, or low-value at a glance.

Why this matters now

The most expensive version of this mistake is not publishing “bad” articles. It is publishing strong articles with weak titles, weak hero images, duplicated dates, or inconsistent metadata and then concluding that Discover is random.

It is not random enough for that excuse. The article still has to become a compelling candidate in the feed, and the preview package is one of the first filters it encounters.

What to look for

• clear high-signal title and H1 alignment
• real featured image, not a logo or generic stock graphic
• clean publish date and coherent OG metadata

What to avoid

• treating metadata as an afterthought
• assuming the body can rescue a weak card
• publishing articles that visually resemble every other feed card on the topic

Final take

If a strong article gets ignored, the first thing to audit is often the preview package, not the prose.

The more detailed version of this workflow lives in the Google Discover technical playbook.