One publisher's Google Discover traffic loss and recovery

A consultant’s Discover recovery story is a different kind of source than the algorithm-update coverage above: it is one person’s account of one client, not a dataset or a Google announcement. Independent consultant Nerijus Masikonis published a case study describing a publisher client whose Google Discover impressions dropped to zero after a core update, stayed there for weeks, and then came back. We read the post directly. Here is what it actually says, in the consultant’s own words, and why it should be read as a single anecdote rather than a repeatable playbook.

What happened, in the consultant’s own account#

The client’s Discover traffic collapsed after a Google core update. Per the post, “the impressions went to zero and stayed there” for about seven weeks. Masikonis writes that his audit afterward found “articles without Discover-ready images, broken images and structured data, a schema bug crediting articles to a different site, missing authors and descriptions, and thin pages.” That is a list of quality and technical signals, not a single named cause, and the post does not claim to know which item on that list mattered most to Google.

What changed, and how Google responded#

The fixes described are concrete: Discover-sized images went out to “almost the whole catalogue,” reaching 97% of articles by the consultant’s own figure. The broken structured data and the schema bug misattributing articles to another site were corrected. Missing author bylines and descriptions were filled in. The thinnest, lowest-quality pages were removed rather than kept. After that, Masikonis writes that he “held the site steady and waited.” Discover traffic did not return gradually, it came back “on the same day” as Google’s next core update, across every title in the client’s network, with the flagship site going on to post its best Discover month on record.

Why this is one case, not a study#

Masikonis states outright that “traffic volumes are hidden at the client’s request,” so there is no absolute click or impression figure to check the recovery against, only the zero-impression period, the 97% image fix, and the relative claim of a record month. This is a single consultant’s self-reported account of a single client engagement, published on his own site, not an audited tracker dataset or a Google statement. Nothing here establishes that the same fixes would recover Discover traffic for a different site, or that the timing next to a core update was cause rather than coincidence.

Why we care#

The value of this case is the checklist, not the numbers. Discover-ready image sizing, clean structured data, correct site attribution, complete author and description metadata, and pruning thin pages are all things a publisher can audit today without waiting on Google. Treat the recovery timeline, the same-day return alongside a core update, as one data point worth watching for, not a mechanism this post proves. Read it alongside our coverage of the February 2026 Discover-only core update, which found that most of an article’s Discover clicks land within a day of publishing, for the fuller picture of how fast this channel moves in both directions.