The verdict
Panda has not been a separate update since 2015, and the helpful content system joined it inside core ranking in March 2024. There is no refresh to wait for and no page score to fix. On the catalogues I audit, half the index being thin is normal, and the fix is a template decision, not four hundred rewrites.
Clients still say “Panda” when they mean thin content, and the word carries a picture with it: a filter that ran on a date, marked the bad pages, and would let the site go once the pages were fixed and the filter ran again.
That picture was roughly right in 2011. It has not been right for a long time, and acting on it now sends the work in the wrong direction.
What Panda was, and where it went
Google announced it in February 2011 as a change to “reduce rankings for low-quality sites”, meaning sites that are “low-value add for users” or “copy content from other websites”. It noticeably affected 11.8% of queries, which is why it got a name. When Google launched Penguin the following year, it framed that one as about 3.1% of English queries, and compared it directly with Panda as roughly a quarter of the size.
Google’s ranking systems guide now lists Panda as retired: it “became part of our core ranking systems in 2015”. The helpful content system, which carried the same idea forward, followed in March 2024, and Google’s note on that update said plainly that “there’s no longer one signal or system used to do this”.
So when a site loses ground on quality today, there is no Panda to have been hit by. There is a core update that re-weighed something that was already true about the site. Penguin went the same way a year after Panda, and its consequences are different enough that I have written up what happened to Penguin and bought links separately.
What it looks like in an audit
What I find is almost never a few bad articles. It is a template.
On two catalogue sites I audited, 466 of 812 indexable pages were thin on one and 93 of 189 on the other. On a classic car marketplace, 8,317 of 9,220 indexed pages were orphans: expired listings nobody linked to and nobody would choose to land on. None of those pages were written by a person who did a bad job. A platform produced them one URL at a time.
That is the part the old Panda picture gets wrong. It frames quality as something you fix page by page. On most sites the low-quality share was produced by a rule, and it gets fixed by changing the rule.
Why “fix the pages and wait for the refresh” fails now
- There is no refresh. Quality is assessed inside core ranking. Google says it can take “several months” for its systems to confirm an improvement, and sometimes the next core update. Nothing runs on a date you can wait for.
- Rewriting is the slowest fix available. An owner told that 466 pages are thin hears “write 466 descriptions” and, quite reasonably, does nothing. Most of those pages should be removed or noindexed, which is one decision per template.
- Bulk deletion has its own failure mode. I have seen a pruning exercise make a drop worse, because the pages removed were carrying the only editorial links the domain had earned. Google’s own core update guidance calls deleting content “a last resort”.
What I do instead
Count the index by template before reading a single page. Products, collections, tags, archives, blog, paginated series. The split tells you whether you have a writing problem or a platform problem, and on catalogue sites it is nearly always the second.
Then stop the generator. One rule for expired listings, one setting for tag pages, one decision per facet. The backlog stops growing on the day you change it.
Then the pages that matter: the ones already earning impressions and the ones you sell from. That is usually thirty or forty pages, not four hundred, and it is where writing earns its cost.
If the quality re-weighting arrived with a confirmed update, this is the part of recovering after a core update that is actually in your control. The rest is time.
The short version
- Panda is a retired name. Part of core ranking since 2015, with the helpful content system joining it in March 2024.
- There is no refresh to wait for. Improvements are confirmed over months, sometimes at the next core update.
- Thin is usually a template. 466 of 812, 93 of 189, 8,317 of 9,220.
- Count by template first. It tells you whether it is a writing problem or a platform problem.
- Stop the generator before clearing the backlog.
- Delete carefully. Check for links before anything goes.
If a site should be ranking and it isn’t, that’s the work I do. SEO recovery covers quality-related drops and what realistically comes back. If you’re not sure which of several plausible problems is costing you, that’s what an SEO audit is for.
