Change in Search Results Due to Supplemental Cleanup?

[g1smd]

This was first posted a couple of days ago, but doesn't seem to easily fit into any of the other current indexing, SERPs, or WMT threads...

Google seems to have cleaned up Supplemental Results in the last 48 to 72 hours, and perhaps had a temporary trial run at least several weeks ago.

I have been looking at a large collection of pages with certain spelling mistakes corrected and certain technical terms rewritten over the last few months.

Google has been slow to drop these pages from site:domain.com keyword searches, where "keyword" is the older version of that content, the one with the incorrect spelling.

For quite some time, these pages continued to appear in the SERPs for the old keyword, even though that word was no longer on the real page, no longer showed in the Google public cache of the page, and no longer appeared in the snippet for that page.

Those are what I called "historical supplemental" results a few years ago. Almost a month ago, they cleaned up those results, and dropped the old pages - but then added them all back in a few days later (hmmm, not sure about "all", but certainly "most").

A couple of nights ago, they cleaned them out again, and this time they appear to be staying gone. So, my site:domain.com keyword searches now return only a few pages, pages where the keyword still shows in the snippet because it is a page that is yet to be edited, or is one that was edited only in the last week or two - and hasn't been re-crawled and re-cached as yet.

[tedster]

Interesting report, g1smd. If I understand you correctly, you're pointing to the fact that the Supplemental Index is not just a url, but a url PLUS a cache date. So now the older cache date versions of a url seem to be flushed from the Supplemental Index.

I've always wondered what the value of those urls would be in the SERPs - since as you say, the keyword would be gone from anything the user could access. Keep us posted on any further changes you see, since you have an uncommon test set of urls - and thanks.

[g1smd]

Sure, but when I say older cache, I mean an older copied stored internally by Google, and not the public cache which is always the latest version they have, and which is displayed with a date caption.

Google keeps hold of older data about the "page" and continues to use that old data to return the page for queries that also match the older version of the content (as well as for queries matching the new content, as expected)... long after they have cached the newer version of the page.

When a URL is returned for a query that matches the old content, the URL is returned as a Supplemental Result (and used to be tagged that way), but you'll not see that old content in the public cache when the URL is being returned in that way.

It's interesting to play with this, and easy to do if you have some unique words on the page that you could reword.

.

There's a related effect where Google continues to show a URL long after it has been turned into a 301 redirect (like when fixing www and non-www issues), or when a page has gone and returns 404.

I think Google has adopted the stance that if they immediately deleted all reference to the page, people who had found that page a few days or weeks ago and were now looking for it again, might think Google was broken if the page in question were no longer in the SERPs, so in this case the cache of what was is useful to still see that content live on for a while longer.