Strange Behavior - Dupe Content Via Query String URLs?

[Alder]

Hey y'all, could use community help with some strange behavior I'm seeing with a particular ranking.

A week ago a high volume keyword ranking above the fold dropped off the map. I immediately thought must be an algorithmic penguin penalty (no manual action message) or panda / dupe content issue. I think it's dupe content at this point because I found my former ranking page in the omitted results section for the keyword we used to rank for.

The strange thing is that without making any changes, Google would momentarily show our domain ranking high page one again, but with a strange query string URL. At first just domain.com/page/? whereas the old ranking was held by domain.com/page/ but now I see several long query string URLs floating around that the engines don't seem to know what to do with. Canonical tags are in place to canonicalize any query string URL back to the top and I have now designated query string URLs as unimportant in Search Console parameter filtering but these URLs persist.

I ended up deduplicating content to a page on another domain we own (think that was the original problem) and there seemed to be a positive effect but now we are top of page 2 with a much longer query string URL as the ranking page. It seems Google wants to rank everything but the former ranking URL even though it's the most authoritative by far, has canonical signals in place, and is now no longer duplicate content. Content checker tool showed 60% similarity to the other piece, which is a ratio I've never known to cause dupe content.

We found the source of the query string URLs to be from an external site that has a link to us but it's a buggy site so filtering on the page adds the string to our URL, so Google can find them and thinks they're significant.

Long question short, has anyone had trouble like this? Getting weird parameter / query URLs to get out of the index in favor of the non-parameter folder? Is it possible the main folder page got hit with Penguin and is "banned?" Still, I don't know why Google would go out of it's way to rank query string copy pages in its place if that were the case. Any help greatly appreciated.

An example of the URL looks like this:
domain.com/page/?CustomerSubscriptionTrack1PageSize=1&CustomerSubscriptionTrack1Order=Sorter_ID&CustomerSubscriptionTrack1Dir=ASC&CustomerSubscriptionTrack1Page=3&WorkOrder_TBLOrder=Sorter_AssetID&WorkOrder_TBLDir=ASC&ID=106

[Alder]

Hey James, sorry to hear you're getting blasted by negative links and appreciate your responses here.

I actually sorted this one out (fingers crossed it stays that way) by having the dev team implement a redirect rule that 301 redirects any query string back to the folder we want ranking. Similar signal to what the canonical tag would send but in my opinion a stronger signal since there is no longer a way to reach those weird query string URLs with a 200 response.

Once that was implemented the appropriate page was right back to its old high ranking position and the query strings are hardly to be seen in the index and are no longer preferred to the old ranking page - so looks like all is right with the world again.

We also disavowed the domain that was the source of many of the query string URLs. I don't think it was a case of negative SEO - just bad coding on their side. I'm not sure what exactly did the trick but I suspect strongly that the 301 redirects is what solidified the index due tot the strong correlation of that change with ranking recovery.

Maybe you can employ a similar solution whereby you can disavow domains where these links originate or set up server side handling to manage URLs of a specific pattern - for example, any URL containing "pornsite.com" if not any query string altogether (in our case we don't have any use for query strings in our URLs so just bagged them all).

Thanks again,
Matt

[Alder]

Thanks for the response, James. The odd thing is that canonical tags are implemented correctly as far as I can tell. In the of each variation you can find the following code:

rel="canonical" href="https://www.domain.com/page/" />

(still using my example so as to keep the site anonymous)

And this code had been in place well before the issue arose. So yes, we are sending that signal to Google to apply canonical back to the top in every case, without query string.

Not sure what you're confused by in Search Console - the platform provides a tool to deal with parameter URLs just like the ones I'm seeing. I used it to mark all parameter URLs as not changing content, which should designate to engines to exclude them from the index.