Different pages presented every day for the same search query.

[Steiner 2012]

I manage a <widget> site, totally dominated the sector for 2 years, solid business, sailed through every penguin and panda without any loss of position.

Site was literally #1 or 2 for any <widget> in this location and sometimes double and triple breakouts.

Now the site has lost a lot of rankings, can't see why it has been slapped at all and since the first week of October a different page has been displaying every day.

Seriously, the most competitive pages that were at #1 thru #3 have all gone. Some had matching URL's, H1' a variation on H2, almost 2,000 words of highly relevant content about <various aspects of widgets>.

Pages were not keyword stuffed and they are daily replaced with pages that had irrelevant URL's, no matching H attributes and 30 words of irrelevant content.

Swings are massive daily: #3 to #105 next day to #12 next day to #84 and every day a different page for that same query.

I have never seen anything close to it.

In WMT; no crawl errors, no manual warnings, no html improvements, nothing!
Some of the less competitive terms are still hanging around top 6-10 (useless)


I went through the site and code but nothing, not a thing that I can see, I really cleaned the titles, which were great before but just to be sure I went through the site and submitted new sitemaps, still the same, last 11 days 11 different results page for 1 keyterm.
Bizarre.

Anyone seen anything as fluid as this before?

[edited by: aakk9999 at 10:52 pm (utc) on Oct 20, 2013]
[edit reason] Exemplified niche and product [/edit]

[JD_Toims]

I've seen something similar and as far as I can tell it had at least something to do with anchor text. I also think some of it could have to with your searching habits -- A few months ago Google announced they had implemented a system that would make it less likely to see results not clicked on in subsequent searches and pages from sites not clicked for deeper searches [or something to that effect].

As far as a "fix" goes, I would make sure you're checking your rankings "clean" and also double check on keywords used to link to the page that should rank, because sometimes when inbound links [or whatever] cause a page to not rank where it should a different page from the same site will replace it in the SERPs even though it's not as topical as the page that's been replaced.

[lucy24]

A few months ago Google announced they had implemented a system that would make it less likely to see results not clicked on in subsequent searches and pages from sites not clicked for deeper searches

Is there a thread about this somewhere? I've occasionally wondered whether test searches might in the long term bring lower results for your own site, because you never end up clicking on it. (You don't want to go there. You just want to see it in a nice Top Three spot.) Unless you scrupulously wipe all cookies, lie about your UA and restart your modem to ensure a fresh IP address --assuming you're dynamic and can even change it.

[JD_Toims]

Is there a thread about this somewhere?

Yes, it was a while ago and I haven't been able to dig it out -- It was around the time domain-crowing was a hot-topic, but I don't remember exactly when or if there was very much discussion on it. There is one though.

[Steiner 2012]

I also think some of it could have to with your searching habits

Wouldn't apply, I have checked the rankings of that site and a number of others using software not searching.

Anchor text is a difficult one, we sell and rent condos, it is in the name of the site, and naturally 'condo' has to be in the url of the pages for relevant condos. That said our URL is the largest anchor text then our company name, then 'visit advertisers website', click here etc.
Couldn't see the density being too high on any one of those.

All of the links we have are from 3rd party property portals in our city, they should be the perfect link and even then a lot are the page url of a specific condo or 'more pics here' etc.

Same pattern again today: 1 keyword down 62 places, next one up 43, next one up 9, next one down 80, next one up 55!?!
Rankings are like water.

[JD_Toims]

...then 'visit advertisers website', click here etc.

Are those links nofollowed? It's fairly easy to determine algorithmically [whether correct or not] those links are not "freely given" to the receiving site/page.

Couldn't see the density being too high on any one of those.

I'd try to throw density out of my vocabulary since keyword density has been obsolete as a ranking factor for a number of years now. Natural co-occurrences of phrases associated with a topic is the "new" keyword density and having "advertiser" in, or surrounding, the linked text to a page is a simple "red flag" to find algorithmically as far as whether the links were "naturally" acquired or not goes.

Also, something to keep in mind is it doesn't matter one bit what you, or I, or anyone else thinks is "ok" or acceptable. All that matters is what an algorithm finds and evaluates something to be and algorithms seem to differ often from a webmaster's idea of "acceptable".

[Martin Ice Web]

@Steiner,

i think this has something to do with the Integration from panda in the main algo. As it now does not run on scedule but permantent this seems to be the flux. Also all datacenters seem to have a different state.
I like to use the search for "bing". There is almost every day new result numbers from 214.000.000 up to 320.000.000 pages.