A few things I can think about:
1) You may be affected by Google's Rank Modifying Patent [webmasterworld.com...]
2) If the ranking of your pages is influenced by user metrics then a change to your page may have reset those and Google will have to build these again.
3) The new content added has changed the focus of your page so it does not rank well for phrases it used to rank, or the page title is not so relevant to the page any more
How long ago have you made the change?
How long after the change did the page dropped?
How much time has now passed after the drop?
Maybe you just need to wait it out a bit longer for the ranking to stabilise?
If it were simply google not trusting pages with content additions, then how would that affect pages that allow blog commenting? Because it would seem to me that it would be a similar scenario: a site ranks well, and (from time to time) new content is added.
Personally, I don't believe that the rank modifying patent is in use by google as much as others do. I am sure it is used, but I don't know think its use is as widespread as others believe.
My first thought is that maybe you tripped some sort of over optimization trigger for a particular keyword or keyword phrase.
Just to be clear:
You did not make any changes to existing content, you just added new content?
How long has it been since Google crawled the pages? Google is really slow these days to rank some pages, and it may just be that their bogged down algorithm needs more time to work through the changes you made.
aakk9999 - I made the changes in early Dec 2012 and they dropped off Google's radar within 3-5 days. There was a big Dec algo update but this was about two weeks before other sites started reporting it. Your #2 and #3 suggestions don't seem to fit but #1 might very well re:Google's patent:
If I got one of these responses it has stuck ever since.
|Bill explains: |
During the transition from the old rank to the target rank, the transition rank might cause:
a time-based delay response,
a negative response,
a random response, and/or
an unexpected response
If these transitional ranking shifts are followed by what looks like a responsive action from the website, the URL or website which was previously only SUSPECTED as spam, might now be positively tagged as spam.
That would be correct, I should have just written that in the opening post. Given that Bing and Yahoo still like these pages I don't think I did anything drastically upsetting and the metrics remain virtually identical for both of them. I have no Google metrics to verify for these pages anymore and so I doubt it's a metrics thing. The rest of the site didn't suffer the same fate.
|Just to be clear: You did not make any changes to existing content, you just added new content? |
If it's the randomized response it would suggest that I don't change the content back since that would be trying to please Google, but how long does it last? It's been 8 months. What would you try, if anything, at this point?
"If it's the randomized response it would suggest that I don't change the content back since that would be trying to please Google, but how long does it last? It's been 8 months. What would you try, if anything, at this point?"
You said you did this on a couple of pages, right?
It might be a good idea to experiment and change one or two of them back and see if it makes a difference.
I personally think that it is more likely some sort of over optimization penalty.
Wikiepedia pages get content added all the time.
Pages that allow comments get content added all the time.
Ecommerce category-level pages get new products added and old products deleted all the time.
Event pages get content added and removed all the time.
So is there some classifier that google uses to denote certain pages / certain sites as being "evergreen" and those sites are not allowed to update their content?
Aside from the thought that the content may have tripped some over optimization penalty, how UNIQUE was the content that you added?
Is it possible it was more of a Panda effect than a pure OOP?
Getting back to changing the pages back or not; I would possibly experiment, depending on how many pages you changed.
I would possibly change one or two back to how they originally were, and maybe 301 one of the pages with the additional text to a new URL. I would probably also try to keep the same new text on one page but try to deoptimize it a bit for its keyword.
I don't really know how much there is to gain from waiting it out and in today's google world, I think we have to take chances and experiment - hopefully not on our main sites but on sites or pages that don't rank as well as we hope.
Hope this helps.
|It might be a good idea to experiment and change one or two of them back and see if it makes a difference. |
Good suggestion from Planet13 - I think I would try the same - pick one page, revert it to what it was and wait a few weeks. I would also suggest that after reverting, you do Fetch as Googlebot in WMT and then execute "Add to index" to try to speed up the indexing of this page.
Yes, I made changes to a handful of pages only, mostly discussing new aspects of an old widget. All original, including some ideas that are my own.
One other thing, the pages have images and 65% of the traffic was coming from image search. They previously received 90% of their traffic from image search up until Google changed the layout of their image search engine. Do you think that Panda/Penguin alter image search results as well since this dried up too? The images are still in Google but down 5-6 pages instead of being up top.
I don't see images being the problem, the image traffic metrics didn't change except in reduction, but nobody has really discussed what Panda/Penguin does to image traffic if a page gets most of it's traffic that. I have however noticed that when a page gets a lot of traffic from image search it can do poorly in web serps and vice versa, as if you can only rank well in one at a time.
I'm grasping at straws as to what it is Google dislikes about these now so I will go ahead and return a couple of them to what they were.
Years ago I had some pages that performed well and I decided to make them even better with some good, additional content.
In making the pages better for the viewer I obviously messed with the key indicators that Google was using to determine ranking and the pages suffered major falls in rankings and traffic.
The lesson at the time seemed to be that adding extra content can dilute the factors that are propping up good rankings... proceed with caution and think carefully before doing this.
An interesting observation: These pages were published years ago but when I set the 'search tools' option on Google search pages to 'past year' these pages show up in results again. I'd say that definitively means Google updated the publish date as I updated the content and I now have a STRONG suspicion that this has devalued the old incoming links for these pages.
A link today is a vote for the page today. If you update the page tomorrow then yesterdays links are no longer as strong a vote for the 'new' content. It's the only thing that makes sense right now. I've changed templates and other parts of pages before without problem so this would seem to apply only to the actual content.
It is a strong correlation but I would be very surprised if that's how Google value links. A page can be modified for many reasons, not all of which would render the existing links less valuable.
I would however lean towards the theory that the material change on the page probably triggered an re-assessment of the page both on and off it, including the backlinks. This may have the same end results as JS_Harris suspicion but the theory of "devaluing the old links because the page content has changed" is a bit of far fetch.
Back in November 2012, I posted a thread on the same topic. I have taken one of my site and updated 1/4 of its content and the results were fantastic not only for these pages but for the overall site's ranking.Actually, the traffic was doubled within 3 months.
Read the discussion here - [webmasterworld.com...]
P.S. - I haven't changes URLs. That might be the difference between JS_Harris case and mine.
Assuming the pages had mostly the same keywords and basically the same content, I would've thought the pages would've stayed but if you're talking complete rewrites, it's just like you posted a new page IMO.
When you have updated content it's often best to post a new page and then simply insert in the old page "UPDATED: MM-DD-YYYY <Link to update>" and leave the old page completely as-is to maintain it's position.
Not knowing what you posted and how it was previously linked with anchor text, perhaps Google simply saw an apples and oranges issue that the things didn't match well enough and it tossed it down the well. Sites and content pages change hands all the time, fall to parking services, etc. and do not maintain their status when these events happen. Perhaps your changes triggered that part of the algo that thought the pages were no longer part of the same set.
|A link today is a vote for the page today. If you update the page tomorrow then yesterdays links are no longer as strong a vote for the 'new' content. It's the only thing that makes sense right now. |
But then probably a bazillion other web folks would have noticed the phenomenon, not just you. Instead of a huge change in the PageRank algo -- which that would be -- I would chalk it up to something simpler, maybe along the lines incrediBill suggests, maybe something else.
One other question:
did the pages start ranking for different keywords?
I know you said that traffic from google plummeted, but whatever traffic you DID get from google after the change, was it from different keywords that HADN'T been sending any traffic to the page before?
The images traffic is a big clue. It is not at all like text SEO. WMT can show you the relative traffic between images and html I think. Look for coincident scapers of the images. We had a #1 image from the 90's that got scraped and took our place on G images search. We sent them a cease and desist notice. They removed it. It took a good 6 weeks for us to get back into the images listings at #3.