|Any Panda recoveries by removing links to parts of a website?|
| 7:54 pm on Sep 5, 2013 (gmt 0)|
I was wondering if anyone has seen page specific algo penalties after Panda? Penalties that seem to affect one part of a site but not the rest?
My site is broken into various sections, each section is linked to from my home page. I will refer to two sections in question as Section A and Section B.
Section A was in the top five results for almost every important related search for over five years until Panda. Now that section of my site is top of page two, or even page three and seems to float around almost daily in the results. Many other garbage sites and low quality pages are now outranking it.
Section B has pretty much the same structure as Section A and has gone to top of page one for terms that it never ranked on that well before. The only difference between the two sections? For years I worked at adding many links to Section A, including lots of links in comments using link titles which repeat the keywords on the page, etc.. All of those links got nofollowed years ago, but they are still present on the pages and indexed in WMT.
I never worked as hard at bringing up Section B, it has many fewer links and fewer self-added links from forums, comments etc. I am wondering if the link profile of Section B seems more natural to google, while Section A has perhaps triggered some algo penalty due to too many exact keyword low level links (even if nofollowed)
My question is...has anyone seen this type of penalty on their own site and has anyone tried removing links to get just a part of their site to recover? What types of links did you remove and what tools do you use to evaluate which links to remove? Thanks for any help!
| 9:52 pm on Sep 5, 2013 (gmt 0)|
Firstly, Panda has got nothing to do with links. It's based on content and user experience.
To answer your question, yes, I had a client who experienced a similar situation.
Two sections of the website lost nearly 60% of its pre-Panda traffic. The remaining sections lost only 10% or didn't lose anything.
It actually makes sense because if you're not a spammer or scraper, it's likely that only certain parts of your site will be affected.
Firstly, you want to determine if this is indeed a Panda.
In your Analytics, choose a data range long enough to show both the period when you enjoyed high rankings and period when the traffic declined. Print the graph.
Now, check the Google Updates Timeline: [moz.com...]
On your printout, draw a vertical line whenever a Panda update has occurred. Do these lines coincide with a drop in traffic?
Also, you want to make sure you know your niche fluctuations. In other words - are there any seasonal fluctuations in traffic?
If you've confirmed that this is Panda, create an action plan and start working.
A few things to look at:
- Do you have pages with thin content?
- Have you got technical issues with your site? Is duplicate content generated unintentionally? For example, do both www and non-www versions resolve in 200?
- Have you been a victim of scrapers? Anyone stolen your content?
- Is your content good quality? (be honest with yourself or get a focus-group together to let them evaluate your content) Is your content really adding value?
| 3:27 am on Sep 6, 2013 (gmt 0)|
Thanks for your reply adder, it's my understanding that your link profile has plenty to do with whether you may incur some kind of automatic penalty. Thats why everyone is rushing to clean up their link profiles and disavow links no?
I have already checked against the timeline and yes, it pretty much aligns with panda updates in late 2012 and 2013. The content is similar across the site so if it were considered thin all sections would go down together rather than some ranking better than before and others being hammered.
I don't have dupe content issues, a least according to WMT, and there have been no technical problems with my site nor any scraped content that I have found.
| 4:42 am on Sep 6, 2013 (gmt 0)|
Do you have internal links from Section A to pages in Section B, and vice versa?
| 7:18 am on Sep 6, 2013 (gmt 0)|
adder - Great post. Yes, you need to align many events on the site with the algo and see what likely caused what. Not all might line up simply with algo changes, though.
In this case, I'm assuming that this is the same site that ichthyous described in this thread...
Keyword albums / tag clouds triggering Panda algo penalty?
|At some point the number of indexed pages on my site started climbing from around 6,000 to 28,000 very quickly in December of 2012 and then started heading back down to where it started in May of 2013. During this time my site has lost approximately 70% of its traffic and it hasn't recovered no matter what I try. |
As I assumed from the description of the problem, the site had a large number of essentially empty search-generated tag clouds, and I felt that, for a bunch of reasons, they'd be a likely Panda disaster.
I'm wondering how the dates, sections, and traffic drops line up with the presence and activity of the tag cloud software. In this present thread, you posted about dates...
|...it pretty much aligns with panda updates in late 2012 and 2013. |
I'm also wondering how the A and B sections affected align with the number of tag cloud pages accumulated in each.
The thought that crossed my mind when I read the opening post... completely guessing because I haven't seen any graphs or stats... is that, if the extra links "worked" for a while, conceivably the extra traffic that came into section A because of the links increased the number of tag clouds in that section (ie, of thin/shallow pages in the section likely to lead to user/Google dissatisfaction), and ultimately caused the Panda hit. This is speculation on my part based only on the little bit I know about the site, and there may be many other issues overlayed, but I thought it was an idea worth tossing out.
Keep in mind that there might be a lag in some dates as things need to work through the index.
PS: Yes, Panda does affect different sections of a site differently. One technique used early on to revive the best pages of a site was to shunt weak pages off to a separate subdomain.
| 8:18 am on Sep 6, 2013 (gmt 0)|
Some links seem to get Panda angry, affiliate links. I used to find that if 20% of your pages had affiliate links you'd face off with the Panda but now it's more like 2% of your pages. Perhaps an affiliate link poisons the content around it? Only Google knows.
|Martin Ice Web|
| 8:40 am on Sep 6, 2013 (gmt 0)|
I experienced the other way. I had a summary page that was working great before Panda. It was pushed down the drain with Panda. This page had one outgoing link to wiki*pedia, nofollowed. I thought why should I write the Information in a new way if it is easy readable at wiki. I removed the link and within two/three weeks the site is back on #2.
Maybe big g does not like outgoing links anymore, because they want to be the only one that refers users to other sites?! Maby big g thinks if this site links to wiki, then it must be similar to wiki and wiki is a more trusted source?
| 11:32 am on Sep 6, 2013 (gmt 0)|
To answer what I think the OP might be asking...
We eliminated an entire section of weblinks. These were not exchanges, just legitimately good sources of information.
We also eliminated around 150 syndicated articles, most were about ten years old - from when we first started. The fear was duplicate content.
We've eliminated around 50 more articles that might be perceived as playing to certain key phrases.
We've gone back over nearly all (1,500) articles and de-SEO'd them. This was really just avoiding the repetition of a phrase. The intent was to make the article seem more natural.
We've introduced around 200 new articles in the last two years. We only publish an article if we truly believe it's the best page on the topic found on the web.
No recovery at all.. from around 1 million visitors a month to around 100,000. The site has around 1,400 pages and around 40,000 links to it (GWT), many from .edu, and some from .gov.
| 12:40 pm on Sep 6, 2013 (gmt 0)|
Zivush - there are links in the footer of every page that link back to the main sections of the site, otherwise they pretty much stand alone
Robert Charlton - Yes it's the same site with the keyword albums (not tag clouds). These keyword albums were in place long before Panda hit my site, but I did notice a huge spike in the number of indexed pages showing in WMT earlier this year. It jumped from roughly 7,000 pages on 12/2/12 to 26,000 by 4/28/13....then it turned right back around and started dropping pages until 8/11/13 when it leveled off at about 10,700 pages. I found the spike odd since I did not add massive amounts of new content to my site and the number of indexed pages was stable for years.
I implemented all the changes that were recommended in that post...I.e. removing all the tag links and no indexing all those pages. So far Google hasn't dropped them from the index though. I can see in WMT that the number of total pages indexed has barely gone down. I'm wondering how long that's going to take.
It also doesn't explain the slow crawl upward of one part of my site and the apparent demotion of another since the site design is a constant. In other words every section of the site and every page of the site contained the same tag links...I don't think the overall number of tag links in each section made much of a difference but I do think that cumulatively they could have had a big effect on my site overall as they pushed the proportion of thin pages up dramatically.
| 1:11 pm on Sep 6, 2013 (gmt 0)|
|For example, do both www and non-www versions resolve in 200? |
Should both return: HTTP/1.1 200 OK. ?
Thin shallow content 4-6 page sites since panda/penguin are now the top ranked sites in my niche. Two which are brand new sites. These sites have replaced much older top ranked for years, local sites that have been featured in numerous local papers, magazines and other publications and articles.
| 1:48 pm on Sep 6, 2013 (gmt 0)|
I am seeing the same thing in my niche...lots of really crappy, poorly designed sites that barely make an effort at content or usability have replaced the old stable of top ranking sites. Lots of news feeds, blogs, and in some cases the same e commerce site taking three or four slots on the first page.
I totally understand why Google wants to shake up the results so they can shake down webmasters for more ad revenue, but what about the effect on the user? Since Google is a monopoly the quality of search results is hardly important to them it seems...nobody is going to challenge them any time soon and they know it.
| 3:48 pm on Sep 6, 2013 (gmt 0)|
|I implemented all the changes that were recommended in that post...I.e. removing all the tag links and no indexing all those pages. So far Google hasn't dropped them from the index though |
I also noindexed several pages about 3 months ago, and they still appear in Google under the site: search. Google is extremely slow at responding to noindex and disavow. I disavowed my links in February and Google finally removed those from the link: results about 2 weeks ago -- that's 6 months! Those links still appear in my WMT data, but that doesn't surprise me as much since WMT link data also contain nofollow links. Those are just some of my observations...