| This 38 message thread spans 2 pages: 38 (  2 ) > > || |
|It's time to delink inbounds urgently - but how?|
Over the last 10 or so years the temptation to take a paid link here or there, and/or build lots of links across crappy sites has been sucked up up by multitudes of webmasters across the globe. Now those sites are in trouble in the SERP's.
Some of those site owners would be reluctant to cull all links, as this will also create big trouble. Site's that embarked on this strategy, that i've seen, have tanked correspondingly.
Links to cull, i believe are:
-overuse of same anchor text
-overuse of links from crappy websites
Links to engage are from high quality sources, connected to reputation. Not just Toolbar Page Rank.
Now here's the catch. Lot's of you have great links mixed in with crappy links pointing to the same URL's. You can get those links changed.
What's the strategy? Do you let go the site ? Do you change the URL's ? What is the checklist and actions for dealing with this ?
What if you just continue what you're doing, but build higher quality links to drown out the lower quality links?
If "negative SEO" doesn't exist as Cutts says it doesn't.....then simply diluting the overall formula should have a good effect upon the website.
I wonder if that would work on key terms [ versus the website]
What i mean is if you have 100 crappy link with the anchor text "red widgets" , how is that term going to be restored.
Is it possible that the tanking is the result of the links no longer working and that the new positions are pretty much where they should rank if those links did not exist?
Personally I don't think it's as simple as an abundance of the same anchors causing the drop, as a single factor causing the effect all by itself. Something I am familiar with contradicts the idea that a single factor (anchor text) alone is enough to cause a Penguin effect. I lean toward other factors on the linking page, possibly in concert with the abundance of similar anchors, working together to cause the effect.
Now we have to define what the effect is. Is the effect simply a site ranking where it deserves to rank (because of low quality signals)? Or is there a penalty that needs to be lifted first?
If the effect is the first scenario (site ranks where it belongs), then the suggestion to build better links may be the direction to take. If the effect is the second scenario (it has a penalty), then removing those links is imperative.
If I Had to bet...
A lot of site owners are going to drive themselves crazy removing, or trying to remove 'bad' links. But after all that, their sites rankings won't improve.
On the other hand, cutts & co. will have gained a major coup - putting a serious dent into the paid links "industry".
|I lean toward other factors on the linking page, possibly in concert with the abundance of similar anchors, working together to cause the effect. |
Would you include poor quality sites e.g. typically blogs that no-one reads, or off topic blogs linking, per MC's interview with Eric Enge
|Would you include poor quality sites e.g. typically blogs that no-one reads, or off topic blogs linking, per MC's interview with Eric Enge |
Sure, of course. Blogs that nobody reads and directories that nobody uses send no traffic. A link that doesn't send traffic is poor quality (?and unnatural?). A backlink profile with many such links is poor and thus gets your site dinged.
|John Mueller @ Google said: In general, if you remove the page that is being linked to (such as a spammy forum thread) and make sure that it returns a 404/410 HTTP response code, we'll ignore the links to those pages. If these links are primarily pointing at threads that you've removed, then there's no need to move the whole forum to a new URL. If you find that there's a significant number of problematic links that you can't remove which are linking to general parts of your forum (eg the forum homepage), moving the forum to a different URL might be a possibility (but I'd only recommend doing that if you're absolutely sure that these links are causing problems -- we're pretty good at ignoring spammy links). If you choose to move the forum, then StevieD has some good suggestions on how that could be done (make sure that your 404 page is useful to users too). [productforums.google.com...] |
It's interesting that he went on to say that those links generally only harm the specific pages they point to and not the whole site.
But I wonder if this is the best solution when you've got great links mixed in with the bad ones, which cannot be easily moved to new pages.
|Is it possible that the tanking is the result of the links no longer working and that the new positions are pretty much where they should rank if those links did not exist? |
I have to agree with you MB. I think Penguin simply "cleaned out" and devalued lots of web pages, and sites (like mine unfortunately) simply lost some "link juice" and are pretty much ranking where they should be.
I still think that removing some really sour apples, coupled with a greater focus on developing more good, quality links, is the best remedy.
Here are points that might help as we have a bit of experience recently with clients in some of these positions.
1. If the site didn't receive an unnatural links penalty then its best to assume, initially that the website did not raise paid links flags. It's not 100%, but its the best base to start on.
2. Download your links from Google and favorite inbound links vendor and mix together in one long spreadsheet. Identify sites that link to you with low scores. It takes a long time (weeks) to get through and hand prospect 5000 links but it is well worth doing if you have a lucrative business model and have lost substantial revenue from Penguin. Good news is most people do not have 5000 links.
Email off and keep track of removals. Anything on websites that are low quality, have poor inbound links or metrics, or likely do not pass traffic should be turfed. Some sites will blur the lines, use your best intuition to make distinctions on those.
Use social signals in your evaluation.
3. Look at inbound anchor text. Variation and anchor text type is the key. Take aggregate snapshots of others in your niche, and develop and understand what makes a strong profile in terms of inbound link anchor text. Brand vs www, Exact match vs long tail.
Review the numbers again and again and know them very well. Then set your bars and go off and email out to sites to change anchors, where possible.
Look at competitors and where they list - especially in associations and memberships / local citations. Much of the time those types of websites will only list you as www or brand, and so those types of listings begin to dilute (aka 'make more natural') a strong over-optimized link profile.
4. Be sure to look at link diversity. Analyze the top 10 competitors in terms of the "link buckets" or link categories they have. Develop a quick working model - or pie charts of link category percentiles for all top ten websites - AND aggregate. Understand where you are deficient, and proceed in working those type of link angles.
Over time you will begin to see improvements in traffic. If you have over-optimized the website, then of course you must begin to de-optimize as well. Internal links seem to be a big one here. Keep internal links natural. Try to stay away from aggressively matching Title, H1 and page-names + internal links.
5. If you received a unnatural links penalty all is not lost. We just helped two new clients dig out of these penalties.
They are not easy to get out of but it is possible. Both of ours required 2 reinclusion requests.
Apart from the fantastic guides already available on the subject of reinclusion, the only thing I can add is:
a) Look for sources and patterns of links across networks - even ones you did not build - and remove
b) Remove any links on any directories that use remotely any wording such as sponsored, pagerank, links, SEO.
c) Read the Google guidelines again
d) Anything that looks remotely paid needs to be removed immediately. Sometimes the difference between appearing paid and not appearing paid is made by using an image link instead of anchor text, brand name or www.
e) Do NOT re-include too fast. Be absolutely sure all links that do not look editorial in nature are removed.
f) When you do reinclude be sure to include ALL work in the request that you have done to help remove links. Be honest - explain what happened.
In both of the cases we submitted we did not blame the links on a previous SEO. We simply mentioned links were built in the client profile and we removed them because we felt they went against the Google guidelines for linking.
Hope this helps.
Great post Cain!
With point D are you saying that image links are a good way to deal with text links that might appear to be paid?
pt B: If G is using this as a criteria, they are being highly presumptive and unfair.
We label on our site "sponsored links" (as G once specified themselves for us to do - now they would penalize for it?), but they are all nofollowed as LATER specified by G also. Furthermore, we specify on our advertising info page that "our advertising is solely for related, targeted traffic and we DO NOT accept any requests for advertising intended solely for pagerank manipulation..." So if their algorithm is going to search for those terms it BETTER be good enough to distinguish between pro and con statements!
BTW, I tend to agree with onepointone and martinibuster (the simplest solution) that post Penguin rankings dropped simply because G "discounted" links from and/or demoted sites which were selling anchor text links thus reducing link juice passed on. Thus removing them probably won't make a difference, and risks removing some which maybe AREN'T demoted after all.
Worse still is it possible to poison oneself? I noticed many of our competitors drop drastically with Penguin and in nearly all cases they are ones which (like us) have more than one domain dealing with the same subject. In this case it is totally natural to interlink them, including in in-line text (example: "do not use our red widget when our (link:)green widget will work better" where green widgets are dealt with on a separate domain). Another example is where you want to link to separate page which explains what in general a widget IS. Do you need to duplicate the same description page on both domains? G doesn't seem to like that concept either. Finally there is the case of text navigation links between domains... sorry, but a widget is simply a widget, sometimes there's just no other thing to call it. Does G know you own all the domains and are linking them for totally legitimate, and user helpful reasons, most likely not, but they should!
If you read back in the early html1.0 references, the whole concept of hyperlinking was invented/intended precisely for this inline use, not for how it has been twisted, misused and is now expected to be used by G. Because early surfers (and early site writers) were inept about clicking on blue text everyone started saying "CLICK HERE". Although not proper, early incorporators discovered this worked very well, until G came along and assumed this was a clue that the linked to page was about "click here"'s.
G's algorithms simply are not smart enough to do what they are trying to do, so they are penalizing those of us smart enough (and perhaps outdated enough) to use the internet as it was ORIGINALLY intended. With great power (a majority of the internet search market) comes great responsibility and G simply is not up to the challenge.
And no, having a domain over 15 years old and many pages with HTML 2.0 in the header does not apparently exempt you :-).
In requiring a confession to get back in, Google is also collecting data about the networks that are selling and trading links. Gaining more understanding of how the links are aggregated and packaged, in what countries the links are originating from, etc.
I saw a manual penalty letter for unnatural links from G where they gave a few samples of the types of links that should be removed and one was due to submitting to a quality directory about 5 years ago when it was still free and now it's charging for submissions. Should those be removed also?
How do would one deal with repackaged DMOZ Clone MFAs in this case?
I just got 800+ inbounds, few month ago, from <snip>)... contacting them is N/A. Pure Junk, site took a dive for few main money keyword phrases...
[edited by: martinibuster at 6:12 am (utc) on Jul 24, 2012]
[edit reason] Removed guess-able specifics. [/edit]
|With point D are you saying that image links are a good way to deal with text links that might appear to be paid? |
Absolutely because images denote advertising more than a text link does.
|pt B: If G is using this as a criteria, they are being highly presumptive and unfair. |
You mentioned you nofollow links. That's what Google mentions they are asking for in your particular position of advertising and that is in fact what you are doing anyway since your links don't pass value.
Regardless of whether Google linking "sponsored" with paid is unfair, it seems to be the current case. If I were a search engine I would have a special "cabinet" for that keyword...
I agree with martini and you on that point (of post Penguin drops because of reduced value of links) At the same time I have seen pure examples of sites actually penalized by having hundreds of directory links (only) post Penguin now.
Matt Cutts announced today the next level of message transparency in which he mentions that websites who received the recent unnatural links penalty would receive next-level followups - and some that we follow did.
However, none of those lost any traffic in Google and got the better of the two followup emails, the one that talks about Google discounting the unnatural links and not the website itself.
If this is really necessary, it indicates a major defeat for Google.
The point of a search engine is to find the best content, whereas needing to do things like this means that SEO knowledge becomes more important relative to content.
There is lots of superb content written by people who have only the vaguest idea how Google works, and who do not use any Google Webmaster Tools or any SEO tools to analyse incoming links.
This is potentially very time consuming, can mean dealing with problems that go back years, and may not be in your control at all.
How do we know that we have an incoming links problem? IS it safe to assume that if we do not get a message through Webmaster Tools that we are OK?
|How do we know that we have an incoming links problem? |
A dramatic loss of ranking and traffic.
Is there any single piece of evidence that cleaning up helps?
|Top Ten Strategies For Dealing With Poisoned Inbound Links |
Any chance you could change that title on the main page? It gives the impression that there is something dangerous about some links. Links are not of themselves "bad" or "poisonous". Google has deemed them so. Remember the Cold War?
"If someone we think is a traitor mentions your name, we will arrest you."
"If a site that does not agree with our ethics mentions you, we will penalise your site."
Heck, you could take Google's mentality right back to "she's a witch! no she's a witch!" if you like.
I think the real question here is, how do we get Google to stop telling us any links are "bad" for our sites and just ignore poor quality links? Should we all buy $9.99 directory submissions for our competitors? Should we all post our competitors websites on forums asking for link exchanges? For $10 I can get 1000 blog comment links pointing to my competitor!
How do we make this situation impossible for Google to keep a grip on? Whoever is with me, please PM me your URl and I'll get to work on damaging your website until Google finally wakes up and stops forcing decent hard-working webmasters to resort to Adwords ...
... wait a minute ... I think I've just ... no, surely not?
|A dramatic loss of ranking and traffic. |
I saw a dramatic loss of traffic to one of our sites - we lost between 50-70% of google referrals. What happened however is a loss of traffic for some certain (optimised) phrases. An URL went from page1 to page 8 for [keyword1], but ranks fine for a bunch of other keywords.
So before you all start running to un-link your sites from any other site on the web, make sure you get rid of proper anchor text links.
Is a loss of rankings only for certain keyword-page combinations an effect of the Penguin Update?
Or is the Penguin Update supposed to work sitewide and my issue may be simply inbound links losing power?
|Or is the Penguin Update supposed to work sitewide and my issue may be simply inbound links losing power? |
Those are good questions and observations. It generally starts with your best keyword phrases and then slowly spreads to other keyword phrases, where you start to lose rankings across a wider range of phrases. If it doesn't spread to a wider range of phrases then it could be something else, many other things.
Absolutely, do not panic. Never panic. If you are certain that the links you built are crap, then you're probably starting to rank where you are supposed to rank should your crap links not be useful anymore.
If your links are not crap, and you continue to rank well across a wide range of phrases, then it may not be your links, it could be something else, it could even not be about directly about your site.
But in general, if you have a dramatic loss of traffic and a dramatic loss of ranking across a wide range of keywords, and you know who ever did the needful on your links did a poor job, then it's probably your links.
[edited by: martinibuster at 12:58 pm (utc) on Jul 24, 2012]
|Absolutely, do not panic. Never panic. If you are certain that the links you built are crap, then you're probably starting to rank where you are supposed to rank should your crap links not be useful anymore. |
Yes, the site I am talking about has mostly crap links and as you wrote:
- a broader set of keywords is affected (pages that were not promoted via link building lost rankings, too);
- seems like the entire domain lost part of its power.
So the question now is: shall one waste time on getting rid of the links or just work on high quality links?
If negative SEO doesn’t exist (Cutts), which I don’t believe, it comes down on proper link building. Google wish “natural” links, which means everybody can add your link. So why punish such? The website where your url is added is (mostly) the one thing you cannot control. The moment your url arrives on the website, it can be 100% according Google guidelines.. but then? Webmasters can add too many OBL, web owners can sell the site, a free site can change to a paid site (as mentioned above) etc. By the way, I question the word “link building” as natural links are not “build” -> they are just suddenly there according to Google.
What I think is best:
Try to remove as much crap links as possible, the links you can’t remove you can try to overrule with quality links. Read Google guidelines again (after all we have to dance in HIS show) and try to act in a way Google feels comfortable with (I do believe when your site has been given a manual treatment by Google, you are not out of his hands that fast and a long headache is waiting for you). <- this happens when you are in the top on most competitive keywords, the manual treatment I mean.
If negative SEO does exist:
Monitor your external links if you are in a very competitive business. It is too easy to throw a 1000 links for $10 for a competitor. I’ve personally experienced with website owners who swore never ever bought any link, we checked the backround of the site and found all cheap Chinese and Japanese bllsht on sites between adult and gambling etc. I would almost think these sites were specially made for it.
Maybe this subject is handled already but I am new here and could not help myself to reply.
|Is there any single piece of evidence that cleaning up helps? |
I don't see any direct evidence, of course this is part and parcel why Google uses so many variables and announces so many updates - to blur the line between changes and increase the difficulty of reverse engineering. Each subsequent update will likely be more complicated in nature making this process even more difficult.
I would say that cleaning up backlinks is a good idea. While doing it, be sure to take an aggregate and individual top 10 snap shot of exact match anchor text links, long tail seo anchors, brand and www anchors and do come comparative analysis.
Ensure your profile falls within those norms, to start.
|A lot of site owners are going to drive themselves crazy removing, or trying to remove 'bad' links. But after all that, their sites rankings won't improve. |
I agree, their rankings won't improve.
One of my side projects is a free link directory and I've had a rash of people writing me in the last month wanting me to remove them. I'll oblige them, but it is a real PITA to search through dozens of pages of several thousand links to find a single link and remove it. So I've started my own blacklist to bar those domains from getting into my directory again.
BTW, I still get around 30 submissions per day to the directory. It may be considered a "crappy site" but it takes a heck of a lot of work to maintain.
Also I guess the Googlebot must visit it a few times each day just to reassure itself that it's a "crappy site" and penalize every domain that's listed.
My strategy for dealing with crummy inbound links has one step.
Step #1 - Ignore them.
All the discussions above reassure one fact that link building is here to stay :) so is SEO :) and google is not so soon tweaking its patented page rank algo.
|They seem to be so concerned about link building activities :p |
If you have been to any discussion on SEO forum for link building. You will find the same discussions doing rounds.
Back to NOW
DON'T OVER DO IT. My simple mantra, just concentrate on generation of contextual links; proportionately giving more weightage to such type of links, of course from high quality sites. Thats it!
|My simple suggestion to google if at all they are manually checking sites then instead of concentrating majorly on links they should focus on usability, infographs and responsiveness of the website as a stronger parameter over others to rank. This way right set of sites would be served to google users |
- Lalit Kumar
| This 38 message thread spans 2 pages: 38 (  2 ) > > |