So yes recently there have been several “new” algorithm updates affecting some sites, and although I hate to be the type of person/writer/blogger to add more chatter to your RSS feeds or bookmarked pages, I figured I’d chime in a little bit on both sides of the equation.
First off is the EMD penalty, which isn’t really a penalty but merely Google trying to remove whatever weight EMDs carry as a ranking factor and trying to push webmasters more towards a branding perspective on things.
There’s no penalty for having an EMD! Poorly structured and designed EMDs will get hit though (and should).
Well-designed sites with good content AND EMDs can and DO rank well on their own, sometimes without needing any off-page SEO (link building) done to them, but this requires more of a build-it-and-they-will-come mindset and doesn’t always pay off for white-hatters and G snitches…
So what’s a poorly designed and structured website that’s susceptible to penalties?
Over-use of keywords, anchors, low-quality content, minimal content, weird non-user-friendly inter-site linking, etc.
I’m going to try and keep this post short, so suffice to say, if you know when you’re looking at an inferior websites, then it’s safe to safe that Google geniuses have an algo for that. If you have an EMD about EMD kws and devote time and energy towards developing reader value and focus on user experience, then your EMD is a good EMD and in a fair world it shouldn’t get whacked (shouldn’t…)
Google’s Link Disavow Tool- a mess they made and encourage…
If you’re unaware of this tool, it’s basically a mechanism by which webmasters (who don’t know any better?) can report inbound links TO their sites to Google as being “unsavory”, undesirable, unwanted, and for Google to NOT count these as links the webmasters want as part of their “credited” inbound links portfolio as seen by G.
A lot of bloggers and SEO types are saying that this is just another fishing expedition by Google to get link neighborhoods “outed” so they can then go ahead and discredit certain link types, methods, sources etc…
Is the disavow tool really a fishing expedition, A LA bad links messages in Webmaster accounts, or do they have a more altruistic reason for doing this?
Sometimes you have to look at things from a search engine’s perspective… With recent algorithm updates, Google made negative SEO even easier to do than before, thus making many people afraid of doing link building altogether and going Extreme White-hat with their link building or embracing the “no Seo is the new SEO” school of thought.
Since recent algorithm updates made negative SEO even easier to do, that means more negative SEO is being done, therefore the mess they created must have a counter-measure as seen by the Disavow Tool.
In a way, it’s brilliant planning and foresight on their behalf because it means that they rolled out algorithms that made a lot of people scared to game the search engines with link building, and part of those algo updates opened the door for the aforementioned negative SEO to become more commonplace.
When people who mistakenly or otherwise had bad links built to their sites, they REALLY don’t know why their rankings dropped so the safest thing to do is un-generate whatever links they couldn’t get rid of by contacting link-giving site owners and if that didn’t work, then just to just use the Disavow tool.
What to do?
1- Watch your Inbound link counts and track which ones are natural, which ones you self-generated and don’t disavow ANY of them
2- watch for link-bombing (negative SEO) done by others to your sites and disavow those only if you’re experiencing rankings drops
Final Tip that makes all this worth reading: Inbound link velocity should be commensurate with the amount of traffic your sites are getting, so ramp up accordingly… add in some PPC or other traffic sources to bring traffic to your sites UP to a level that seems natural with the amount of inbounds coming in/being generated… 
{ 8 comments… read them below or add one }
Great tips Dan.
The overall value of links (per link) is greater than ever. It just requires fewer links of a higher quality.
The data suggests that links are still THE most important factor in gaining high serps, and that’s not going to change any time soon.
As for the LDT. Bing had this a few months ago. I personally would only do it if I was link bombed, outside that I won’t touch it. I don’t want to give “Big G” any more insight into my SEO methods than I have to.
Hey Paul, nice to see you on here again…
Google has a link algorithm team and a content quality team, so they don’t really need this Link Disavow Tool (LDT)
Same thing goes for people that thought having a site in webmaster tools was BAD. Google knows where your links are and what your traffic is like, so having a site in webmaster tools isn’t going to cause bad things to happen based on merely having a site in there.
Some people claim this is a Recovery tool and maybe it is who knows
Paul…excellent point about the quality of links now being of utmost importance, and I agree 100%. In developing our new post-Penguin link-building system (MasterLinker), we did a LOT of testing to see what did and did not work.
We tested new sites, old sites, EMDs, branded domains, sites that were hit hard by Penguin, sites that escaped the wrath of Penguin, etc. We also tested sites that we felt were “punished” by G (outside of Penguin) for having bad links.
In summary, here’s what we found:
1. Gradually and consistently building High Quality Links (HQLs) over time (at least one month…preferably longer) had a HUGE impact on improving rankings. Just a handful (15-25) of HQLs moved sites in ALL of the above categories dramatically up in rankings VERY quickly. And continuing to build HQLs in a controlled manner continued to improve rankings.
NOTE: By HQL, I mean links from PR2-7 sites on inner pages with content that is relative to the linked site. Plus, the linking page only had 1-2 OBLs.
2. Link velocity is VERY important. We found that building HQLs gradually over the course of at least a month yielded the best results vs. adding all of the links within a couple days (makes sense). In general, HQL velocity should be gradual and increasing vs. all at once and then stopping.
NOTE: We had a couple sites completely drop out of SERPs when we added too many HQLs too fast. However, those sites came back anywhere from 7-16 days later and ranked much higher in SERPs than before.
3. Building Low Quality Links (LQLs) to support your HQLs is very important. By “low quality” I mean links from low PR sites, but the links are still on pages with content relative to your site. (Please note that by “low quality” I do not mean spammy links, which I discuss below.) While the LQLs don’t send a ton of link juice to your site, they give context and validity to your HQLs.
Not having LQLs to support your HQL growth appeared to be a red flag to G…at least in our testing. Fewer of our HQLs were indexed and it took G longer to index them when they were not supported with LQLs. Our guess is G figures if well-respected sites (i.e., sites with high PR) think enough of your site to link to it, then there should be plenty of other sites that feel the same way.
We found that building at least 10 LQLs for every new HQL worked well; however, we did not test for the optimum number of LQLs to support HQLs. So take the multiple of 10 with a grain of salt.
Building high volumes of LQLs without building HQLs definitely gave the sites a bump in the rankings, but the impact from hundreds of LQLs was far less than the impact from a handful of HQLs. So while HQLs are most important to peak rankings, LQLs are a critical component to maximizing your site’s rankings because they appear to magnify and accentuate the positive impact of the HQLs.
4. A steady flow of social signals from different social sites helps your rankings, but they don’t have a lot of staying power. We consistently built links from 25 different social sites every day, and Twitter, Facebook, LinkedIn garnered the most immediate attention (which is to be expected). However, G+ didn’t get as much attention (at least in our testing), which absolutely surprised us. In addition, getting links from multiple accounts on each social site worked best (which makes complete sense).
While the flow of social signals helped bump rankings to their peak levels, our sites dropped a bit in rankings (although not significantly) once we stopped building these social links. Also, when we started to build lots of social signals to sites after we built new HQL and LQL links, it seemed like the social signals encouraged G to give us credit for the links more quickly.
So while social signals won’t give your site huge movement in SERPs, they appear to play an important role in the speed at which your site rises in SERPs, and also in the peak position your site attains.
5. Varied anchor text in the links turned out to be very important. To test this, we used the same “keyword-rich” anchor text for all links for some sites, we used 100% naked/generic URLS as the anchors for other sites, and we mixed the two types in various ratios for other sites. We found that having keyword-specific anchor text for 35-50% of the links and generic/naked anchor text for 50-65% of the links to be most effective in our testing. I’m sure there are some optimal percentages here, but we didn’t focus on tweaking the anchor text ratios (that’s a future test we have planned). Therefore, take these ratios with a grain of salt because we didn’t isolate this in testing, and the results could have been impacted by other factors.
6. Building very high volumes (tens of thousands) of spammy links gave sporadic and short-lived results. Sometimes they helped sites jump in rankings (especially newer sites), but then the sites quickly dropped back down in rankings…sometimes even farther down than where they started.
We can see these sites dropping back to their pre link-building rankings as soon as G discounts these links, but the fact that they dropped even further tends to make us think that G applied some sort of penalty to these sites as a result of having these links. If that is the case, then Negative SEO is definitely valid. However, none of these sites had catastrophic drops in rankings as a result of using spammy links (at least not in our testing).
Our feeling is these links can give your site a quick boost (kind of like a 5-hour energy), but once the effect wears off your site will not continue to benefit from the links. I guess you could say these links have a shelf life…in our experience they “expired” in a couple weeks. Maybe continuously feeding your site a steady stream of these types of links will enable your site to keep its rankings, but we didn’t test this so we aren’t sure.
While these links did not kill any of our testing sites, we also didn’t build massive amounts of links to any of our testing sites over a prolonged period of time. I know this strategy used to work very well, but I am very curious to see how G will react in the post-Penguin era. Will it still work, or will Google now kill (i.e., de-index) a site that does this?
To satisfy my curiosity, I plan on doing this to at least one of our old sites that have a PR3 and some very solid rankings. I want to see exactly how hard G comes down on a site that uses huge quantities of spammy links, and I also want to see if we can get a site to recover if it gets killed. It should be interesting…and telling…to see how G reacts.
Anyway, sorry for the long post Dan. I just figured this would be good data to share with your readers, and I’d love to hear what other experiences people had or what testing they have done.
Best Regards,
Tom
Wow, nice share. This should be a guest post rather than a comment!
Did you try any of these methods on sites that had been hit REALLY hard by penguin or just on fresh/untouched sites?
Nice post/comment Tom, keep us updated with the testing
EMD is not really a penalty. Look at it this way; before the update because you had an exact match domain Google gave you 10 points, after the update Google has reduced that number to 3 points. You simply have lost the advantage you had of owning an exact match domain.
When you write a quality article on a white hat, important site, what’s the best way to promote it (get it ranked)?
Agreed, EMD is a leveling of a bonus which didn’t make a great deal of sense in the first place and just came to perpetuate the “First in – best dressed” attitude of Google where it was assumed that if you were quick enough to get an EMD, yoursite was likely the best vlaue for visitors. Since 80% of the EMD’s I visit are pretty ordinary (commercial EMD’s here – I understand everything is an EMD of itself) and EMD’s have been actively targeted by the micro niche community for several years, this has long since been a poor way to prioratise a sites value. It’s increasingly become the case that “EMD = 5 pages of randomly spun rubbish with an Amazon affiliate link in it” EMD snipering has been dead for almost 2 years, at least the old way of doing it.