Sunday, June 1st, 2014...6:46 pm
What Happened to Press Release Sites?
One of the recent big news of the SEO world that some connect to the Panda 4.0 update has been the drop of many press release sites. Indeed, if we look even at something as generic as “press release” SERPs, this is what we see:
Wikipedia, Wikihow, news and images, Apple, Walt Disney, Airbus, U.S. Department of Education (I am looking at google.com), Hewlett-Packard, Berkeley. Out of the top 10 sites, there are only 2 actual press release sites rankign for their title keyword: pr.com and prlog.org.
Compare this to April SERPs for the same keyword as seen in SpyFu’s cache:
There used to be 5 press release sites out of the top 10: PRLog, PRWeb, PRNewswire, PR.com and 24-7PressRelease. I checked the organic visibility data for 5 large press release sites in two tools, SearchMetrics and SEMRush (just to be sure), and got this picture:
PRWeb – definite drop confirmed by both sources
PRNewswire – both sources also showing a drop
SBWire – same here
PR.com – a small improvement reported across both sources (indeed, that’s one of the two press release sites remaining in the top 10 for “press release” – infact by the look of it, it has even moved up from #6 to #4, unless it’s the effect of the difference in the IP/personalisation of the SERPs)
PRLog – a drop reported by both sources (this is the second site that survived in the top 10 since April but barely so – #9 currently vs April’s #3)
So what has actually happened and why?
Until lately, press release sites have been something akin to Wikipedia – many of them have been ranking for just about anything. Below is an example of keywords PRWeb used to rank for before its current drop:
This has been going on for quite a long while – Danny Sullivan wrote about the curious assortment of “press releases” posted on PRWeb back in 2012 and even then it was already old news.
As Google made it easier and less risky to rank content on high authority third party sites (read: parasite content), you would be likely to spot various press releases, Facebook pages or even Google Play entries ranking for all sorts of competitive keywords every now and then. Using press release sites in particular has got so mainstream that pitches like this one started making rounds:
Apparently, if you’re trying to make a press release on a press release site rank, you can just spam links at it like there’s no tomorrow and forget the threats of Penguin and manual penalties, right? With a link profile like this, can anything ever bring it down?
Just to be clear, that same PRWeb’s enormous authority did not happen overnight and only due to people linking to their own press releases. Even after all the parasite mayhem, the site’s home page still remains the most linked to URL on the whole domain. But to see those most willing to gain exposure at PRWeb’s expense, it would suffice to take a look at the top linked to pages in MajesticSEO:
Skin care, real estate, investment opportunities, etc. What is this site about again? Why would it still rank for “press release” if the majority of external link targets (and anchor texts of links pointing at them) suggest otherwise?
Make no mistake: nobody goes out and builds thousands of links manually, nor do they happen naturally. Majority of these links are nothing other than automated link spam. How bad is the automated link spam problem for PRWeb? To estimate approximately, let’s do a simple math:
Total links: 27,395,196
Links to the home page: 419,662
Suppose the home page links are the natural ones and the amount of automated link spam to the internal pages is no more than 10%. (27,395,196 – 419,662) * 10% = 2,697,553. And that was a very conservative estimate. Likely the percentage of link spam is greater than 10%. The sum percentage of “here”, “click here” and “read more” alone (typical generic anchor texts usually used to dilute the commercial keywords by just one automated link spam tool I am aware of with the default setting of 10%) is 1.85%:
Another example of the site topic shifting due to parasite content and link spam intended to rank it: looking at “competitors” suggested by SEMRush for PRLog, one would think it’s a site about loans, not press releases:
So apparently, backlink anchor texts are still a very, very strong factor answering the question, “What is this site about?” to Google. So strong that Google makes it a special point to penalise sites exploiting this strength/weakness in the algo too aggressively (this is what Penguin was all about). It looks like for whatever reason Google just cannot do without them, for years its algorithm had been built around links as votes and anchor texts as pointers of the topical relevance. For years SEOs have been told it’s bad taste linking to a site with “click here” – not just for usability reasons but apparently because it doesn’t tell the search engines what it is you’re linking to. Now we see the backlash of all that power.
But wait, there’s more to it!
It’s very much worth noting that 3 of the sites dropped – PRWeb, PRNewswire and SBWire – use the same rather nasty trick, namely place an iframe under the press release pulling in the press release submitter’s site. Like this:
Needless to say, this creates a duplicate content issue that not only harms the press release sites in question (which may or may not be an issue due to their enormous domain authority), but, what’s much worse, may cost rankings for the sites being iframed (as their domain authority is most likely much weaker).
Oh, and btw, neither PR.com nor PRLog use embedded site preview iframes. Coincidence?
Conclusion: Panda 4.0 did not target press release sites as a class. Panda or not, the press release sites that did drop have been guilty of high amounts of link spam in their backlink profiles, could have been targeted as part of the Payday Loans 2.0 update as a lot of their content did target those highly competitive SERPs (if only there actually WAS a Payday Loans 2.0 update as an algorithmic action – as we’re led to believe – and not a simple manual or semi-manual sporadic cleanup which is what it looks like 2 weeks after), and in addition to all of the above in many cases had duplicate content. It is my perception that the drop has been brought on not by any one but a combination of these factors, presence of each of them on a site clearly not helping with the others.
“But how do I manage my online reputation now?”
I hate to tell you, but if your online reputation management depended largely on press release sites you probably have been doing something wrong all the way. If the only positive content about your brand is the one you put out yourself via press release sites and nobody else has anything good to say about you, it’s probably time to sit down and rethink your whole business practices.
So have these PR sites fallen prey to a collective negative SEO effort?
It may be regarded like this, but I tend to call a thing negative SEO if it’s deliberate. What this case does show is that negative SEO is theoretically possible against any site, no matter how strong it is, but if the amount of effort it would take to bring a site down is more than a person contemplating a negative SEO attack is willing or able to contribute, chances are they won’t bother or will try something on a small scale, see it not working and give up (I talked about it here, slides 54-55). In this case, it worked because it was indeed a huge collective effort over the course of several years – but this surely wasn’t a deliberate effort intended to bring those sites down.
Making predictions in the ever changing SEO world is thankless job, but if there’s any pattern in the above case I’d keep my eye on Facebook and – gasp! – YouTube (well they did penalise their own Chrome site at one point didn’t they). Used as parasite content hosts for easy rankings? – check! Happen to have a bunch of duplicate content? – check! So this guess is as good as any other. I know of at least one more site with a rather “interesting” business model worth watching but don’t want to disclose it in public.