• http://www.v2interactive.net/ Josh


  • http://www.facebook.com/greg.jarboe Greg Jarboe

    I think this is a valid test. But it begs for more testing on terms that face more competition. And it also begs for more testing on a variety of press release distribution services.

  • http://searchengineland.com/ Danny Sullivan

    It’s not taken out of context. Back in November, there was renewed attention about press releases being used to drive garbage into Google search. Remember, the how fake acquisition story? You can go back and read here:


    In that article, I mentioned Matt’s post from 2005 talking about how press release links supposedly don’t count:


    “A legit press release can get you written up by reporters, or editors/sites may subsequently choose to link to your site. But the actual content of the press release itself doesn’t directly affect a site. For example … those hyperlinks don’t help … (in Google).”

    So that’s Matt saying press release links don’t count eight years ago. No qualification. No, “press release links from PRWeb don’t count but other smaller PR sies might be OK.” Just, links from press releases don’t count.

    That got repeated in 2011:


    “Someone noted that while paid links violated the search engine guidelines, you can pay a press release service to distribute your release to places such as Google News, so don’t those links count? Matt clarified that the links in the press releases themselves don’t count for PageRank value, but if a journalist reads the release and then writes about the site, any links in that news article will then count.”

    In December 2012, he reiterated that statement yet again:


    “I wouldn’t expect links from press release web sites to benefit your rankings, however.”

    Note the slight qualification now, to “I wouldn’t expect” rather than it’s just not going to help.

    That’s at least three references to press release links supposedly not working, the latest coming after renewed attention and concern about press releases in the context of how they potentially may pollute Google’s results.

    So no, there’s nothing out of context here. I’d see this press release go up on Friday announcing the test (look at that, a press release that caught the attention of the press), noted on Sunday that it actually worked to rank Matt’s own blog for a term, and that seemed pretty solid proof yet again that the “press releases don’t impact rankings” statement that Matt gave out recently, and in the past, clearly wasn’t the case.

    That’s the “transparency” of how this particular story came to be. As for the “transparency” of why it’s relevant to cover, that seems pretty self-evident. If Google tells you something doesn’t work, and yet you can demonstrably see that it does, that seems kind of noteworthy.

  • http://searchengineland.com/ Danny Sullivan

    When Matt says that links don’t help, I think it’s more than reasonable that most people assume this means anchor text in those links aren’t going to help another site rank. My response to you above covers this more.

  • http://www.facebook.com/adam.boulton.71 Adam Boulton

    Mr. Cutts claiming something doesn’t work when it actually does? If I had a dollar every time this occurred.

  • http://twitter.com/BWInfoDiva Sandy Malloy

    I really liked this comment — and I work for one of those press release wires.

  • http://twitter.com/BWInfoDiva Sandy Malloy

    You would know, Greg, as you have been involved in some very high quality research that I am still citing (thank you, thank you…!)

  • http://twitter.com/forefront1 Todd McDonald

    While I really like your creative example and thinking, I’m not sure if SEO is really about anything other than search :). More specifically, I don’t think SEO really goes beyond optimizing content for searches conducted through search engines.

    Again, fantastic idea though. Love hearing about people capitalizing on what is going to “drive” Google via peoples’ searches.

  • http://twitter.com/Aaron_Kocourek Aaron Kocourek

    Nope, this is something I have been thinking about doing over the past year but the time and effort it takes … I just don’t think with the extremely limited about of time I have that is makes sense to do. Maybe I should just outsource the actual post creation. Plus the fact that there are so many bloggers already out there sharing great information, not to mention the private groups and forums. Anyways we always have a large amount of tests running regularly and it is not uncommon to see results for tactics that are ‘dead’. It is all about how you use each tactic and if you are doing it right or wrong.

  • http://twitter.com/Aaron_Kocourek Aaron Kocourek

    Thanks Sandy and yes, Business Wire is one of the companies we use.

  • MrAlexMiranda

    Thanks Todd. At the end of the day it all ties in. It is more off page SEO. And yes being creative always helps.

  • RyanMJones

    I hate posting links here because it might be confused as spam, but I expanded upon my comment with better examples here. http://www.dotcult.com/rant-seo-tests-cutts-statements-the-algorithm

  • Doc Sheldon

    I wish that only those people that possess a decent level of reading comprehension were allowed to leave comments online…. my blood pressure is suffering from all the misquotes of Matt’s comments.

    He specifically referred to “… links from press release WEB SITES” (emphasis mine). Read it again, folks:
    Matt is a company man, obviously. It isn’t his job to tell us where all the potholes are. It doesn’t take a lot of analysis on the part of anyone in SEO to realize that his statements are generally very well thought out. In fact, what he DOESN’T say is often more meaningful than what he DOES say.

    But I’ve been following his words pretty closely for several years, and I have yet to ever find that he’s made a statement that was false. Mis-read? You bet? Incorrectly interpreted? Sure!
    Most often, though, what happens is precisely what has happened regarding this comment… a handful of people see imaginary boogeymen and holler it from the rooftops ’til a large number of people believe he said something different.
    “…links from press release web sites

  • https://www.alarmgrid.com/ Joshua Unseth

    It’s ridiculous not to think this is important. So many SEOs avoid pressers because of these announcements. But in industries like the security industry, if you analyze the backlinks of the vast majority of competition, they are ranking as a result of daily press releases. If it’s not their press releases, then I’m really not sure what else they are ranking from, since those are the only links they have and their co-citations are also all press releases. So, yes, this is an important subject for anyone who gives a shit about how to rank right now.

  • https://www.alarmgrid.com/ Joshua Unseth

    I see people ranking for difficult keywords all the time as a result of press releases. I see it especially in e-commerce, old-guard industries where the only links that are available are in pressers because they don’t “do seo”. And these are definitely keywords that you would consider competitive. But Stuart, really? You don’t understand how this is very scientific? He used a word that doesn’t appear on Matt Cutts page, but was really just a random string of letters, and suddenly matt cutts ranked for that random string of letters. That sounds like a very good experiment.

  • https://www.alarmgrid.com/ Joshua Unseth

    This is the real issue with press release links. I have a client whose previous SEO did nothing but press release links. Every month, if those links aren’t refreshed, they lose like 30% of their rankings. The strategy requires anyone who takes the account to issue irrelevant pressers in order to sustain their rankings. That said, it’s absolute evidence that press releases work, especially if they are released by the right distributor.

  • http://www.seo-theory.com/ Michael Martinez

    But it IS taken out of context because youre equating “links passing anchor text” with “links helping with rankings”. Search algorithms have to deal with the law of diminishing returns just like everything else in life. People go out and hammer links all over the place then visit the Google forums and complain that they’re not seeing any progress or they have “lost rankings”. Occasionally a Googler shows up and says, “Your links may not be working the way you expect them to.”

    To argue that a link passing anchor text means it should help your rankings is naive. Maybe the link will help rankings and maybe not. Nowhere did Matt Cutts ever say that links which pass anchor text have to help rankings. Furthermore, there are alternative accounts on the Web about what Matt actually said at SMX West 2011 — Search Engine Land’s own article (written by Vanessa Fox) merely presents a summation of Matt’s comments without providing the question he was asked (Debra Mastaler provided a better context on her LinkSpiel blog, for example).

    What really makes this whole situation annoying, however, is that you have spotlighted a clearly unscientific and unrevealing test from someone who is obviously just trying to draw attention to his business. I just expected better of you guys. I’m disappointed to say the least but we all have our bad days. This was one of yours.

  • http://www.seo-theory.com/ Michael Martinez

    That anyone would have thought this silly test was significant underscores just how far the SEO community has to go in learning to debunk the nonsense that is passed around the Web every day. Try searching on “Danny Sullivan” and “Daniel Tan” and you’ll find some very interesting content out there (so far as I know, Danny has nothing to do with it and no knowledge of it). This guy was the last person whose SEO test should have been singled out for attention on Search Engine Land.

  • https://www.alarmgrid.com/ Joshua Unseth

    Cutts makes a claim, someone tests it with a very simple test. What is unscientific about it? SEO isn’t a robust science. It’s not like it’s principles are extensible to the real world .The world we’re running experiments in is Google. Google’s search has rules. If someone tells you it operates in one way, and you test it, and it seems to operate in another, you have proven that the Google world operates differently that someone claimed. This isn’t science. But if you have a problem with the methodology, it’s better to describe which part of the methodology was problematic. You’re just flailing your arms and saying it was unscientific. Congratulations on having an opinion. I have plenty of those too.

  • http://twitter.com/MarkusBaumann Markus Baumann

    If anything, this test actually proves that all of those statements from Matt are correct!

    In the image, Matt’s blog listed at 7 of 8 results when searching that term, yesterday he was listed at position 20 of 21 results and today, 40 of 41 results. Additionally, it’s easy to find reasons that Matt’s site is listed above the one news mirror without suggesting that PR was passed from the press release.

    This test only proves that Google crawls the press releases and can index content based on them. It in no way shows that any of your current rankings will benefit from press releases.

    So, for SEO professionals that want to get pages indexed for terms that are unrelated to their content, this test proves that you can use press releases to accomplish that goal and you can expect to get placed at the end of the results. Since it would need to be a non-competitive term, there seems to be little-to-no use for this, other than misspellings and there are simpler ways to deal with those.

    Even going back to the 2005 Avatar Financial example from Matt, the linked terms were ones that Avatar should have already ranked for and I have to assume that their rank for those terms did not jump after the press release.

    If you want to disprove Matt’s statements, you’ll need an example where a page has been ranking at one position consistently for a few weeks and then, without a change in Google’s algorithms and without site changes, the rank jumps up after a press release. There are certainly more factors involved, but at least it would be relevant. The test above is not.

  • http://twitter.com/workingnomad Working Nomad

    Links are still everything with all the search engines, and Google is the one where they matter most!

  • http://www.facebook.com/profile.php?id=512945366 Jaimie Sirovich

    re: “links passing anchor text” with “links helping with rankings”

    I think you hit the nail on the head here. Google uses a mix of the boolean model and the probabilistic (I know it’s more than this, but let’s keep it simple) model.

    So just because it matches the document in boolean mode doesn’t mean it contributes to rank. If you point leasreepressmm at a document and they model that as a term in the document, but it’s flagged either as NO_RANK or BUPKIS_RANK, or put it in a UNTRUSTED_ANCHOR_TEXT document section it would do what Matt Cutts says while still returning the document.

    It would be similar, I guess, to taking a BM25F function and giving said document section a weight of 0 or 0.0000001.

    Most people who don’t know IR in depth won’t understand the subtleties, though. It could be totally possible that all those leasreepressmm links contribute a weight of 0, but this is very difficult to know because it’s just 1 factor baked in to or summed with 200 other factors.

  • http://www.facebook.com/profile.php?id=512945366 Jaimie Sirovich

    Danny Sullivan Rank != Recall. Google uses a mix of the boolean model and the probabilistic (I know it’s more than this, but let’s keep it simple) model.

    Just because it matches the document in boolean mode doesn’t mean it contributes to rank. If you point leasreepressmm at a document and they model that as a term in the document, but it’s flagged either as NO_RANK or BUPKIS_RANK, or put it in a UNTRUSTED_ANCHOR_TEXT document section it would do what Matt Cutts says while still returning the document.

    It would be similar, I guess, to taking a textbook BM25F function and giving said document section a weight of 0 or 0.0000001. So it will return the document, but that’s pretty much it.

  • http://www.facebook.com/profile.php?id=512945366 Jaimie Sirovich

    It’s not an opinion. He’s correct. You can’t assume that recall = rank. This doesn’t prove Matt Cutts lied because the fact that it returns the document does not mean it contributed anything to its ranking. It’s a subtle point, but it’s true.

  • http://www.facebook.com/profile.php?id=512945366 Jaimie Sirovich

    No. It proves it recalls the document. Then non-text factors boosted said page because that domain has huge authority to begin with. It never contributed to rank, or at least does not prove that it contributed to rank.

  • http://www.facebook.com/profile.php?id=512945366 Jaimie Sirovich

    Actually, tf*idf can’t what’s at play here. It’s a 1 term query for one thing. But you’re on the right track. It’s more likely they just throw untrusted anchor text like that in a section that contributes little to no weight in the final score.

    At least then, if it’s a very obscure term, it will recall the documents.

  • http://www.facebook.com/profile.php?id=512945366 Jaimie Sirovich

    You can read it that way yes. It wasn’t lost on many people. However, you would think Google would also try to cluster the documents from the various sites and credit the link from.

    1. Only 1 of them
    2. From the original one.


  • http://twitter.com/ReidBandremer Reid Bandremer

    The way I’ve always taken Cutts’ statements on Press Release links is this:

    1. I’ve never heard him specifically say no value is passed at all from a press release. (please correct me if he has – but the Google forum certainly doesn’t count as an explicit statement of no value being passed ever.)

    2. He is saying he doesn’t expect the links in the releases themselves to be much help.

    3. Even if Cutt’s said nothing on press release links, you shouldn’t expect the links to do much — under typical circumstances…, where:
    A. most of the sites that syndicate press release authomatically apply the nofollow attribute.
    B. any exact syndication of the press release is most likely considered duplicate content and would devalue the power of the link
    C. Most press release hosting is done on services (like prweb) that links out soo much and has soo many pages that the auhority able to be passed on by a link on a prweb (or similar) -hosted press release has to be very low.
    d. your target keyword is probably at least moderately competitive (you’d think so if it had enough value to consider justifying the cost/effort of a press release)..

    So the tests certainly prove some value is passed. However, it doesn’t prove that enough value would be passed for the release links themselves to be “expected” to make a noticeable rankings improvement under “typical” conditions for most of us.

  • https://www.alarmgrid.com/ Joshua Unseth

    Just to clarify, I’m not asserting that anyone lied, only that he is apparently incorrect. Or, at least it’s not as black and white as he made it seem.

  • johnycre

    Hey Joshua – Any suggestions on which press release services work best?

  • https://www.alarmgrid.com/ Joshua Unseth

    Depends on what you’re trying to release. I believe in a holistic approach to content. I think you need to create content on-site, build links the old-fashioned way, and good press releases are part of that. If you are issuing pressers that can affect the markets, I prefer globenewswire. Its a bit expensive, but it’s a very good wire. The other paid wires do what they do. Basically, most any reputable press wire is going to have a similar schedule of releases, getting them on pretty much the same lists of sites. But as has been discussed here, pressers are not hugely effective when it comes to trying to rank. Their benefit is in their getting picked up. If they get picked up and turned into actual articles, you’ll get some actual benefit from them.

    It’s like the early days of linking. Getting 1 great link has always been far better than getting a million bad links. Likewise, issue good, quality pressers. 1 release that gets picked up by a big paper is worth a ton more than a pointless release every day.

  • http://www.devdigital.com/ Dev Digital

    Good one Barry. Matt are you reading this post..!!

  • https://www.alarmgrid.com/ Joshua Unseth

    You can critique it anyway you want. Whether they are non-text factors or the links in the document itself, something contributed to it ranking for the keyword in those documents. I’m confused by the semantics your arguing. Whether the link itself contributes or not, it’s irrelevant. The understanding that press release links won’t contribute to rankings assumes that press releases themselves won’t contribute to rankings. In the comment you are replying to, I said that I am seeing tons of companies rank without incredible authority. Their entire link profile is 1) obviously purchase links, 2) press releases and 3) internal links. These are low PR low AR sites. They rank basically as a result of those 3 kinds of content building/link building, whatever you want to call it. It’s a nice assertion that in order to see gains like this your site has to be an authority site. My anecdotal evidence does not support that claim. I generally assume that when Cutts says something, that a strategy is nearing the end of its life. But he was saying content farms didn’t work for like 2 years before they did anything about content farms. Directories worked for years after he said directories didn’t work. Strategies definitely need to change as Google changes, but right now, pressers are not completely ineffective.

    You can argue why press releases still work, discussing boolean vs probabilistic factors used in Google’s algo. But it’s pretty hard to deny, press releases still do something. They may not contribute to page rank, and they may not contribute to long-term, healthy rankings, but they do do something, and it’s a different kind of something than people decided Matt Cutts says it does. Sometimes we don’t need to understand why it works, we just need to know that it works and how it acts.

  • RyanMJones

    right. i’m not saying tf-idf is the factor or the cause. just using it as an example of how the algorithm can change based on the resultset. It’s definitely the case that the result set is so small that factors are treated differently.

  • http://shazidamain.in/ Shazida Khatun

    Every thing is fair when it is occurred from Big G but when we do the same thing then must face AXE of matt cutts ……………………

  • Scot Small

    Interesting reading these comments. The only thing this proves is that no one really knows much about what causes rank to occur. Funny thing is that everyone thinks they absolutely know. It seems to me that this kind of pissing contest is for ego building only. And it is exactly what Google wants to have happen – confusion.

    What I know for sure is that not everything works all the time and that what I know this week will have to change if I am going to be around for the following week.

  • Hemanth Malli

    word “leasreepressmm” have no searches on Google adwords and this
    word Google result is only ‘8’. Here the main point is Matt Cutts blog having
    High ranking – ‘6’. For that reason why obviously his blog is at 7 position.

  • Doc Sheldon

    I don’t have any doubt that they could, Jaimie. But when evaluating whether an item is really newsworthy, they might well use a metric that print news editors have used for many years – who else sees it as newsworthy. Anybody can put up a press release, but it probably will gain more notice from a real news site.

    You raise an interesting question, though… if a PR is picked up by 5 or 6 verified news sites, with the links to the PR originator intact, will all those links be individually weighted the same as if only picked up by one of those news sites? I’m inclined to think yes. But as the Shadow said, “Who knows what evil lurks… “. ;)

  • Joanna McDonald

    On the time investment — definitely hear you. I’m sure I’m not the only one who’s interested in test results, so I hope you do choose to do that sometime in the future!

  • http://twitter.com/Koozai_John John Waghorn

    Interesting post, I personally think that there is plenty to be gained from using the press release. Although it’s debatable as to the strength passed on through any links, there are plenty of other benefits from utilising this form of content within your overall marketing strategy. I recently wrote a blog post on this to highlight some of the additional advantages beyond the link/ranking debate:


  • http://www.guukle.com/ lucas bowen

    Ok. 2 weeks down since my comment and thought I’d share a test status:

    1. ahref is now picking up 8 BL’s. 7 do-follow, 1 no-follow

    2.WMT is picking up 34 BL’s from 28 diff. domains!

    3. No SERP placement (>#50)

    all from just the mentioned PR. Nothing else

  • Johan Hedin

    That test may not be valid cause there is no competition probably for it so it would rank anyways once it’s spidered.The way Google handle competitive keywords with search volume may be very different from the ones without any…. If you could test a keywords that is competitive and even if it improves from 350 to 299 that show that it may work….