Video Transcription
"Howdy, SEOmoz fans, and welcome to a second beardless edition of Whiteboard Friday. I apologize. The beard should be back by next Whiteboard Friday, but this time we're gong to go clean shaven, which it was for Halloween and Movember. I hope we'll get some good donations for that.
I want to talk today about a prediction that I've got around not the complete death of anchor text, but that anchor text is actually diminishing as a signal and being replaced by something else, something Google has gotten very clever about. I think Bing is using it as well. That is co-citation, so co-citation of terms and phrases along with a brand or a link, and I'll show you what I mean.
First off, some examples that you can check out for yourself. One of very favorite things to do is to go find relatively competitive keyword phrases and find sites and brands that are ranking without appearing to target the keyword, meaning that keyword isn't in the title tag. It's barely even on the page. It doesn't seem to be something that they're going after. They're not getting much anchor text for it.
How do they rank well against folks who are doing all of these classic SEO best practices? Well, here are a few examples.
For the query "cell phone ratings," coming in at number four is a web page on ConsumerReports.com without the words "cell phone" or the word "ratings." Actually, I do think they have the word "rating," and they might have the word "phone." But it's in the text. It's not even in the title. Really remarkable that they're ranking so well for such a competitive query. I'll talk about why in a sec.
Number two, "manufacturing directory." Again, another very competitive phrase, and ThomasNet is ranking number three without mentioning any of these terms, without seeming to try and target that phrase at all.
Number three, "backlink analysis," where Open Site Explorer, SEOmoz's own tool ranks number two, and yet not in the title. It's not anywhere on the page. Neither of these words are anywhere on the page. In the snippet, Google is actually using some text from another article that they found that mentions backlink analysis.
So it's just fascinating to see these sorts of rankings, and you kind of have to question like, "Boy, there's a ton of people with a lot of good anchor text, with a ton of linking root domains who are all trying to go after that phrase. They put it in the title tag. How does Google know? How does Google know to associate these terms with these websites if the classic signals that we think about as SEOs aren't there?"
The answer, in my opinion, is co-citations. Let me show what I'm talking about.
You can see a lot of articles on the web that mention cell phone ratings and reviews and mention Consumer Reports. They don't necessarily link to this page. In fact, very few of them link to this page. But many of them will do exactly this. If you look at a text snippet on the page, it'll say, "Cell phones as rated by Consumer Reports." This doesn't even link. This is not a live link. It's not even pointing to their website or to that specific web page. But Google is noticing the association. They see the words "cell phone." They see the word "rated," and they see "Consumer Reports." They put two and two together and say, "You know what? It seems like lots of people on the Internet seems to think that Consumer Reports and cell phone ratings go together."Same thing happening here. Directory of Manufacturers from ThomasNet, it's not even linking to the manufacturing directory page. It's linking just to the home page of the website. But it's mentioning ThomasNet, and it says, "Directory of Manufacturers." So the words are in there, and Google is kind of going, "Oh, yeah, I see this association happening a lot. It's not directly in the anchor text, but you know what? I'm getting smarter."
Google is getting a lot smarter about this. A ton of articles that mention backlink analysis, how to look at backlinks, talk about Open Site Explorer. Some of them link to it. Some of them don't. But because Open Site Explorer is very commonly cited in addition to the keyword phrase "backlink analysis," you're seeing OSE do really well for that query term.
This, in my opinion, is one of the kind of future looking elements of how we're going to do SEO, brand association, having people write about us and do PR about our brands, associating those terms together so that very frequently when you see an authoritative, high quality source mention a keyword phrase, talk about a keyword phrase, they're mentioning your brand. They're linking to your site. They don't even necessarily have to link to exactly your page. This type of SEO is something that's not very practiced today, but it certainly should be on a lot of people's minds for the future.
I would urge you, anytime you see something ranking that doesn't have the classic SEO targeting types of things, the anchor text and on-page text and the title tag, you take a deep look and try and figure out whether co-citation is what's causing it to rank higher.
All right everyone. Look forward to seeing you again next week for another edition of Whiteboard Friday. Take care."
Just to point out, Open Site Explorer [edited] has the keywords "backlink analysis" mentioned as part of the ALT text for the logo.
I think you mean OpenSiteExplorer has it as ALT-text, dont you?
That would be what I meant. :-) Edited. That'll lurn me to post early in the morning. ;-)
Not Open Site Explorer but seomoz logo at the top of opensiteexplorer.org has alt text = Link Popularity Checker + Backlink Analysis Tool
On seomoz.org the logo has alt text = seomoz logo
Interesting use of anchor text.
You know what this means?... Everyone, stuff as many keywords into your logo alt text as possible... Definitely a winner.
None of this co-citation rubbish! Just stuff those keywords
:-) nice spot by the way.
Can I rank for more then one keyword in my logo? lol How many keywords would you recommend? is 200 ok
200 is okay. 404 is not. ;-)
Ha!
I didn't find any keyword ALT text for thomasnet.com. So what about that?
No, but the thomasnet.com site does have the keyword "manufacturing" several times in it's on page content and in it's inbound link anchors. It also has some high authority inbound links with the keyword "directory" in the anchor text, such as these:
https://www.epa.gov/epp/pubs/case/dod_pave.htm
https://sbplibrary.org/research/index.html
https://dir.yahoo.com/Business_and_Economy/Directories/Companies/
It's not much of a leap to think that Google could say "Well I know this site is related to the keyword manufacturing and I can see lots of people are referring to it as a directory in their anchor texts, so for a query for 'manufacturing directory' this page had better rank).
OK, debunk time folks. Sorry, Rand. I am a big supporter, as you know, and continue to be, but I think we jumped waaaay too far this time around... There are plausible reasons for each of these that have nothing to do with co-citation.
First, let's start with OSE.
The old title of Open Site Explorer was "Open Site Explorer: Link Popularity & Backlink Analysis Tool" - https://web.archive.org/web/20100213052000/https://www.opensiteexplorer.org/
The phrase "backlink analysis" is part of 200+ links from around 50 unique linking domains pointing to OSE (according to AHrefs). OSE doesn't pick them up because they just aren't very valuable links, but they can certainly influence Google when a site has such huge authority as OSE does. Between the anchor text links and the historical content of this page, I don't see any reason to be surprised that it ranks for backlink analysis.
Second, Consumer Reports and "Cell Phone Ratings"
The page I get is... https://www.consumerreports.org/cro/cell-phones-services.htm
This page has both links with highly related anchor text (ratings of cell smart phones), has the keywords on the page, and the domain as a whole has backlinks with the exact phrase ("cell phone ratings"). It has the word cell phone right in the URL, which makes up a good percentage of its anchor text as well. Those, combined with the general authority of the site seem sufficient to explain the rankings.
Finally, Manufacturing Directory - https://www.thomasnet.com
ThomasNet the exact anchor text "manufacturing directory" from at least 2 root linking domains. It has 49 root linking domains with anchor text that includes both the "manufacture" stemmed words and "directory". It has over 800 root linking domains with anchor text that includes the word "manufacturing" or the word "directory".
The funny thing is that I actually do think co-citation is likely used by Google. It just makes sense, but I don't think that these are very strong examples of it. Because each factor presents such a small portion of Google's algorithm, it is highly unlikely that we will ever stumble upon strong examples because no single factor could push a site over the edge in a clear, indisputable manner.
As an SEO who sees the work of a lot of SEO's then I can't really say that reducing the weight/metrics of anchor text would be a bad thing.
I try my heart out to get anchor text wherever possible, but it's usually the spammers who always win this battle. When the spammers win, good, honest people will turn to black hat SEO in order to remain competitive.
What makes co-citation any less difficult to manipulate? Generating content is easy as hell. For Google to stop that from being spammed, they would only trust content on trustworthy sites...... which is what they are already doing with links.
I know very little about co-citation, so I couldn't say.
I am just pointing out that Anchor Text as an SEO strategy is seeing a lot of abuse, and therefore I could see why Google would want to make changes.
That was my first thought. How would this be anymore trustworthy? I would think a lot of damage could be done by using co-citation as a strong factor.
its wish washy stuff. its always about what Rand "feels" and not what is really happening
Nice detective work Russ! I think all three of these probably have some input from co-citation based stuff (mentions with links, mentions with the brand name, queries that include the terms, etc), but interesting to know they're not purely ranking based on that - some classic stuff is likely included too.
Russ is spot on. Remember that there are million of SERP results directly tied with links and supporting anchor texts, and the very few examples that is mentioned here still have Anchor's Supporting them as Russ showed us.
Rand, I think your prediction could be true, but it is a long time (many years) until anchors are disregarded completely. It is and has been a major ranking factor for a long time, I doubt they will just ditch it over night. Will it drop in strength over time? Definitely.
Great WBF. At face value, I agree that co-occurrence is growing as a factor, especially with the growth of named entities, brands, and semantic relationships between words.
I'd like to see some data behind this, across a larger data set. It got me interested, so I spent some time digging around.
For the examples given, I think this is mostly old school anchor text (with a bit of co-occurrence thrown in). This includes anchor text that includes phrase and broad match targeting, as well as targeting semantically related terms.
I think some data is harder to see because of redirects.
ThomasNet has thomasnet.com, thomasregister.com, and thomasregional.com. Between those domains, there is a lot of broad and phrase match anchors for manufacturing and directory. Their older links for the register and regional domains includes a lot of old school link building tactics, including directory submissions, link pages, and what appears to be a few text link ads (paid links).
For Open Site Explorer, I think there is some redirect stuff there as well. The URL https://www.seomoz.org/backlink-analysis (a product that shares its name with the keyword) redirects to OpenSiteExplorer.org. The top links to that URL are phrases like "Backlink Anchor Text Analysis", "Anchor Text Analysis", "backlink analysis tool" and "Backlink Analyzer". It has links from sites like Bruce Clay, Outspoken Media,SEO Chicks, Business Insider and SingleGrain. I think that redirect, and the Linkscape redirect, are playing a role in OSE's rankings.
The cellphone keyword is a bit tricky. I think that site is having some problems with cannibalization and duplicate content. There are some pages that are decently targeted on-site and with internal anchors, but the page I'm getting (https://www.consumerreports.org/cro/electronics-computers/tvs-services/tvs/lcd-and-plasma-tvs-ratings/ratings-overview.htm) is a duplicate of a page that should/could rank (https://www.consumerreports.org/cro/electronics-computers/phones-mobile-devices/cell-phones-services/cell-phone-ratings/ratings-overview.htm). The TV ratings page has a higher PA than the duplicate cellphones ratings page.
After spending time looking at these rankings, it made me feel the opposite about anchor text. It showed data that phrase and broad match anchors, even through redirects, are still highly relevant and can drive rankings, even when there is no, or poor, on-site targeting.
What it did make me conclude is that on-site mentions of keywords are diminishing in value.
But I still agree that co-occurrence is playing a role and is increasing in importance, and that SEOs should align their brand with keywords as co-citations.
Great detective work Justin, really helpful.
+1 Nice sleuthing sir :-)
Yes, yes, a thousand times yes.
I think the big difference is that there is still some anchor text including "backlink analysis" and "backlink analysis tool" but they're coming from very few sites (a dozen each) compared to your leading phrases, which are mostly brand-focused and come from a lot more sites (2k+ in the most popular instance). Now if it were the other way around, you'd have an issue with Penguin...
At the end of the day, it's common sense. It's not impossible that someone will link to OSE with "backlink analysis" as the anchor text, but it's a lot less likely. If we say hypothetically that it's 1% of people, then it should be 1% of your anchor text, not 30%, 40% or 50% (as is the case with those who've gone too far). IMO exact match anchor text shouldn't be ruled out entirely per se, but it needs to be a part of a larger, much wider mix of anchor text including brand and 100% natural miscellany.
Hi Rand,
I've noticed that kind of "absurd" results (absurd in the sense they are not following the so classic rules of SEO).
Surely co-citation is at the base, but I personally consider that other elements are playing, for instance:
1) entity association (brand is a subset of entity), so if you are an entity which was able to stand as such in your niche (i.e.: Manufacturing), then Google will start associate almost naturally to all the queries related to manufacturing.
2) an improved "semantic" and contextual analysis of pages by Google, so factors as proximity (brand + keyword) and LDA (maybe) are playing a bigger role than in the past also toward the Entities cited
3) search volume or queries' volume. I consider that if people is searching a lot "ThomasNet manufacturing directory", then Google will start to present the ThomasNet site also for the simple "manufacturing directory" search phrase.
What do you think?
Ciao
Your last point is very interesting Gianluca. It did not even pop into my head until you brought it up, but it seems very plausible that a ton of searches for brand + keywords could trigger Google to think about those two things as being related. This could also be very closely tied in with the autocomplete function and how Google is able to take keywords or brands and predict what you may be searching for in relation to those words.
Always appreciate your insight there sir
Totally agree - when I say "co-citation" I just mean any form of association of term+brand or term+website or term+URL. All the ones you mentioned are ways that this could be accomplished (and I probably should have included them in the video).
if copy/paste by email sharing, aka dark social, as Addthis stats have time and again shown, is co-citation analysis of gmail content as a viable search signal, in the Goo works ?
This stuff's pretty easy to measure. Try using xpath to search all links containing a keyword NOT contained in the a element.
This would be a useful starting point:
https://stackoverflow.com/questions/5074469/xpath-find-text-in-any-text-node
Similar method here: https://seogadget.com/how-to-find-lost-links/
Looking forward to the raft of YOUmoz posts this will inspire! Good call Rand.
Yeah - only point I'd add is that, as Gianluca noted below, co-citation probably comes from many more sources than just raw mentions near links. It's when brand names are included, when links to the domain are, when citations happen in search queries (e.g. "cell phone reviews consumer reports"), etc.
Maybe in the ranking factors next year, we could try to include some simplistic mention-based co-citation data. That would be fascinating to analyze!
Rand,
Did you forget that many brand names are keywords and often EMD'S? Or are you saying this only applies to frequently searched brands?
Good call Rand, I personally see no reason why anchor text should be viewed as a ranking factor any longer. Especially when it has been so abused. Google's understanding of semantics and context must be sophisticated enough now to understand the meaning theme of the content/page in which a link is placed.
Libertine, I totally agree with you. On one hand I feel like Google is going in the right direction but on the other hand, I feel like the SERPs are as messed up as ever. What do you think?
Certain SERPs are still messed up, but maybe that's because Google hasn't taken things far enough yet. If they turned the dial down to virtually zero on link anchor text then this would surely have the biggest negative impact on spammers and others that have tried to influence the rankings. I say Google should do it.
Also, the SERPs are getting better imo. The EMD update was a big shift in changing the SERPs for the better.
I do agree with you completely, I just wish that there wasn't any collateral damage. For example, so many companies rely on SEO agencies to get results. Not everybody is a thought leader so many agencies have been leading their clients off a cliff. It's unfortunate, but they weren't all spammers, they were just caught up with what worked up until Penguin.
So many completely deserving companies have heavy EMAT, not because they are spammers, it's because of how the industry was. I agree with you, if Google turned down that dial all the way, it would have been sheer madness.
Great points, Libertine!
So many completely deserving companies have heavy EMAT, not because they are spammers, it's because of how the industry was. I agree with you, if Google turned down that dial all the way, it would have been sheer madness.
Not sure if I'm misunderstanding you, but I'm saying that I believe that Google should turn the anchor text dial all the way down. I don't think it'll be madness. If agencies have been doing their job properly then it won't matter if a site has lots of exact match anchor text - different signals will have to kick in and provide greater weight, signals such as the theme of the site, the context of the content in which the link was placed etc. Good agencies will have been placing links in the right places anyway, so it won't matter if they've used too many exact match anchors because that's the method that used to work.
I agree with that, but despite EMAT being against Google, companies have been trusting SEO agencies to do things right. Even the most legitimate SEO companies were using EMAT in the day so I would just hate to see legitimate companies get dinged when they weren't doing anything *knowingly* wrong.
Regardless, there will be collateral damage with any update, but maybe going the route of no EMAT links is justified for the sake of advancing and deterring spammers?
I don't think that EMD update changed much. Check 'coach outlet' in Google, it's still spammed by those fake stores.
EMAT should die. Or the spammer will never stop their spams.
Jasonwelsh You are right. Need to change little something.
No, the SERPS are messed up because they continue to let big sites rank for irrelevant longtail queries. Who has the button to turn off ehow?
So instead of a Backlinkchecker you will need a "Co-Citation Checker" :-)
That could be a good idea for a tool...!
... Although you may be interested in Whitespark' Local Citation Finder!
(Wondering if that comment is actually co-citation... ;)
Whitespark is very reliable and a great tool to use.
It's possible there's a small team at Moz working on something like this right now :-) Even more possible that they're trying to roll something out by the end of the year!
A lot seems 'possible' in the Mozplex ;-)
Perhaps,
but anything is possible at Zombo Com.
Of course you guys are.Why am I not surprised. :)
Now i'm actually interested in the Seomoz pro.
I'm a bit late to the party on this one, but.
If you can't wait for the Moz team and have more money then sense (seriously, this next tool is expensive), you might want to consider using radian6.
You can monitor social media and mainstream news around your brand terms and certain keywords. It's a great way of seeing what your share of voice or reach might be, which in term might lead to any possible co-citations.
Thanks, Rand. This was the best WBF in a long time.
I think that how co-citation triggers domain queries is very relevant here...
For example when people read about the Consumer Reports cell phone ratings report then to straight to their toolbar and type in "cell phone ratings consumerreports.com" or "site:consumerreports.com cell phone ratings".
IMO finding ways to generate domain queries is the Rocket Fuel.
Highly actionable citations might be more powerful than a link??
Any of us would probably use a site search query, but I would highly doubt the general population would. I think this citation theory is fantastic and in some of Rands examples of the link not going to the target page, it makes onsite optimization even more critical so Google can serve up the right page.
IMO.
Totally agree EGOL - queries that include the terms are certainly going to be a part of this.
We all are waiting for the new tool "Co-citation" finder ;) Nice idea for a fresh start-up! ;)) Good luck!
+1 That would be an excellent tool!
We're working on it!
This is so interesting. I thought about this concept years ago and wondered if Google secretly did it, but no one else was talking about it and there was no information about it, so I kind of forgot about it. I didn't even know it had a name.
I've always thought it was taken into account; it must be a great way for Google et al to match relevant websites - and they have the technology through image searches.
Other than that reasoning, a quick Google search to find evidence of why I might have thought this only brings up things like keyword proximity and semantic closeness. Perhaps I thought keyword proximity was equivalent to 'co-citation' in my early SEO days and it's stuck in my head since, who knows...it's still interesting that this could become more important a factor than anchor text matches (if not already in some cases).
Great finding but what about Domain Authority and Domain trust. Sometime these two factors help as well to rank a page without same anchor text.
I think both DA and website theme will help Google to determine what’s the page all about and where should it rank. (Correct if I am wrong)
This. I think if you have domain authority, all sorts of other factors come into play which allow you to rank for terms you arn't even trying for.
But if you are a little website with no domain authority, no amount of citations will help - because in order to get domain authority in the first place you need backlinks...
Woo hoo! Great news for big brands. Google stomping on the little guy, though, as always.
Frustrating, right? The more Google is going this way the harder it's going to be for new brands to get started. It reminds me of the election with people were accusing the gov't of picking winners and losers.
I wouldn't say it's stomping on the little guy so much as making sure he gains traction on his own and not through SEO trickery. When I search for cell phone reviews, I'd much rather see ConsumerReports before John's Cell Shack blog. That's not to say that John's site should be buried on page 20 and there's always the chance that I find his site, love the reviews, and share/write about it on my sites, thus increasing his rankings.
A question for you Rand:Open Site Explorer is popular with the short form as well..that is, OSE. Have you seen anywhere where you feel that maybe a lot of folks did mention open site explorer, but mentioned OSE and Google awarded a good ranking for that keyword?If you can share insights it would be great.
Excellent point. That's a case where I suspect Google has seen "OSE" inferring reference to "Open Site Explorer" and thus connected the acronym with the name. Another interesting use of co-citation!
Good call dude :-)
"ose" is the anchor text coming from 100+ domains. There's also partial matches in the dozens: "check inlinks ose", "open site explorer ose", and so on. So it could be the case due to inbound links and their anchor text more than co-citations (or it could be a combination of the two...)
Search ranking just becomes more and more complex, I don't think it's a surprise though that Google would be using contextual clues and pages that link to it.
I agree it is getting more complex but it is also getting more interesting.
Okay another debunk! I love this video (and SEOmoz and Rand for that matter :). BUT the examples used are not strong.
I'll take up one example with OSE. Yes OSE doesn't have the on-page text "backlink analysis". But look at OSE's HTML (view source). You'll find "backlink analysis" in the alt tag of an image. In fact this is exactly where Google is getting OSE's snippet from - at least for the one I see in the SERPS.
Also a previous commenter (russirante) mentioned that the old title tag contained the keyword in it. I'm a firm believer that the title tag holds a lot of weight and I've seen this before, when I've changed a title tag the site still ranks for old keywords that were in the title tag. I guessing this is also a factor at play here.
Check out russvirante's comment for other debunk details (towards the top of the comments list).
Even with that all true, this is still a good post/video as it gets you thinking about how to shift gears toward co-citation. And I do think co-citation is being used by Google. But again there may be some better examples out there. Feel free to reply back to my comment with other good co-citation examples you've found.
Happy Friday all! Excellent WBF topic with great, illustrative examples. I imagine that this Co-citation effect may apply to blog comments that link back to a website. It's incredible how much SEO practice and Google algorithm has changed since the 2011 Search Ranking Factors report. I eagerly await the 2013 report from SEOmoz in the coming year.
Very intriguing! This would explain why I am ranking for keywords which I am not even trying to target. It does sound a little bit like an exploitable system though, because it certainly isn't hard to write a bunch of text-only comments around the web "mentioning" your brand.
So this could mean more webspam? humm...
Im wondering how long till seo's start doing co citations without links as the new link building (maybe they are already)
I thought of that too. People will now start kicking out hundreds of spun articles that have their keyword and company name in the same sentence--even easier than link building!. Although Rand did stress in the video that the citations should be from high authority sources, and I'd hope Google has learned its lesson by now, and wouldn't count citations from crap sites.
Since we work with a lot of firms, I feel like I can give some good insight to this question. Some are, most aren't. We're moving in the right direction though. Instead of exact match anchor text, so many SEO's are going the route of branded AT, URL AT, some keyword AT and a bunch of hybrid AT such as "Logitech wireless keyboards and mice" as opposed to just "wireless keyboards".
(Yes, I was searching the room for an example and saw my keyboard)
Interesting thoughts. It makes sense to me and seems like a good direction for Google to go. Hopefully Google will block against people abusing this, because I'm sure they will.
I totally agree with this post!
As for local SEO and Maps-Results in the Serps citations and co-citations are the anchor texts of the present and the future!
I found that shocking too!
Zeppelin, I've edited your comment about Rand's looks. We do ask that comments remain professional and TAGFEE concerning appearance.
If you're interested why Rand was beardless, we're raising money for men's health issues like prostate and testicular cancer through Movember.
Great post by Rand. But it really worries me as @clayburn says it is great for big brands but the small business's will suffer. The big brands are the ones that rank and also pay Google loads of money for Ad Words> The big brands also get all the the citations they need. How about SEO companies advertising (or not) and having citations? I'm not sure it will work. I wont even mention the cost of PPC for SEO companies and the quality score. Google does what it wants at the end of the day. They created this monster. Without links or highlighted text how will the bots be able to categorize? It will get very messy and once again the big brands will win.
I think this is a great discussion and very valuable, I am sure there will be more to come...Perhaps we should call this Rand Panda update #22?
Virginia
the semantic link?
Using this without anchor text as a signal would necessitate the brand being distinct enough to be credited with a citation. For instance, our brand is Domo, and it shares a name with lots of different things, both in English and Japanese. I think it would be difficult for Google to associate Domo with our business analytics software with so many other uses of the brand's name.
100% agree
my sentiments exactly! I cant see how it will work for small business
Rand,
I always interpreted 'co-citation' as more of:
If sites A and B link to sites C and D, sites C and D will be interpreted as having a relationship even though they do not link directly to each other.
In your video, it sounds like you're more referring to the words found around a link on a website (vs. the words IN the anchor text) as co-citation.
Maybe you are thinking about things from both perspectives and it's too early on Sunday morning for me to be watching SEO videos...
I would guess that depends on whether "Blue Corona" is already associated with the site you have in mind... if Google is trying to use all relevant signals, they'd still grant some credit to the fact that the post author includes a link, whether it's anchor texted or not.
The whole concept of co-citation makes so much sense, and is clearly a better signal than pure anchor text because it deepens the analysis of the content of the linking page. It's always been true that pages that are not relevant from a content standpoint don't offer as much help to the linked page as a page that is on the same topic as the keyword, so this just gives that whole concept more weight, and gives decreased importance to the less relevant "scattering" of links across the web.
All good things for people who care about getting credit for content.
It could be co-citation and it could be a simple association algo as well.
For example, one of my test websites(image niche) already about a year is found in bad neighborhood - porn niche ;) Needless to mention but there isn't any content related to that but users constantly find it by porn related keywords. And not only that - they searching/looking in to the site for porn as well.
So, i find this co-citation/association algo in place already for some time but newer formulate it clearly until present moment. Now I see - I'm not alone ;)
Regarding big and small brands/websites - Just try to think how Google 'could' work:
Google's task - get rid of the Black Hat SEO, optimize user experience, journey etc.
Google's action/execution/insight thoughts - 'we' need to find a test objects! Which one we will use - small or big!? Of course small as big almost always paying us and in case something bad will happen - small website will newer make serious problems for us - it's just a safe side. So - a few small is always better than one big!
Mainly this is the psychology that I could think of them to implement - all experiments innovations etc. firstly will hit small fish.....If big fish will not adapt to the new rules - it will die as well but a way later.
.............................................................................................................................................................
Algo for co-citation:
Personally I think It is simple algo's number's game!
Suggestion for open-minded big brands with reputable profile that are after rankings for generic keywords:
1. Buy 10 private/elite proxies,
2. Each day Google your desired key term plus your brand name about 10 times per each proxy,
3. After some time you must see exactly this term in Google Suggest,
4. After some time you will rank high for the same term without your brand name attached.
You will ask me - why this could work greatly only for big brands!? Because Google is taking seriously into consideration (in order them to change something) only brands with power! Brands that can influence the traffic and search as a psychology and brands that are changing a lot of things around the search - specially how the algo understand the word - BRAND!
Personally - I think they distinguish those guys by traffic volume and business tape. Ether it is e-com or blog - Brand should be highly visible and influential in any niche.
They filter them let's say - by best 10.000(by catalog/niche) - and after that they looking how their(Google's) changes will be affected on those websites. Of course - split/multivariate etc. testing is present there and that is one more - why!? Google implemented Hidden keywords in GA!
regards to all and thanks for the post Rand!
Jungles
Very interesting, this is similar to how they rank images. It's not only the alt text but the text that surrounds the picture as well.
Does this mean that adding tradition seo (both/and) would easily boost consumer reports to #1?
Nice post! I immediately think of what happens when you Goggle someone's name. Unless you're some crazy SEO, you're not using anchor text to try and show Google what should show up for your name. Somehow Google knows the appropriate pages to show up in the SERPs, using through what you've coined as, "co-citation."
I really hope this is becomes the staple in search results both as a website owner and a frequent searcher.
After watching WBF this came in to my mind ( New SEO = Great Product + Marketing )
I did a test search for "Search Engine" in Google https://bit.ly/101YPPJI
I cant find Google or Yahoo homepages in the top 10 results. I guess both have large number of co-citation for the same keyword
But Bing is there, and they only used the keyword in Meta Description
WHY?
Yeah, the search results for 'Search Engine' on Google are as poor as most other search results on Google after the gazillion of updates in 2012.
I couldn't agree more. Coincidentally, while reading this I was doing this search on Google:
"2001 ford f-350 7.3 turbo diesel" : https://goo.gl/XexMt
Results:
- 8 YouTube videos
- 2 organic results (1 of them sucks at all)
- 9 AdWords Ads
- 5 Google Shopping Ads
WTF? this is a worst user experience than Ask.com!
Rand, I'm afraid I don't see how this is really new.
I certainly can accept that Google is getting better at recognizing the ties between entities via non-linking citations, but the use of citations (not really sure why adding "co-" is significant) - is something that many of us have been implementing for a few years now.
Its real significance, in my opinion, is that it's evidence of continued progress on the front of their learning algorithms to improve the graph. But we've known for some time that much of their work is taking us closer to Semantic Web capability (painfully slowly, but hey... we take what we can get, right?).
Frankly, what has concerned me for some time is the potential for abuse, which will necessarily drive them to be very aggressive with their filtering. That can't help but slow the algos' learning processes. Maybe in the long run, that will be a good thing, but for now, I can't help but see it as yet another speed-bump, while creating another feeding-frenzy around the placement of crappy articles.
By the way, good on ya for your Movember support! ;-)
Hey Rand, What if the company name is a keyword? Very nice post, but my question is how does this citation will work if the company name is same as keyword. For example lets say keyword is Australian shows and company name is also Australian shoes!! How does this co-citation work in that case ???
Hi Rand,This was also mentioned in your former post about how to target/create the keywords and since then it was on my mind. But got stuck with regards to the implementation.. How can we just assume what Google would place at the top? How should we actually target such "Smart Google Keywords" within our page? Please do share your views :)Regards,Prajz
I was under the impression that co-citation had been happening (particularly for brands) for quite some time anyway, though perhaps the effect is being magnified more as time goes on?
Hei Rand, you look great without beard, no need to bring them back. I was looking for such a topic you just discussed. I also want to be a visionary. Anchor text link is losing its importance from few updates and i am also not concentrating in linking with keywords. I link with those words best described or indicates the page. I always tried to be natural. I am getting feedback from Google and it is positive.
I want to thank you for highlighting a topic like this. In recent past i also observed such websites ranking without any keyword in TITLE or description but never analyze the reasons behind it. I thought that the sites have a higher domain and page authority and have words or phrases similar to the keywords. But your post revealed the reason. From now on i will analyze my site with targeted keywords. Bye and waiting for something more exciting in next white board Friday.
This makes sense - time-earned authority should carry more weight than stand alone tactics.
That means all plumbers with the name Smith rank well because one is mentioned a lot in co-citations? How would Google know the difference between two companies with the same name?
Maybe I should become a plumber then...
Rand! Nice Shirt! Did you go home between last week's WBF and this week's episode? I hope so. Anyway, I was hoping maybe you just wore the shirt again this week because it won some sort of split test, lol. After all, as I mentioned in my comment last week, my 9-year-old son liked the plaid shirt much better than the khaki shirt you wore the week before.
...but I digress....Does this mean that there's now potential for "Rand Fishkin" to rank for "plaid shirt" even though I'm not linking to either one of your posts? :-)
I hope you know I am being facetious. This was, as usual, an awesome WBF. It was also very timely. My company happened to release a new product from Roland this week. We busted "you know what" to get videos, blog posts and product pages up before any of our competitors. We succeeded and ranked in the top spots pretty quickly. But now here come our competitors. We noticed Sweetwater.com ranking right away for the product, even though the page ranking made no mention of the new product. There was only a very minor mention of the Roland brand. We scratched our heads and wondered "How on earth are they ranking for this product when the product is never mentioned at all or even linked to from the page?" We did a little research, and you are right. We found them heavily associated with Roland via citations.
Great work. Thanks for all you do for the SEO community. We appreciate you.
I thought Movember was for growing your mustache or beard. I can't hardly recognize you Rand without the beard.
I like the idea of co-citation. It makes sense that this would factor into Google algorithm as they get more advanced. Great examples, as always. So, does this mean that black hat SEO is going to shift from anchor text spamming to co-citation spamming?
I know this will help brands that are mentioned a lot. My question is how subpages can/will be affected by this. It's easy for Google to associate the mention of a brand directly to that brand's own root domain.
However, if I am searching for a specific product that can only be found via 3-5 links into the site, and I am taken to the homepage, I might not find what I searched for (I think co-citation will have to be broadly applied).
This is a good transition overall for Google to implement, because it will generically rank brands higher based upon what other people are saying about them online. And it will in turn give the subpages better chances of being ranked due to a higher DA.
I'm curious to see if others have detected co-citation to affect subpages?
Another reason to pick a good brandable domain when starting new sites. The hits just keep on coming for exact match domains...
That was going through my mind too. Wonder if big domainers are sweating yet?
When you think about it, it makes a lot more sense for Googlebot to identify the text around a brand name or link instead of identifying the Anchor Text, though I am a young lad to many, I still can't remember when the last time i made an actual link, which Anchor Text, that was not intended as a linkbuilding like. As a consumer, I rearely see people using anchor text instead of just copy+pasting www.seomoz.org (the URL). In that sense, I think that the "oldschool" SEO tool-kit is about to go through a serious re-work, co-citation, being a part of the future (Also, this further empowers the relevency of "the death of link building"-Whiteboard Friday) - I couldn't agree more! Good call, Rand - and good call Google!
While I do agree with the idea that Google is starting to become much better at reading the context of links relative the words around them. But, I am not sure I am ready to say anchor text value is going away just yet. I think I agree with the sentiment of this WBF and most of the comments, that anchor text should be phased out more because it is a weird thing to count as a signal. Most of the natural editorial links that I see are branded or just the domain name. Most of the time when I see something that is obviously some keyword rich anchor text, I assume an SEO was the driving force behind that
Great (beardless) WBF as always Rand!
IMO, I think this co-citation idea is probably more prominent in major keyword SERP's to favor big brands. Smaller volume keywords might be more influenced by EMAT.
This still raises the question of what happens with co-citations if Google doesn't recognize your company name as a brand? Or, if you're a new website?
Just my 2 cents. I would love to see co-citation become bigger, but like all algorithm updates, there will be collateral damage on those who truly deserve to rank.
If you're interested why Rand was beardless, we're raising money for men's health issues like prostate and testicular cancer through Movember.
I came back to this post to check some stuff and I do see the text "Backlink Analysis" as an alt tag. Not sure if it was added after the video. As now I see OSE ranks no.1 for the term. Also upon looking at the history..I see this term was used on all important areas of a page.
Also if you look at the anchor text distribution for this page - https://www.seomoz.org/backlink-analysis/ which is 301 to OSE, you will find
26% - backlinkanalysis
17% backlink analysis tool
and rest were also centered towards the phrase "backlink analysis"
So for OSE, this makes perfect sense to rank for that phrase.
I am not denying this fact and I do think we will see co-occurence playing an important role..but may be we need to test more and need a better case study...
Having content has always been the food for search engines, but having informative content is the main thing and content syndication is now more often used strategy. So if you are targeting content for your ranking strategy then you must use your brand name with keywords sensibly.
Title for example Consumer report listing:
<title>Top ${productoverview.singularName} Reviews | Best ${productoverview.singularName} – Consumer Reports</title>
view-source:https://www.consumerreports.org/cro/cell-phones-services.htm
Hmm...
Appears that is a template error at consumer reports.
Great topic this week, totally agree that this would make good sense, thank you for the examples, time to do some more digging on this one!
Wow, this is amazing. Anchoring may die, but content will still be king... even without links! Nice.
Thanks for such a breath of fresh air Rand; it made my friday morning!
To be honest, I've just gotten back to SEO, even though I've been an SEOmoz member for 2 years now (as an entrepreneur starting his IT company and the fact my background is principally technical, as a PHP developer, I had to put the SEO stream on pause), and I felt a bit lost with all of those updates, algorithms, tools (disavow links)...SEO landscape has changed SO much in the last 18 months!! I almost can't believe it!
I understand that co-citation is only a prediction, but still, it's something I can start working on with my clients! :) Thanks for that!
Btw, WBF quality also improved A LOT (even beardless)! :) Keep up the good work at Mozplex guys!
If you're interested why Rand was beardless, we're raising money for men's health issues like prostate and testicular cancer through Movember.
Yeah but I thought the point was to grow a mustache?
Co-citations make lots of sense ... Links are not used (created) by regular users.
I have a question .... Say I have a widget that is presented on thousands of pages on websites other then mine . Just below the widget it says: "Powered by Company name" !
Is this a co-citation ?
Also, "Company name" currently is a link to our website.
Would it be better to remove the link and make it something like that:
"Term+domainname.com"
Would that be the perfect co-citation, posted on thousands of pages where the widget is present ?
Of course the pages have some content and the are relevant to what we do.
Great Post. Learnt Some New Techniques and one the morething Looking Handsome Rand in Clean Shaving :) :)
Hi Rand! Thank you for enlightening us. I really like your topic. For sure many people will learn a lot from your post. Good anlysis about co-citation and I totally agree with your prediction. Anyway, Google is indeed changing everyday and Google is changing the SEO world.
just logged into say that
"O My god rand has shaved"
I made a google search for gillette shaver and the image came up was from seomoz :)
I can't tell you how many times I've linked to/referenced this article. This was so spot on. We all knew anchor text was dying after Penguin...but no one knew what was coming in to replace it. Now we do : )
Very curious to see how this develops but I have to agree with several of the comments above in regards to my concern for spammers tweaking this system. Other than that, I think this should prove much to be a very interesting development for ranking in the search engines.
anchor text devaluation is eminent prepare yourself, the least you could do is create appropriate variation like a good seo :)
Rand is the man! Predicted this months and months ago and is spot-on!
Yes it is more natural - but hyperlinks (for better of worse, white or black) make the web what it is - if co-citation becomes the 'in thing' then usability is the first victim. They (Google) should be smarter than that.
Really amazing what has been happening with SEO these days. Gone are the days of anchor text links. Great video.
I would not say that I don't know this, I am using the same for last 3 years and not only while doing off-page activities,
I do take care of it while doing On-Page as well,
Not new to me at all.......
While I never coined the term "co-citation", I have given this a lot of thought over the past few years. It seems that every iteration of Google algorithm changes, this becomes more and more apparent. While I don't know how concrete it is, as there is a severe lack of evidence...I think it definitely warrants further investigation. One has to think - why WOULDN'T Google implement something like this into their algorithm? It just makes sense.
I have analyzed and experienced this situation before after changing the links on exact anchor text. All the projects on which I applied this, ranked well in Google SERPs and I recommend to all the SEOs of the world to please apply this. It is a clear path which indicates Google that we are linking purely 100% natural. It is a great technique of link building according to the smartness of Google.
This seems to go against a previous moz article about domains aimed at specific keywords. This may be a poor example of the top of my head but you should get what I mean.
Someone builds a website called "shoparoundforthebestdeal.com" and the title on every page is "Shop Around For The Best Deal > Computers" or something like that. Now how often do you think that phrase is used in articles? Pretty often i'd guess and it wouldn't be a link just as in your examples. Rather it would be phrases such as "When buying a computer it's good to shop around for the best deal" now by your reckoning google will think oh "Shop around for the best deal" often goes with "buying a computer" so now when someone searches for "buying a computer" our shop around site will get highly ranked without targeting or being linked to with the keywords "buying a computer".
If that is the case then there seems to be a whole new range of ways to target new keywords or key phrases
Awesome WBF, good sir! This should alleviate every writer's dilemma on targeting exact match terms. It also reminds me of a comment I made on a post here some time ago about the impact of authoritative links, mentions and shares on rankings. It has been somewhat still left clearly unanswered. To quote myself:
...I'm quite new to the field and am curious to know your stance on the relevance of the content of the linking page to the external page it is linking to.
Say you have Page A being a blog post talking about the top 5 running shoes the author (who's into running) recommends that links to a product page of a specific type of running shoes on your site.
Then there's Page B, also a blog post but from another site of a somewhat different/non-specific niche like a personal blog where the author shares about what happened to his day and at some point mentions that he bought this type of running shoes (the whole post was too broad, having no specific topic of focus), linking back to the same product page on your site.
Assuming both pages have comparable authority (maybe due to decent readership/following, etc.), and using the same branded anchor text on the links back to your site's page, which of Pages A or B do you think passes a better vote? Does the relevance of the content of a linking page to the external page it is linking to matters?...
The post author, Philip Petrescu, had no definitive answer but was kind enough to cite a post from the Google Search blog, and added that "the Penguin update seems to penalize sites with outgoing links that are completely unrelated to the actual content."
well, I just started to manage a website and there are already lot of changes of how Google works.I will definitely be visiting this website often. maybe on a daily basis since I just found it.
I noticed this about a year ago. The only citations that matched search queries were;
1. in the url
2. in the content
nothing overwhelming in back link anchors, inlinks, title, any tags and no on site back links. It was a subdomain injection into an existing negative link farms doing negative seo. the kw was in the sudirectory and in the (terribly spun) content ONCE.
I think Google's Knowledge Graph is at work these days.
4 years later and I'm still blown away by this video
Back links are never going to lose power. That would be like Google shooting themselves in the foot, the algorithm would be chaos. If you are a real SEO then you will already know how to rank sites quick and efficiently and not be bothered by penalty's. Period. Paragraph. End of story.
Anchor text is definitely on its way out. I have done some sample testing with just citations and just creating major social profiles with zero anchor text and Im ranking for some slightly competitive keywords.
It's my first time to hear this term "co-citation" . I even went to a dictionary to check what it means. However, I have noticed when I check rankings of my sites, several sites are ranking very well without having the target keywords in title, description or article body. Co-citation should come into our mind now when we build links.
Very good point, Rand!
I totally agree with you Rand. Co-citation may be be the future seo practices. I have practised with some of my clients before 3 months back. If I can't able to place my links in forums I have used saying these items were sold by them. And I saw those links in Google first page when I checked link:"xxx" I saw those post ranking rather than the links I have given through anchor text but it was not so effective at that time. But after seeing your video I guess we need to keep an eye on that one to make sure these things gonna work for our future practices.
I have a feeling I'm buried way to down the comments to get any exposure of this but I'll try.
Couldn't this feature be absolutely torn to bits by spammers? So instead of getting anchor-rich backlinks from websites and spamming 40,0000 blogs, and thus get easily penalised by Google, they could simply get hold of an EMD, spam the crap out of comments with and blogs and websites with brand rich copy and reap the rewards?
I don't think the anchor-text dial should be turned down too much. Isn't it a great honey trap at the minute to ensure that obvious spammers aren't getting away with things?
I hear ya Jason!
I think that's definitely a possibility, but what I imagine as well is that, in order to get "co-citation authority" (for lack of a better term), you're going to need those citations to come from quality, user-friendly sites. By that I mean, active blogs, mainstream news, public social channels and so on, that are relative to your niche. As a really simple example, Joe's Socks company might get co-citation credit from LaundretteDirectories.com, but not from best-e-cig-provider.biz.
That's the way I see it anyway.
May I say that the title of this post is a bit miss-leading. Anchor-text referes to the actual link (in the form href="bla-bla").
None of the examples Rand presents have to do with anchor text but actually only with the (target) keyword density, relevancy and proximity. I do agree that anchor text is getting weaker as a ranking signal, but nothing in this presentation has to do with it.
Could please provide some more detailed examples?
Thanks!
This Video is awesome.. I started working on this co-citation and just awaiting for the results.
How does this work on a very large site though?
Say I have a site about finance. I have expert authors writing in their own sections of the website for their specific area of expertise (e.g. specific types of mortgages, insurance, loans and investments). With absolutely tonnes of pages of info it won't be very helpful to the user to be sent to the homepage as they'll never find what they're looking for.
Will co-citations help a specific inner article about the future of interest only mortgages, for example, rank well for a query related to that?
"How does Google know?!" Rand, Google knows everything! :)
Interesting and informative post Stop Torture .
Thanks for helpful article. Me pleasure to read your thinks. Cheers
https://www.adoreinfotech.com/softphone.html
This is indeed intriguing. I've noted since the EMD update, probably only because I have been more careful in watching the movement, that a lot of sites that are targeting grammatically incorrect phrases are loosing rank. Sites that don't even have the phrase on page, in headers, or in description are gaining. It is kind of like the sites that once targeted misspellings. Those have been filtered out with google's spell check. This could be related.
It doesn't seem like the anchor text is going to die anytime soon. Just an example is for payday loans / auto loans / loans market in general. You'll see a lot of websites now(more than ever) spamming just one keyword and getting a good ranking. Yes, they may not stay there for long, but what they do is they have other back up sites that they just boost again once they get removed from the search. Either that or hacked websites is what you're going to see on the top right now. So, I don't really believe that it's going to die anytime soon. Just a thought.
I am only about three months late on watching this, but great wbf as usual. Rand, I appreciate you laying out the seo treasure map again. This is extremely useful and provides great insight on seo strategy should and could be changing.
It seems you are not the only one late watching this .. me too. Seo and google are getting smarter folks !!
Great video once again Rand. Came across this particular item many times over by now and never associated this in this particular way. You opened my eyes on this one.
Hi RAND,
I did some tests and I saw that after penguin update (5thOct) the results changed a lot in Spain, actually one of my eshops that used to rank in TOP3, now this keywords are in position +50...
I agree that the exact keyword on the anchor text, now is a problem, but to be practical, what is your suggestion for websites that have been doing this for years?
Do you think is better to use the disavov tool, and clean all the exact anchor text links?Start a new campaign of link building using the co-citation phrase?
When you suggest co-citation (Brand + Keyword + Url ), what happends for cases where the keyword is in the url? Example: SexShop www.sexshop.com
I would like to share this report that I found very interesting:
"The real estate industry varies in competition when you look at local factors, but in this case we’re just looking at it on a national level. The keyword real estate is competitive with an estimated CPC of $0.96 and approximately 1,920,000,000 search results.
Realtor.com (ranked #1)
RealEstate.com (ranked #2)
Zillow.com (ranked #3)
Complete report: https://www.searchenginejournal.com/anchor-text-analysis-post-penguin-emd/50484/
Looking forward to your suggestions...
Regards!
Very enthusiastic video - loved it!
If co-occurrences become a major player in SEO, PR firms will have a huge role to play as often the types of press traditional PR firms get clients may be great at brand awareness and may even contain keywords that could drive business, but rarely contain anchor text anywhere. Co-occurrence would solve this problem and make PR firms a huge asset to SEO firms. Or maybe the SEO firms will all do PR :-)
So you are saying that even when people just mention our brand on a relevant page with no link it will still help with rankings/ authority? Similar to how citations help rank local?
Interesting predication, and it makes sense. When I searched the examples I didn't get the same results and did see evidence of optimization, but either way, Google is definitely headed toward these associations, understanding related terms, etc.
Great Prediction! This is the first time hearing about this and yet learned something new! Thanks for the information Rand and I shall look out for these type of SEO.
Another challenge added to the already competitive SEO market.
Hi Rand,
Amazing stuff! I'm actually not really in SEO, but more of a blog writer. Still, I find it helpful to learn a bit about the related subjects.
I never thought of this co-occurrence phenomenon, but it makes sense that Google did. Here's my question:
One of my company's competitors has a brand name that fits very naturally with the product. (In other words, if you searched on Wikipedia for the product, you'd probably find the brand name in the article as an adjective describing the general product.) So, do you think the (presumably abundant) amount of natural co-occurrence of the brand/product would help this company or hurt i?
At first I though this must give them great free SEO. But now I'm wondering if Google will count the co-occurrence for this specific company or just see the adjective, not the brand.
I hope that this makes sense and I'm looking forward to hearing what you think.
Thanks,Effy
Thanks for a great WBF Rand. I have been noticing some absurd results on a couple of competitive keyword phrases for our customers, No keyword in title, backlink profile doesnt have any strong signals and very little in content yet the company is still ranking for the keyword phrase.... After doing a little scratching around, I think you may have just answered this riddle for me, thank you! Roll on co-citation-explorer!
Thanks Rand! I always leave with some knew acquired knowledge after reading your posts.
Very Interesting Article... Interesting informational articles like this are rare.i agree with your point of views. i think co-citation has started working in these days .
Great video Rand!
I Agree. Back links are still popular and important for bloggers and therefore internet marketers in general today.
Thanks for sharing
I Agree. Back links are still popular and important for bloggers and therefore internet marketers in general today. We'll see how whole thing plays out.
I'm waiting for a post about co-citation in a deeply way. Very good job and thanks for teach us.
Rand, why not ask Dr. Pete to do a experiment about co-citation, setting up three test sites with similar DA, DT and MozRank metrics, to see if your prediction can be confirmed.
Thanks,
Awesome foresight Randfish. It only makes sense. Backlink profiles have been a quantitative way of evaluating authority for some time, but with the evolution of Google and Bing's algorithms, it's only a matter of time before anchor text no longer plays a role in SEO.
Build your brand, get people talking about you and it will help with SEO. This makes total sense since it is a lot less flawed than anchor text which is been so abused. This is great news from a white hat perspective. However, it means that big brands will dominate rankings even more since most people talk about big brands...We'll see how that plays out though.
I think link building remains to be big so does anchor but it is all about trust and authority. For my little consulting company in San Diego (sefati consulting), I can get links with exact match of "san diego seo" all day from sites like NYtimes and CNN and won't hurt me a bit only help me. But if you get that from cheap sits then you do have a program.
And as for citations, nothing new. Maybe Google is putting more focus on them and if you ask me that is long overdue. And as social media, I think Google needs to step it up and heal the relationship between facebook an twitter cuz more people are and active there than their g+
For opensiteexplorer , there is a 301 from seomoz with high quality backlink with "backlink analysis" anchor text : https://www.seomoz.org/backlink-analysis/
anyway we are testing these strategy with my team for 2 week , we will see
Interesting prediction Rand... Are you saying you have 0 external links pointing to OSE with the anchor text "Backlink Analysis?"
This makes perfect sense in context with a video I saw for the first time yesterday. In April of this year (2012), GoogleWebmasterHelp posted this YouTube video with Matt Cutts.
It's a bit long for them, nearly 8 minutes, but pay close attention at about the 4 minute mark when they discuss indexing. There is no discussion of anchor text, just the co-occurrence of terms and authority of the site/page/domain. Obviously, the explanation is simplistic but it is pretty much this exact notion.
With the shift towards quality and away from practices that can be automated or manipulated this is completely in line with that. Not to mention the rel="author" and rel="publisher" attributes Google now advocates (to G+ pages) and data from Google Places etc. Other search engines aside, Google has an awful lot of context and content besides links.
Further, I know I have posted detailed comments on quality blogs/forums where I did not include a link for fear of it's potential ramifications even though I was interacting authentically...inspiring to think that may actually be rewarded.
What a perfect way for Google to rank whatever it wants and make it harder for people to determine why a site they favor is above theirs. Too bad small businesses prefer mentioning and linking to the big brands instead of the best but not as easy to find independent alternatives. Perhaps the growing awareness that Google is systematically handing the converting keyword phrases to their big brand buddies and intentionally putting online and local small businesses under might eventually make at least those whose brains still function wake up and start promoting each other. They need to stop competing and start collaborating. We must identify and get behind or build the best independent sites that will send you traffic or small business and freedom will only be found in history.
The other trend is that much more local business is getting exposure in Google SERP if the client/browser IP address is located in this local area. This change I found good. So a small business, if it has a strong local optimized focus, can be at top in SERPs over big brands. This balanced somewhat the overexposure of national or international big sites with brand.
Its true, Google's Matt Cutts stated himself that Google was going to be spending more time looking at the quality of your content, not your anchor text.
Thanks Rand,
but i think your ranking examples is because of the strength measured in authority, trust and mozrank (citation flow / pr). After this the anchor text comes in play. Nowadays it is better to avoid over-optimization, which of course the money keyword as part of an anchor text is part of it.
I am trying to test using instead of the target money keyword using synonymous terms.
It would be interest if someone of your staff, like Dr. Pete, could do there a experiment, to see, if this helps to improve the ranking :)
For such big sites like your example, it is much easier to rank. A google trend since a while, which have messed up the SERP reducing the diversity of the search results, where the same sites ranked at the first page, comes up in the following google pages again and again with other pages.
I know people, who think this must be an software error ;) If you want research something, the SERPs are an google nightmare.
Take care,
The first title of this WBF was too affirmative, it's a fact. BUT, we all know here that google is a little bit smarter every day. For me, the main information in this WBF is not about anchor links, dead or alive. No, it's about the fact that Google is or will soon give a part of his positive assumption to a website based on this co-citation system.
What makes me sure about that ? The fact that you've told in a comment that MOZPlex engineers are working on an analysis tool that will check main co-citations for each website. If SEOmoz can do this, Google can definitely do it too... It would be so stupid for him to leave this kind of information in a cupboard that I cannot imagine only one second those criteria being unused, now or soon...
What you gave us today is probably a chance of taking some kind of time advantage, in consideration of a necessary update of Google Algorithm (If it's not done already).
PS : Sorry, English is just a second langage for me, hope you get the point anyway :/ )
Excellent article. So you are saying if I write a post about SEO and mention seomoz without a link, text link, or anchor text link Google will now give SEOMOZ some serp weight in it's next algorithm?
If so will this make it harder for spammers to succeed with more spam? I am sure they will eventually find ways to make it work for them.
Wow! Excellent analysis Rand :)
Re. another website's description of OSE being used for that result.
Check out what Matt Cutts says on how Google determines which text to use (given, this video's over a year old BUT he does say that the decision on which text to use is on a query-by-query basis, still valid today)
https://www.youtube.com/watch?v=v6xxcQdAHbI&t=3m13s
Someone please build a tool allowing you to input a whole load of TLD links and find which market/niche/money terms are in the text of the page upon which each link sits. Then run a ranking report for those terms and see the level of correlation - it's those sites you wanna get a home page link from. And at the moment, home page links using brand name are far easier to secure than keyword rich deep links.
Hi Rand,
Thanks for sharing the findings. I personally don't think that the anchor text will be dying anytime soon but rather it is not the single most important factor to rank in SEO (which never is). With the co-citation, I look at it more like they complement each other.
Cheers!
I just can say WOW! Beardless Rand answered an important question. How to rank well in a word without even using it. Maybe it shows another use of guest blogging!
If you're interested why Rand was beardless, we're raising money for men's health issues like prostate and testicular cancer through Movember.
I think Rands got Movember all wrong.. It's "no shave" November. BTW I'm not buying it.. What really happened here is Rand grew out a stache and then unevenly trimmed it to the point of a reset.
Pssst... the WBF's don't actually get recorded on the Friday they air. He recorded that a few weeks ago, and now his mustache (and beard) are growing out quite nicely.
For once I am SMH on SEOmoz. You guys are all about marketing and timing is a very important aspect. You couldn't synchronize Movember with WBF. Who's in charge over there anyways? JK
Regardless of the stache its all for a good cause and I respect Rand and SEOmoz for raising awareness.
It's the thought that counts!
(Unique content + keywords + brand/domain mention) Makes sense to replace anchor texts and that frequency and the locations where these associations exist will be the deciding factor - if the algo does take this path it could generate a necessity for national/global online retailers to get their products on power houses like Amazon in order to see any non branded visibility due to the sheer volume of content with product genres within the Amazon. Not to mention the growing importance of social indications, another area where retailers with an extensive client base are basically out of reach for start ups.
Great post Rand, I also have noticed many sites ranking higher on Google without targeting keywords. So now I find the reason behind that. That's a strong signal now we need to target this instead of Anchor text distribution.
Doing local SEO in the sense of local citation, is very similar to what Rand is suggesting.Some of our local cite does not have a link, but mentions.Very cool observations.
Great spot Rand!
I did have a feeling that Google might take this approach. As I've seen comments on blog posts that help the post to rank for a keyword which may not appear in the title or even the post itself!...But does appear in the comment section.
That's why I get angry when admin's don't allow comments on their blog posts. "You're missing out on traffic god dammit!" :)
Hi Rand,
Your point of view is interesting, but in all of what you say seems to bring us to the LDA. Do you remember it? of course, yes. Let me explain why have to take a look back.
The LDA correlation taught us that web pages could rank better if we were able to build them with co-occurence (including TF*IDF)and that is a semantic way to think and writing a specific content.
Don't you think that the co-citation is a part of the semantic field that Google (much more smarter) uses to understand a specific piece of content?
Tks for this whiteboard
Calculating an LDA-based score for every keyword/phrase and every backlink is implausible. Co-citation could be as simple as taking the document, breaking it down into n-grams, determining those that occur in a statistically significant disproportion to the natural occurrence and then apply those as tags, so to speak, of linked to documents. Proximity to those links could play a part as well. Google probably already does something similar to this in weighting any page, and now they merely would apply those discovered attributes to the outbound links.
That's what I think. Proximity, Co-citation in a phrase that contains an outbound link is plausible in my opinion. And even the link in outbound has more power to pass link juice to the page the link point at.
I strongly believe that Google is going to be more semantic in the years to come.
I feel its more than just co citation. It's more like the experience and the type of analytical information Google is receiving from there own program(in this case Google Analytics) from millions of websites. User experience has always been left out until recently where people understand that you can't just impress Google with onsite seo and backlinking, but impress the user with gorgeous and simple design as well as providing relevant information to what they are searching for.
I think Google is taking a look at bounce rate and time on page for specific keywords and giving websites the opportunity with better user experience to pop up higher.
Makes a lot of sense Rand, thank you for the insights. I think structured microdata is going to really help fuel co-citation in the very near future if it already is not.
Structured micro data is working extremely well in searches related to recipes, product reviews and geo places. Structured micro data will be crucial in increasing serp click through rates too...I am pretty confident in future, even the organic listings will have a quality score similar to one in Adwords.
This explains a lot. A good friend of mine runs a company in Toronto, ON and performs very well on search with very little to no classic SEO practices. In fact, they have not once adjusted a title tag or diversified anchor text in their existence (just a few examples). I always wondered how they do so well. They have a decent ink portfolio which certainly helps but did not explain why they ranked well for terms that were never used in their SEO (which technically does not exist). They do not even partake in social media (until recently) even though they have a pretty inherently viral product.
However, said company is very active offline and is involved in a lot of community engagement. As a result of this, they have been mentioned (not linked to), countless times on the web from various sources with respect to certain things (query related things).
This sort of stems back to your Whiteboard from last week which talked a little about "offline SEO" and how things like community engagement can result in better SERP. It appears that they may be the beneficiaries of co-citation (albeit on a smaller scale than your own examples from this Whiteboard Friday).
Very insightful and thanks as always for the masterful work you and the SEOMoz team do.
Rand - you are right on.
Post like this - is why I belong to the MOZZ society!
Another reason why I happily pay me member fees every month.
Your pal,
Chenzo
mmmmm - lots of blogs comments with text only citation. that will generate some interesting results. A whole new era of blog comment spamming on the horizon. Anyone have any thoughts?
I think that the quality and the amount of the real content is pretty importent. This way a blog comment can not count too much... At least i hope so! :)
I've been seeing this trend of ranking for words we hadn't even optimized for, just because of where they appeared on a site's page. Its defiantly a reality now in some shape or form and I'm glad because it forces us to shift into more realistic patterns; creating a relevancy within the page rather spewing it all over the face of google with overused anchor text. As always, love the insight Rand, looking forward to seeing more about co-citations in the future from the community!
Really interesting Whiteboard Friday!It does look like Google is delving deeper into the searchers psyche and drawing deeper connections. This is a little scary actually and if they continue on this path at an exponential rate, it won't be long before they can accurately deliver the precise result we seek. OK, maybe I'm going off the deep end. Very valuable prediction and I look forward to seeing more evidence in favor of your hypothesis, as well as experiments from the community.
Hi Rand,So is this bad news for companies with generic company names or exact match domains? I will give you an example. I have and SEO company in South America. You use the word "Posicionamiento web" instead of "SEO" down here. My company is called exactly that "Posicionamiento Web". How would Google tell the difference? Tks CarlaBTW...
Carla, I've edited your comment about Rand's looks. We do ask that comments remain professional and TAGFEE concerning appearance.
If you're interested why Rand was beardless, we're raising money for men's health issues like prostate and testicular cancer through Movember.
Thanks Erica for letting me know :).....Sounds like a great idea
Excellent!
I think all of this would only make sense. The only downfall I see is for new websites or companies who's names are so vague that Google can't recognize them as a brand to give them credit.
If this goes through as one of the more prominent ranking signals, link building is still alive but it'll be more like brand building (without links). Personally, I would love to see this go through so that it's a step towards quality-focused SEO.
Great theory, Rand!
Really interesting, and more indicative of what a true editorial link looks like in the wild, right?
Reminds me of some latent semantic indexing discussions I've stumbled across (and only sort of understood) -- the idea that writing 'around' your head keywords is a powerful way to contextualize your content for the engines while simultaneously setting up to leverage the long tail.
So linking context is coming to the fore in more sophisticated ways than wut-words-do-the-label-say, maybe. Which has to be a good thing. I can drink coffee from a mug that says 'Kung Fu Rockstar', but that doesn't necessarily say anything about my martial arts skills.
I have been seeing this myself and with the industry I am in it helps ALOT. Great video Rand
So what does this mean for ecommerce?
IMO, it means that SEO is going the direction that legitimate SEO's have been touting all along. Build your brand and let the rest follow. Just, do it smart and don't rule out exact match anchor text and especially links just yet.
Very interesting... Like Richard, I'm also looking forward to more posts about this idea and also some good analysis.
Even way back, when I saw Google offering synonyms and phrases with very similar meanings or even some predicting user intentions, I figured that some type of connection is being made between keywords. The thing that the database is ever growing and search intentions are followed more closely only suggests that similar connections will be more and more dominant and after all associated with rankings.Many people are highly skeptical regarding this one, but you've made some pretty good claims with the examples (and there are many more out there, some of which always puzzled me up until now).
The brand association is what I consider the next step. Google aiming for brands being the ultimate signal of trust and authority is almost obvious now. With EMD's almost out of the picture, shady link building techniques being devalued and even punished as well as anchor text being probably next on the list, we can clearly predict the direction of movement. More trust with brands and authoritative sites, more stress on figuring user intentions by making associations between queries, social signals being more dominant than ever and still on the rise... Once again, what makes the most sense from the user point of view is ranking higher and higher.
I hope that these predictions will soon become reality Rand.
Oh, and such serious predictions require the presence of the beard, so grow it back for the next one.
Take care
Great post. It seems like this fixes a fundamental problem and question in search. That being, if you are the largest directory of manufacturers offline, should you be reasonably easy to find online by those terms even if you really don't work your SEO? So, good.
That said, will Thomas' rank for the other terms they would like to rank for from co-citation alone? I don't think Google will get there any day soon, but I have been wrong about Google numerous times. Great that this is happening though.
Agree Rand!
Currently at our organization that we are also employing co-citation strategy instead of rich keyword anchor text. Due to Google updates that we have changed strategy which comes with co-citation in order to rank in Google. I believe this will be honest SEO strategy which will be good as per Google guidelines.
One more time for WBF
This is a really good point and one I never even thought of Google doing. Thanks for the insight.
How does this work in regards to sub-directories of sites that target more specific keywords?
What a bold prediction! This does not follow the 'classic rules' of SEO but after your explanation, you have me thinking the same!! The co-citation or words around your company name could definitely play a larger role in ranking as anchor text has been used and abused. Maybe this is when the dreaded article marketing makes a comeback?
Thanks for the beardless Whiteboard Friday post
If you're interested why Rand was beardless, we're raising money for men's health issues like prostate and testicular cancer through Movember.
Really great content as always Rand! I really love the idea how Google determines ranking by seeing coherence with what people are saying that they like, and not what the title or anchortext lists.
I often see websites ranking pretty good for terms that they are not incorporating in the basic SEO elements. Well that is good for the goal of searchers intent, but annoying for SEO's (at least SEO's following common SEO best practices).
It is also very nice to see that the current way of ranking well is shifting away from self-made tweaks, towards true opinions made by real people!
This is really interesting, I shall certainly track it with interest... as I have been since 2006. Yes, this article is from the good-old/bad-old days when directories still rocked the SEO world...
https://www.searchenginejournal.com/co-citation-seo-linking/3199/
Sure this is not exactly the same thing, as it has obviously become a lot more advanced as you rightly point out, but it's a development of something that's been around for a while.
Isn't doing this kind of good practice post Penguin anyway?
Great video. G is making a good move on this. It will really help the end user. It takes most users a few queries to get to exactly what they are looking for in the first place because of all the crap that can come up due to the older method of linking (glad for the change)... and interested to see how this will effect the web in the mean time.
Also I must have missed it, but what was the Halloween costume that required a shave Rand?
https://www.seomoz.org/blog/moz-movember
That blog post isn't about Halloween, but it has a pic of Rand in his costume.
Thanks! lol love it!
Good analysis about co-citation, but do you think social media has in any way have any relation to these co-citation keyword rankings for websites that does not have the keyword on page or in any meta?
and this co-citation is going against for those who have links from blog commenting, paid links directories etc etc because on the same page there are different websites links like buying shoes links, cell phone links, and lets say cold drinks links.
So having relevant content on the page you getting links from is much much important.
It might not completely go against but ignore those links in future. You have to wait for this to occur for some more days, until that no one can predict anything :)
This makes sense in the grand scheme of things.
Running a quick query: "backlink analysis" + "Open Site Explorer" yields 103k results. I'm curious to see how many of these results have links back to OSE.org.
A move away from anchor text as anything more than a good/bad "neighborhood" checker makes some sense. If this is actually going to/already coming true, making content for users, not search engines is more important than ever. If it's shared enough, talked about enough, written about enough, active links to the actual page with exact anchor text lose their relevance.
There will always be ways to game Google, as we've seen with the recent posts by Dan Petrovic on Dejanseo.com. It's a cat and mouse game that some SEOs enjoy. I mean, who doesn't like finding new, novel tactics and strategies? The problem is that these tactics and strategies aren't sustainable. That's why I'm personally pumped for these kind of paradigm shifts. We now have to consider non-linked text. Context matters now more than ever if this indeed is true. PR and Reputation Management campaigns may now have a huge influence over rankings where anchor text once held a tight grasp. (Apologies for the dramatization)
On a sidenote, I'm glad to see that a lot of the stuff that Mike King and Wil Reynolds talk about, doing #RLS, is coming true. We're now much more than SEOs, we're becoming, or as Rand illustrates, we have have already become Inbound Marketers (apologies for all of the buzzwords). We now have to actively integrate ourselves not only into web design/web development and content writing, but also into branding, PR and reputation management, traditional media channels, etc.
I'm excited for the future.
What Rand is calling co-citation isn't exactly new from my experience, but it is definitely a stronger ranking factor than before - in comparison to anchor text. I assume this is from anchor text abuse from over-optimization and spammers. Whatever ranking factor determines the co-citation, I suspect that the authority of co-citation sources also play an important role.
Google is really changing the SEO world. The one million dollar question is where´s Google is going....
Better user experience ( while ruin analytics keyword data) and to take control over Google listing.
Thank you for your info Rand..
Its Google getting smarter Day by Day...... it was a nice explanatory video regarding evolution of co-citation.
How about tons of blog commenting, where you "brand" yourself while talking SEO (or whatever you want to rank for)?
my guess is that it should work both-ways. But again, I hope Rand is right here, because for SEO's the death of anchor text is one of the most extreme statements! We should re-organize our entire tactic.
I started to see this on one of the projects we were working on at search eccentric where a reputation of a Name of the owner of a company was showing some results which didnt even had it name as he had left the company more than a year back. So result B was showing instead of Result A (where he is working now and the content states it).
When we looked deeper of the links we had for B we saw that a lot of text and titles and few links of the same name of our client being used for B. Were being used for this specific domain which made search results change to show page A.
I believe Google is taking these steps on SERPs and Rand prediction on this can become very true.
If this is true, can we not expect Wikipedia to drop quickly in results? I think I agree with your assessment but am having a hard time jumping on board entirely because I think it plays out differently in some areas. Wikipedia may be a bad example because they have content on just about everything, but, if you cite a credible reference, I often think the search engines reward you for that more than they do the site you referenced.
For those who troll the patents all the time looking for info about algo workings, any mentions there of cocitations as ranking factors? I think that would be a good indicator.
@Mark - At SES Chicago people were talking about this, but I spent part of this AM looking and couldn't find anything on Google Patent Search.
As mentions above by @VirginiaC and @clayburn this great for big brands but what about small business? How does the little guy compete when there are very few times they will be mentioned in this manner?
Great WBF. Google has been taking into consideration the text around the links for a while already. Nothing new here.
So basically it is about brand building, which is founded on genuine content as well, isn't it?
I've heard and seen people say, "Content is KING" - Which it should be, right?
Hi Rand,
This Beard-less look suits you :)
Coming back to the video, I do believe that this sort of seo we will see in coming days. I guess, Google has realized that Links, which used to be the most effective matrix of evaluating website's performance is dieing. Social Media, co-citation and everything which related to "being real" will be a key to success. Having said that, we must appreciate Google's machine language learning and AI which made this possible.
Thanks
Salik Khan
Thanks for the post, very interesting!
YES I agree with this Google now knows what is the suitable page to rank within a website so sometimes we can see the ranking for another page but not for pages that we are targeting.
Man, your passion is causing goose pimples ;-)
Rand, I cannot say if it's a great WBF because I have no idea if it's true or not. are google really letting off the historical Anchor text signal in favor of the Co-citation?
If it's real, then you are awesome for sharing this right in time to prepare ourselves for a whole new tactic for our clients.
Thanks again!
Yoav Vilner, i think Rand is saying the fact. I also watched many sites ranking good without keywords in their sites title and description. Co-citation is a common sense i think. If you are Google yourself, don't you think you will also give importance to that co-citation which looks more natural?
Rand, we experience some of this here, in Canada. Pitstop Media ranks on the first page for SEO consultants though that was never our focus. We've never built links with such anchor text and barely display the words on the site. That seemed a bit strange so we've looked into this and why it might happen. We don't know for sure yet what's going on, but might be related with the overall theme of the sites who link back to us.
could you give further details on your case study? whats the URL?
You've got "SEO" and "Search Engine Optimization" in your Title, H1, H2, body text and anchor text...
Another great WBF! Nothing to say between the video and the comments we know where this is going. Best thing about this is seeing you without a beard again, just weird!!
Rand your article very helpful for us. I have seen all these things happening on Google but I never noticed them. I was thinking about that why Google showing but now thanks to you, You cleared all my own doubts.
I wonder how they does it in Google. Do they expand the query [cellphone ratings] to [cellphone ratings | consumer reports], OR they add the phrase "cellphone ratings" into the keywords of the goal page during the indexation?
Hi Rand,
I am already using this method. Also I liked you with the beard more. :)
If you're interested why Rand was beardless, we're raising money for men's health issues like prostate and testicular cancer through Movember.
God Bless you guys. BTW, Erica, the link you gave is broken, you might wanna fix it.
Ugh. Don't know what happened. Fixed now. :)
Hey Slava, do you have any studies or anything you care to share? I would love to know how effective it's been for you.
Thank you for your question and interest. So far I just dropped links here and there, like for example this one. I have too many websites that I am responsible for and I am trying to organize them all now, so it's hard to tell if there are any evident results, at this moment. But, I really stopped focusing on anchor text itself about half a year ago, that's for sure.
100% in agreement with your Assessment here, I also have started to leave out keywords in some article titles, ad the phrases in to the text & link from Brand Names/other key words, when combined with standard methods the results are superb.
Dammit, there you go again - giving away something I have been doing for a while! But seriously, this is a much better explanation of how it works then I have ever been able to give. And when people start understanding just how well Google can figure out relevance without old-school SEO stuff like anchor text, we'll start seeing a real move away from webspam.
Spill it, from a person that has been doing it for a while, how is it working for you? This would be a good Moz post if you had the data.
You rock Rand! I have been wondering that for a little while now and you nailed it.
Thanks!
I believe this also has weight when it comes to search suggestions.
Rand you are awesome! I had this question in my mind for ages.. Answered today!!
Thanks Rand for this huge help! I see this happen a bit with the travel industry, not with the fat head terms but for 3 word keywords.. yes! I was not sure why but this can be one of the reason..i have to look more in to it.
Why I think your idea make sense is because it’s more human. I mean link to the website with text around it taking about what it is makes more sense to me as compare to a link in the author bio… I am not saying that author bio link is shit but at least its less valuable as compare to the link within the content!
Just wondering if SEOmoz can run some test on this to prove the point with the experiment no?
Please, don't use cursive for text formatting in the Video Transcription (it's very hard to read)
This isn't rocket science. I Google wants citation then give em citation !
Not Just from any website , but from High PR website / Blogs.
Google loves links from High PR website / Blogs.
Either you build High PR webpage and link your main website from it.
Its a real slow and painfully boring way............ or ...you can get
contextual links from High PR websites for faster ranking . I've been
using this for last 3 years for my clients with consistent improvements
in ranking.
One of my reliable sources is at https://tinyurl.com/cjoorob . Sign up
for free and get a Informative Report on what works and what does
not. If you are serious about SEO get the links and test it out.
You'll start seeing improvement in ranking right away within few days
or weeks depending on how competitive your niche is.
Rand you were really awesome, I am really surprised that how you manage to find out these kind of topics.Thanks a lot....Surely gona consider this aspect from now in all of my SEO projects..
Achor text already dead.. No More Linkbuilding, Try Relationship building..
By
Infant Raj
onwebtoday.com
What do you mean exactly with relationship building?
At the end it is also a link, right?
Google is goin in right direction . I have heard first time co-citiation
Good to hear about the co-citation, but in whole article you have mention only Text link dying. here people want to know more about co-citation with Example other wise same topic white Friday is not worth. Be specific and Clear about the post title. for Example i am also ranking on 3 Place on "seo ahmedabad " keyword in India without doing seo submission or other seo stuffs.
[link removed by editor]
I think you're right Mr. Fishkin!
- And I think you can put "Title tag" under the same umbrella. #usedtobecool..
someone please tell this to greencowseo.com - because right now anchor text is killin it for them. These white board Fridays are always off base and completely speculation.
This is one of those examples where I feel like Google says one thing, and does another. We're with you on this.
But ther is nothing new!