In the podcast:
01:38 – Can this harm your rankings?
05:00 – Here’s the SEO experiment
06:06 – Is it a good idea?
09:22 – Some takeaways
10:56 – What is proximity?
11:48 – The test and result
13:22 – Big takeaways
14:02 – On image linking
19:10 – PR links – do they help?
23:59 – Recap and action steps
Powerful business resources await you inside SuperFastBusiness membership. Click HERE
James S.: James Schramko here. Welcome back to SuperFastBusiness.com. You are actually listening to part two of a three-part series called Counterintuitive SEO Experiments Tune-Up Series, where James Reynolds from SEOPartner.com and I have been looking at some of the tests that have been run regarding SEO results and seeing some of the surprising results.
If you haven’t listened to part one, please go and listen to part one before you listen to this because it’ll set you up nicely for where we’re at. In that episode, we looked at if click through rate affects rankings. We had a look at Mobilegeddon and how that impacted search results. And we were seeing if Google displays more rich answers in search results.
But in this episode, we’ve got a few more theories to test. And the first one is exact match anchor text and if it harms rankings. So welcome my guest, James Reynolds, to discuss this very topic.
James R.: Hey James, how are you doing?
James S.: Doing well. So, interesting feedback from our previous episode. A few myths exposed. I’m sure that there’s been some new mobile-ready websites and some better calls to action being put up on the internet of late.
James R.: Let’s hope so. Yeah, let’s hope so.
James S.: Alright. So exact match anchor text. Can it harm rankings? First off.
Exact match anchor text
James R.: What is it?
James S.: You know what I’m going to ask you, right?
James R.: What is it. OK. So exact match anchor text. I guess first, determine what anchor text is. Anchor text is the text, clickable text element that’s in a link that points to your website. An example, if you have a third-party website mentioning you within that content and there’s a hyperlink from that content through to your website, the actual text component of that is what’s called the anchor text.
James S.: Yes.
James R.: So an exact match anchor text is when you have, in the text itself, the exact keyword that you’re trying to rank for, let’s say pest control Sydney was the keyword that we were wanting to rank for. The actual text, the clickable text, which our website would say “pest control Sydney”. Does that make sense?
James S.: That makes perfect sense. I’ve heard it described before as it’s like a voting system back in the early days of SEO. It was pretty easy to manipulate the search results by literally getting yourself more votes. If you had 100 people on the internet pointing back to your website and they all said, “Sydney pest control,” then Google could easily figure out that’s what you’re all about. You must be a good Sydney pest controller because all these people say so with their links.
The evolution of that was there were a whole industry set up around, ‘Hang on, well let’s just put some more links out there with those exact words,’ and it became a little bit gameable and of course entire networks got detonated when that game stopped. So I’d be interested to see what it looks like in 2016 now.
James R.: Yeah. I guess just to add some context to that, it should absolutely be something that Google do look at. What is contained in that piece of text should give Google a clue as to what to find on the end of that link.
James S.: Well, it’s still one of the only things they can look at.
James R.: Exactly, yeah.
James S.: It’s a pretty major thing to get a profile for a website is, so what does the link profile look like? And it’s definitely going to be a factor.
James R.: Yeah, absolutely. But as you rightly pointed out, it is very easy to manipulate these things especially if you’re using automation, or scripts, or networks where you can sort of submit content and it just blasts it out en masse.
“Can there be too much of a good thing?”
James S.: Can there be too much of a good thing, one must think.
James R.: Well that’s pretty much was this test was looking to find out. I mean the general consensus has always been that it’s better not to have too many links all saying the exact same thing and the reason for that is it just doesn’t look natural if all of your links to your site all say Sydney pest control and nothing else. Well, have those links actually been created by real people endorsing your websites? It’s highly unlikely because generally, people would use related phrases. They would use things like “Click here,” or “Visit this website,” the brand name, etc. You’d expect to find a real varied mix of link types pointing to a site that is established and trustworthy.
So generally, everyone thought exact match anchor text is bad. Of course, SEOs, they should put this to the test. This is what Rand Fishkin tested. So there’s another test from Rand at Moz. Basically, what he did was he set up a couple of pages. He pointed 20 links at both pages. One had varied anchor text links and the other one just had links all saying the exact same keyword term.
Surprisingly, what happened was the exact match anchor text like completely obliterated the varied one and outperformed it. I think it went up 19 positions for the exact match and only went up five positions for the varied anchor text. So it goes completely against what was thought to be true within SEO, which was of course quite interesting.
James R.: And did it have any other types of links pointed to it? Like was that a very high ratio of the links coming to it, like 80% or 90% or 100%?
James R.: Yeah. In fact, I believe it was 100%.
James S.: Wow. You know, it sort of does go against what people would say things like, “Penguin was supposed to stop oversaturation of back links.” Would you say it’s a good idea to try that campaign on your money maker?
James R.: No. I think there’s some context that you need. I guess the first one would be Google Penguin. Google Penguin is the element of the algorithm that gets updated really infrequently, pretty much yearly at the moment, until it gets rolled into the main algorithm, and that’s the component that looks to weed out manipulative link building and find out who is actually going against Google’s terms and trying to hoodwink them into better results.
Because it only gets updated once per year, there is a chance that in this particular test example, Penguin wasn’t at play. Google have taken the signals from the anchor text link and they’ve returned that into higher rankings. But it’s very likely that when Penguin next gets rolled out, that test site would in fact be impacted. So yeah, absolutely, I don’t recommend people go and try it. You do want to have natural linking patterns because ultimately that’s going to give you the longest term results and give you a very good baseline from which to work.
James S.: Alright. So when you’re doing SEO service provision, you’re definitely paying attention to the ratio of links coming to a site that have the exact text. You are going to make sure that it helps Google understand what the page is about. But you do want to leave a little buffer of margin there because unfortunately with SEO, one of the great dangers is if you happen to take brinkmanship one step too far, you can be out of the rankings altogether at least until the next update.
In some cases, that could be a few years if you don’t know what caused it. In the worst case, it could be forever. If you’re lucky, you just get a little warning and you can submit it to be reinstated. The name of the game is to not push the envelope. Let the R&D team do that on test sites like Rand did. He does it on a test site that won’t matter if it blows up. In fact, when you are testing, the whole point is to see what you can blow up and what doesn’t blow up and then try and work out where the line is so that you don’t have to touch the force field with your own site.
James R.: Yeah, absolutely.
James S.: Well, that was interesting. So it’s a result that is surprising, but it’s not necessarily a prescription to go out and emulate it.
“No one ranking signal is finite.”
James R.: Absolutely not. With all these tests, you have to take them within context because no one ranking signal is finite, if you like it. It doesn’t sit in its own silo. There’s other factors at play, such as the other ranking factors and how they relate to this individual one. It’s very difficult to get a really, really accurate test result, and you always have to think about, ‘Well what else could be occurring here?’
James S.: There could be so many things. It could be the server it’s on, or the domain history of any, or the time that it’s put out, or the competitiveness of the term. And just because it works now it doesn’t mean it will work in the future. Certainly anyone scientifically based would want to see at least 30 or 100 test sites perform in the exact same way to get a little bit more confident around it.
“Just because it works now, it doesn’t mean it will work in the future.”
James R.: Yeah. So I think probably some takeaways here would be certainly what is contained within the link to your site will absolutely give Google a clue about what that content should be ranked for. Somewhere within your link profile, if you want to rank for that keyword, either if that’s an exact match link or a close to exact match link, that’s going to be helpful. But ensure that it’s part of a rounded link profile.
The way that you can influence that is with your own content, if you want to rank for a specific term, make sure that that post is optimized for that keyword. As an example, the 16 SEO experiments post, the term that I wanted to rank for with that was SEO experiments. It’s contained in the post title. So when people start referencing that post, they’re going to link to it through 16 SEO experiments or something similar. But they’ll do it in a natural and varied manner because they’re individual people creating it.
So there’s ways that you can influence that naturally. What you should absolutely not do is go out and blast out some content, all with the same exact match anchor text using blog networks or something similar because that is likely to get you into trouble.
James S.: Dicey. And just the fact that you are using a blog network spells danger.
James R.: Absolutely.
James S.: So if they get one, it becomes a magician’s handkerchief. They just keep dragging the whole thing and thankfully, that game ended quite many years ago. In fact, a lot of the people in the whole SEO industry are gone. People who were there six years ago are no longer there. It was just too hard and they moved on because they weren’t paying attention to this stuff.
Let’s talk about text surrounding a link impacting search rankings. I guess you call this proximity.
James R.: Yeah. You’d call this proximity, and it’s almost like an extension, if you like, of the exact match anchor.
James S.: It is, isn’t it? It’s not the words that are linked, but it’s the ones right next to them.
James R.: Exactly, yeah. So the belief is that Google are going to look at what surrounds that link to get context again about what the links to page is about. Does the text before or after the link affect the ranking of the target page? So if you have the keyword not in the link itself but next to the link, does that help Google determine what the page is about and does it ultimately affect rankings?
James S.: And how did the experiment go?
James R.: Well this one was run by Dan Petrovich. What he did was set up four variations. One was the keyword in the text that’s surrounding a raw link. So raw link would be like “http:” and then the website name. That’s actually the raw text link. The second variation he set up was the keyword surrounding the anchor text “click here”. So just generic anchor text. The third one was actually the exact match anchor text itself as we were just describing with Rand’s test. And then the fourth one was an image link. So an image on a third-party website that had the keywords that he wanted to test actually within the alt text or the alt tag of the image and then that image linked through to the target website.
What he was expecting to find was that the exact match anchor text would work best based also on the testing that Rand had done. The close proximity would probably perform second and the image results would come fourth. What happened was almost the reverse of that. So the image link actually performed best, the exact match anchor text was second, but the text links with the keywords surrounding the hyperlink actually didn’t get indexed or ranked at all. So this one was quite surprising to say the least.
James S.: So what does that actually mean for us? How does that tie back to the point we just made about exact match anchor text supposedly being better?
James R.: I guess the biggest takeaway here is that image links, at least in this test result, they work really, really well. So maybe in a moment, we can talk about how we might develop those. But I guess it just adds fuel to the fire that actually a varied profile is going to be best for you. So here, image links worked best. We’ve seen the exact match anchor text can work really well. You want a really varied and rounded link profile, links from many places, many different sources, many different types is ultimately going to work best for you. Doing what you can to create those is the obvious step.
James S.: So is it time to talk about the image link because they sound appealing?
James R.: Yeah. So image links can come in all sorts of different types. But ultimately, if you want to get images linking through to your website, the way you’re going to do that is to create linkable assets, image assets that people might want to place on their own site and when they do, reference your site for the placement.
Great examples would be things like what you do on your site, James, where you do sort of custom images for posts with quotes and stuff because those are beautifully designed and people like the content. I’m sure they want to share those or maybe even place them on their own site where relevant. Now if they’re doing the right thing, they should then link from that image through to your website. So certainly create linkable assets.
Infographics are another great example of that. Really share worthy content. And if you’ve mentioned your brand or the target URL on those infographics, when people start to place them on their own sites and share them, if they’re good marketers, they should be then linking back to your site from the image itself.
So create linkable assets. And then the second action step that relates to that would be to ensure that the keyword phrase that ultimately you want to rank for for that piece of content is included in the alt text of the image itself. And alt text is basically the descriptive tag that you give to an image to help a searcher better understand what that picture is actually about.
James S.: Yeah. Like you said, we use a lot of images. Firstly, our standard operating procedure is every post has an image. That’s not negotiable. Often, we’ve linked the image back to the post because we know the image will get ranked in the search results because we name our images properly and give them the right properties. So people might find the image but not the post. So they click on it, they get taken to the post. And the same for when we send it out in emails. They click on the image, it will go to the post. And lately, we’ve been really doing some great stuff with Pinterest, where people love to collect images and pin them. That has quite a good result. We built up 1.3 thousand followers since we started pinning stuff even from our own site, which is a nice way to leverage it.
But yeah, images are big. You look at things like Instagram, Facebook, how visual is it, it just makes sense. Often when you’re searching for things, you’re going to see mixed results popping up in the search page. And people do often click on the image tab to see what something looks like, and they can click on the image, and they can visit the site. So using images as a front door is such an amazing technique. So naming your images properly and always linking them.
I learned this also from when I did heatmaps on my site. All these spots, they just light up on any kind of thing that you would think might be clickable. Any kind of image is going to get clicks. So hyperlink your images, and you name it properly, and you can take advantage of internal linking, like linking between your own content and you can also crosslink between your site and other people’s sites.
James R.: There’s one thing I would add that’s also important, is to actually see who’s used your images by doing a google image search. And then approaching them if they haven’t then given you attribution for that picture. If you’re creating assets like you are James, frequently for all of your posts, it’s great to actually then do a search and find out where they’ve been used and then you can approach those website owners and say, “Hey, we really appreciate you sharing my image on your site. Would you care just to reference us, say image source by SuperFastBusiness.com,” and encourage them to link. I think that’s fair exchange for using your pictures.
James S.: It is a fair exchange. And another site is TinEye.com if you want to find out where your images are hiding. It’s much like the technique of finding out where you’re mentioned on other sites and just asking if they mind putting a link where they mention your name or your website.
James R.: You’ll find really high returns. It’s a little bit of effort to go and find those people and then email them. But I’m sure a member of your team could do that. But you’ll get high returns. I mean the majority of people that you contact will then put a link because as we say, it’s a fair exchange. That’s really going to help your search results. Those mentions and those inclusions of your images, etc. If they start linking through to your site it’s going to give you a really nice rounded link profile that really will affect your search rankings.
James S.: So on the topic of text surrounding a link, you’ve told me that it probably does a little bit.
James R.: Yeah, it does a little bit.
James S.: I would’ve thought it makes sense. I would’ve thought, if I’m Google, I’m going to have a look at what’s around that link to see if it all makes sense. If they’re talking about spaghetti, are they talking about tomato sauce, or grated parmesan cheese or pasta on that page? Or is it just someone’s injected a link in the middle of nowhere to scam me.
So I think that that does make sense, and it’s a good tip, which leads us to the next one, and the final one for this part of the series, press release links. Do they help rankings?
Do PR links help rankings?
James R.: Yeah. So this one I guess relates a little bit to web spam also and the way that this test came about was as a result of a statement that a guy called Matt Cutts from Google made. He’s the spam czar, if you like. He’s the one that is sort of heading up Google’s spam team, trying to find out what manipulative tactics people are using and then eradicate them. He’s really a spokesperson for that whole team.
He announced I think back in 2014 that essentially, the press release links do not work and it’s not a valid strategy for people to implement to try and gain more links to their site. And the context for this is if you put out a press release and you syndicate it through a press release service, that’s essentially you building links to yourself, right? If you’ve included a link in the press release, it gets distributed to several places. Those links point back to your website. Those are not really earned links. They’re essentially bought links for a want of a better description. So those types of links shouldn’t have really any effects or at least not anywhere near as much effect as unrelated third party, creating a bit of content, liking your website, endorsing it and then linking through to your website from their own content.
James S.: But what if it’s a newspaper that’s writing an article about Trump and then they were to quote something from a Trump-produced press release? That’s just getting a quote from the source, isn’t it?
James R.: Yeah. So I think there’s a distinction to make here. If it’s the actual press release that you have produced and syndicated, then that really shouldn’t have any SEO value to it. But if you’ve produced the press release and that’s got you attention, “We’re the leading newspaper,” and then they do an interview or they write a column and then they believe the story is so interesting that they then want to link through to your site and reference you then that should be a search influencing factor.
James S.: Well I’ll give you an example in my world. The world surfing league just purchased Kelly Slater’s wavepool. So there’s an article about it on all the news blogs actually. Some of them quote the official press release. So that’s sort of a hybrid, isn’t it? Where the press release has been issued and engineered and pushed out there. But then people pick up on it and report on it.
James R.: Yeah, exactly.
James S.: So what was the results of press release links? Do they help or not?
James R.: They do. The actual environment for this test was kind of funny and ironic actually because the guys that were running it produced a press release about this whole statement that Matt Cutts had made. In this press release, they used some sort of random made up keyword and then in the press release, they link to of all places, Matt Cutts’s blog. What happened was despite that fact that Matt Cutts had not got that keyword mentioned anywhere on his site, he was then, as a result of that press release, ranking number one in the search results for that keyword, and all it came from was one single press release.
So they do work. But again, giving some context, there’s probably far more valuable link sources that you should be focusing your attention on as opposed to trying to build links en masse with press release publication. I guess secondary to that as we just pointed out, you should be using press releases in your marketing, but you shouldn’t be using them as a link building strategy. They’re a way to get the message out and get attention that then could spin off on to all these other benefits that will ultimately affect your search ranking and your business as a whole.
James S.: So did they use the exact anchor text for this made up word?
James R.: They did. So there you go, a little bit more proof on that one I guess.
James S.: Backs up the old anchor text theory, hey?
James R.: Yup. Absolutely.
James S.: OK. Well I mean that’s how the whole anchor text thing started, from some campaign where they were linking to a president or something from memory, many, many years ago. That’s how I became aware of it. Now, it still works with made up words. I think that’s a great way to test what happens, is using something that’s never appeared before.
James R.: Yeah, definitely. That will give you a true and rich answer, it’s nowhere else, and it’s only this one influencing factor that could have reaped the results then that’s going to be pretty definitive.
James S.: I did that too. I actually invented a word that doesn’t appear anywhere else and I have the only results for it and then I started seeing who scrapes it and picks it up and manipulates it in a really fascinating way. That’s the sort of stuff an R&D team does for a full-time living, I used to observe. So it’s like running a trace dye through your body to see how it’s flowing through your organs for x-rays and stuff.
So that wraps up this part two of our three-part series. So far, we’ve been covering all sorts of counterintuitive SEO ideas so that you can tune up your website. What do we need to do now James for our own campaign when it comes to press releases?
James R.: Well I would encourage listeners to utilize press releases when they have something that is genuinely newsworthy and interesting, because the secondary benefits you get of that if you get wide exposure of your press release will be things like interview requests, and blog mentions, and news stories and stuff like that. But you have to have an interesting story that is going to yield those second tier results. Don’t just announce the opening of an envelope. Make sure it’s interesting, you’ve got a real valid point to say. And case in point, we still use press releases from time to time at SEOPartner.com. We did one last week for a client and within a day or so of it being released, they had several interview requests coming in and that was because it was a valid news story. So those secondary benefits will be useful. But yeah, as I say, don’t use it as your primary link-building strategy. It’s not going to be the best use of your time.
James S.: Right. So if you’ve got real news, then it’s something you could put in your armory.
James R.: Yeah.
James S.: Right. So in our next series, we’re going to be talking about negative SEO and about outbound links, that’s linking away from your site, and what happens when links are removed. So that’s what we’ve got to look forward to in the next episode. In the meantime, if you need some help with your SEO, head over to SEOPartner.com and see what James Reynolds’s team have for you. They do good work. We’ll catch up with you on the next episode.
James R.: Thanks James.
Grow your business faster with help from our community of entrepreneurs
Enjoy more episodes by subscribing on iTunes