This month’s explosive dossier is discussed. We discover another ranking factor and we go into discussing this further. This is the last test of the bunch and we recommend that you sit through all the tests. Aside from the second ranking factor that we have discovered, we also discuss part 2 of the mobile first index test and what you should do if you receive the notice from Google, if stop words in the url affect rank, if h1 in Javascript is read as a normal h1, and if Google indexes numerically sequenced URLs.
We also do a rerun of these two tests from two years ago using our improved methodology, checking on the strongest infographic sites, and if LSI still affects rank. We also discuss the best photosharing website since we found that infographic sites aren’t that good, plus another tool/program test, this month, testing Social Monkee links. Excellent tests this month to get your page up to the top and hopefully beat all your competitors to the top. Read on and watch on! 🙂
Tests this month
Test 214: Is Google’s Index Mobile First (Part 2)
Test 215: Do Stop Words in URLs Affect Rank?
Test 216: Which are the Strongest Infographic Sites?
Test 217: What is the Best Photo Sharing Website?
Test 218: Do You Need a Unique Page for a Canonical to Pass Juice?
Test 219: Do Social Monkee Links Positively Impact Rank?
Test 220: Is an H1 in JavaScript Treated as a Normal H1?
Test 221: Will Google Index Numerically Sequenced URLs?
Test 222: Does LSI Affect Rank?
Test 223: Do Keyword Matches or “Variations” Affect Rank?
July 2018 Dossier
Transcription Below:
Is Google’s Index Mobile First (Part 2)?
Last month, we looked at if just across the board, everything was in mobile-first index. Many people in the SEO community has assumed that the whole index was mobile-fist. A lot of people thought that because there was a huge Google Algorithm updates that have happened kind of subsequent to Google’s announcement about the mobile-first index and it really jacked-up a lot of things in the SEO world especially when you are trying to get pages indexed. A lot of people has assumed and they were saying that you might need to get some notice on Search Console, they were saying that the whole thing was mobile-first now and that was the whole update was.
Last month, we found out that that wasn’t true. We had sites that had not received notifications and we have found that Google as not indexing the mobile version of the page but actually the desktop version. Last month of course was the the last Thursday of the month, when we were talking about this.
I think the very next Tuesday, Google rolled out a whole bunch of notices and I would say, out of the sites that I was handling, about a third received notices. Prior to that, it was only 1/2% has received any notice of being a mobile-first index site. About a third of my sites got that notification and a bunch of test sites did which was great. That allowed me to be able to test it.
What I did was, one page was created and then using a WordPress plugin that allows mobile display, I put in a fake keyword. The idea is, if this test site that received the notification that it was not in the mobile first index, if it actually is, and this content that is only displayed in the mobile view, if the page can index for them, then we can see that that is true.
This was what fetch and render looked like and if you did notice, you can get into Search Console and you can actually fetch your page in its mobile view and in its mobile view, you can see the keyword and in the desktop view, the keyword does not exist. After the page indexed, there is the fake keyword and there is the site and the cool thing is it only appears in the mobile version. So that fake keyword was not in the desktop version.
Also, you can see that when you receive that notification from Google, that they are telling the truth – your site is in the mobile index. If you did not receive the notice, then your site is still in the desktop version of what is being displayed. The main point here though is if you received that notice, you should really go in and take a look at how Google sees your site in mobile version and make sure that you do not have anything missing like where a dev decided that they need to strip our your htags to make it look right, or this content was not important, or perhaps these images weren’t important and they do not show in your mobile version.
You should really take a look because they could be factors that you are relying on for SEO and something that you did just looking at the desktop version as you were editing and creating the page and something maybe funky in your mobile version. So, go into Search Console, see that Google can see and if you ever see the notice and you are in mobile-first index, that is what Google will be indexing and using for your ranking factors.
The page cache also showed the mobile version, at least it showed the keyword although you can see that the format was the desktop version which I found kind of interesting. So the format itself was the desktop version but the term that was only in the mobile view was there.
Do Stop Words in URLs Affect Rank?
A stop word in English are a, the, in, for, at, about, around, lots of articles, prepositions, those type of words. They are extremely common and many programming languages ignore them. If you search for something within a programming language, they will ignore stop words because they are essentially not important, they are not the main words. Stop words are often ignored. If you like to see a nice list of English stop words, Yoast has a nice list right there and you can check that out.
Where this comes into play with SEO is let’s say you like the term “best plumber in Dallas”, that is what you want to optimize your page for. You are going after that and this stop word “in”, you like that and you are optimizing for it. What should your URL be? Should it be .com/bestplumberindallas or .com/bestplumberdallas minus the stop word.
In my personal opinion and what I have always been thinking is I have always liked stop words, I always liked my full keyword in my url simply because that is what I am going after, the URL is a ranking factor, I should get my full phrase in there especially if that is something that I want. I want this in Dallas concept for what I am going after for this particular phrase so that is what I would put in.
What I decided to do with this one is because we are using English stop words, we need to use English in this test although it would be in absurd phrase. The absurd phrase that we have optimized for is crispy rhino for key and the idea is that the for is the stop word that we are going to be testing. Because we are using English and that creates a lot of variables, what I decided to do was I created the same page on ten different domains. Five of those domains have a URL with the for in it and five have a url without the for.
Set 1 has .com/crisy-rhino-for-key and set 2 has .com/crispy-rhino-key, so the stop word is missing. I thought this was a good way to do it and I wasn’t entirely sure how the results would come out but I was thinking that I could do some regression analysis.
For you science nerds, you know what that means, for others, regression analysis is the idea where you get a lot of different data points and you could put them in a graph and you can see if a pattern emerges and we have the straight line like this, this is linear regression analysis, this is what is done in real science. We have an advantage in that Google’s algorithm is really a closed system, there is not really anything outside the algorithm, obviously there are a lot of variables within the algorithm but there is not a lot outside of it.
In real life, if you were to do a science test, the earth is not a closed system, as a result, when you run a test in the physical world, you have to often do this sort of analysis where you have to plot your results and see if a pattern emerges in. People are like, how can you do tests on Google’s algorithm, we are actually in a good position because it is a closed system with a certain set of variables, nevertheless, when multiple variables are thrown in, this is what is done. It is regression analysis and this is how you do this and see if a pattern forms and that is how science is done.
If anybody asks what we are doing, this is one of the things that we could do. That is what I thought we could do in this particular situation. We will be doing some regression analysis, I’ve got ten different domains and they all go to those pages and there’s the different set of urls, I’ve put them in a chart and we have a nifty line that make for a fun graphic for this right up. Well, that wasn’t needed.
When I launched these things and I put them in, notice that, no stop word, no stop word, no stop word, no stop word, stop word, stop word, stop word, stop word and then stop word is down there. So we lost one domain and I do not know where that poor boy went but you can see that the no stop word urls were the definitive winner. We are throwing this in, obviously, there are other results that have happened like hey there is this rhino 3d printer and other things that come up for this particular search but I was pretty surprised to see that the no stop word url pretty much went across the board.
This is not something that I’ve implemented into the wild yet, this is just results within the last couple of weeks for what we are seeing but this is how I will be doing my urls moving forward. When I have a long tail phrase that contains stop words, I am going to remove them and I am going to do a url that does not have the stop words in it.
This blew my mind that it was so straight down across the board. In order to see all the results, you do have to click for the duplicates but this one right here, the no stop words, was the number one and has always been the number one, once the pages launched. So, that took that one away.
This is one of those things that I would not tell anybody about. You might talk to other friends, I am in SIA and they do these tests and here are some fun things that we talk about, you might say that you agree with this or don’t agree with this but I don’t care. On this particular one, I would just put this in your pocket and take it and use it and feel free to not share that with anybody.
Why did this happen like this?
It’s because computer programs ignore stop words and I think that is what is going on. In the situation where it comes to reading the url that I think the stop words are being ignored or that the non stop words version is preferred by search engines.
Which are the Strongest Infographic Sites?
Back in July 2016, which I can’t believe was two years ago, we looked at infographic sites and I can tell you that the test was a complete disaster and the reason is that infographic sites are awful, we had tons of problems getting them into the sites in the first place and once they ran, getting them to show in the sites, some sites had payment issues, etc. but it was something that I wanted to try again and one of the reasons is that we have seen in the last year that a lot of what was old was now new.
The speculation and a lot of these thoughts come from Bill Hartzer who is now doing the Thursday Chats. He has this great theory and I really think that it is true, that the guys who are running Google’s algorithm now, the engineers that are doing the code, they are not the original engineers. So they will come across a piece of code and it looks superfluous, it looks un necessary and they cannot understand what it might be there for and because they are trying to create cleaner, better code, they simply remove it. The code is cleaner.
Well, that piece of odd code was blocking some old school spam from 2005 and they just do not know that so when they remove it, those old school things work and we were seeing that in a lot of our test environments and in a lot of the things that we are doing – old school things are working again and I really like Bill’s theory for why that might happen, that there is a streamline in the code, they take out what was plugging the dam on some form of spam and now it is working again.
So, in this period of old as new and it was two years ago, I wanted to give these infographic sites a test. The reason is two-fold. One is, there are many people who build tiered backlinks and infographic sites can be a fantastic way to get a whole tier that you can use to fire links through. The other thing is that, if you are into guest posting and you are into creating cornerstone content or if you are into posting things into your sites to silo, infographics are still a fantastic thing.
If you can get people to share them, it is a great way to get links into your silo or however you are formatting your site and you can drive things through that. Getting some quick wins by posting your infographics to infographic sites would be a great thing, it would provide a mechanism to kind of jumpstart that viral fuel.
So, I dug up the infographic from 2016, that is what it looks like there and we gave it a go. I can tell you without a doubt that infographic sites are still a mitigated disaster. These were the ones we fired to, this is the sheet that I was working with when we were trying to get this done. Of the 18, we are able to submit to 4 of them. A bunch, you could see emails and no responses and that is like email your submission in, sort of thing and I never heard back from these guys and I did multiple emails to them.
One is $350 and I told them to pound sand, there is also an $89 one but I am not going to spend 89. I tried to spend this $19, $25, and $20 and I actually tried to pay these clowns and I could not even get that done so that was pretty insane. A couple of them, like this site, once you get into their submission page, it wasn’t active and this one had a recapture problem. It just never let you prove to this robot that you are not a robot.
Side note, if you were thinking that you would really like to make a directory site and you don’t know what to do, this is an opportunity for you to create some sort of directory site for infographics cause these all are absolutely terrible. You could probably make a couple of bucks and if that is your thing, you are welcome.
I ended up being able to do four submissions, only two of them cached and the one I think barely counts because I actually went with it because it is actually a BlogSpot – infographicsdirectory.blogspot.com – so this is basically a Web 2.0. It barely counts. The only real “infographic site” was this ucollectinfographics.com, that one cached and for the target keyword, it hit the index. I then took a quick look, well maybe there is some image results here because we’ve got those other ones, maybe that would be of value.
The blogger infographic got two times in, the rest of these images all come from that ucollect and none of them was the infographic which I thought was hilarious but it was associated to the fake keyword that we put into the infographic.
Infographic sites, not the best. Kind of taking it into an advancement, I thought, there are now a bunch of photo sharing sites and so we decided to run the same thing on the photo sharing sites and that is what we go on this next test.
What is the Best Photo Sharing Website?
So we have our handy dandy infographic and decided to give it a go to what we considered photo sharing sites. To be honest with you, this might be the latest and greatest infographic site. But, I’ve got your Flickr, imgur, deviantart, photobucket, shutterfly, 4shared, etc. I was able to post to all of these except for photo.net.
If you are doing this for link building purposes, to build tiers, that is the only one that I would really avoid. Not too many ended up indexing – the deviantart did, the efello which I never heard of before but it is a Canadian site, and ello.co, those all got in and you can see the results actually got multiple things. The deviantart, efello, and ello, that is what the pages look like. They are getting a lot of things in there.
If you are in the business of using this for tiered link building, these photo sharing sites are certainly the way to go. I would put the deviantart, and those other two that cached at the top of my list because you are going to get them into the serps and if you feel that is a good signal for driving links through, then there you go. But the rest of them all, I was able to create those pages on. I did not play too much with which ones I could and couldn’t manipulate their description to get a link into so there is that to play with but I think this is a start. If this is something that you are looking into doing for both of the reasons that I talked about on the infographics one.
Whether you are trying to spam or not spam, link to if you will, or if you are trying to get something to go viral and you just kind of want to jumpstart that, this is the direction that I would go and these are the ones that I might use especially the three that show the most promise.
I did the same check on the images and the deviantart one is the only one that got the image into an image search. The rest were everything but the image so kind of keep that in mind if getting into an image search is important to you, the deviantart is your best bet of those.
I guess there was something in the efello, probably the sidebar and somehow the fake keyword got associated with all these other images of things that people have posted. That could be a good way to spam your competitors if image art or images are a good thing for you or something you might want to do, you could flood your competitors’ keywords or images here and associate them with other things. If that is what you are into.
Do You Need a Unique Page for a Canonical to Pass Juice?
This is a two-parter, we have been trying to nail down the canonical links pass juice, I feel that they do but kind of want to get that for sure but the next part of this thought was, does the page need to actually be a duplicate?
The idea of a canonical is you would use a canonical if you’ve got two pages on a website that are the same or mostly the same, what you would say is that Google mostly understand that these are the same, don’t worry about this one, I am going to canonical it to the other one. I would say that this is the primary canonical, just simply means that it’s the primary source. So you point from one and acknowledge that this is a duplicate, don’t worry, please index this one over here, this is the one that I like as the primary page.
But here is the thought then that the canonical links pass juice so we kind of want to look at that with this and then also, the concept that the page actually needs to be a duplicate if it is not a duplicate, it won’t pass juice.
Ryan Hough set this test up and ran it and had two different versions. Test one is an exact duplicate page that was created, it has the canonical pointing to the target page, to our test page. Test two, 51% unique page was created and then was canonicaled to the target page. I hope that is clear. So test one, we have a completely duplicate page, the traditional reason for a canonical and it has a canonical pointing to the target page.
Then we have a unique page, it is 51% unique and it is pointing to a target page, why 51% unique? Because we know that that gets past Google’s duplicate content filter. If you have a page that is 51% unique, it will not be seen as duplicate by Google so we know that that percentage gets us above that mark and Google will see it as a unique page. Maybe we can still get some canonical juice passage out of it.
Test one, there is the initial page and I can tell you and this is how we did it, just through WordPress, setting up a canonical url over there, easy peasy. The duplicate page that was sending the canonical link didn’t index and it has not impacted the test page which is super interesting so it is possible that it did not show up because it is actually doing what it is supposed to do, you told Google that this page is a duplicate, that we are sending the canonical link to somewhere else and they are duplicates and they are not putting it in the index.
What I find more interesting though is that it is not impacting rank, it didn’t move the needle at all, getting the target page to move. We are keeping an eye on that and you can see that this last check was on the 20 so this has been going for a little bit but we are still keeping an eye on this just to see if there is a movement that happens.
Test two, we have our target page and then a 51% unique page is created and the canonical is set. What was crazy about this one is that the canonical page is beating the target page. So this page is unique in the eyes of Google. IT is 51% unique, it indexed, and it is beating our target page. It actually pushed it down a spot so nothing changed in terms of rank but our target page moved down. Our target page getting the canonical isn’t doing anything here.
The unique page has moved itself to the number 2 spot. It is entirely possible that it is getting a benefit out of the source code where the target keyword is sitting in a url in the source code and that is actually giving it a ranking benefit over the target page. So we are going to keep an eye on these but kind of going the original thing, we are not getting a lot of love over these canonical links which is very surprising.
I really had the mindset that the canonical link would pass something but we are not just getting a lot of it here and we are keeping an eye on this, we are going to keep you informed, we are going to run this one more time because canonical links go to a lot of SEO theory and a lot of SEO techniques so we are keeping an eye on that there and we will let you know what happens but this seemed really good to me to present at this point so that is why we want to get this out to you guys now.
Charles has a good point that he has found in doing similar stuff that canonical links can take a long time to affect rank so these have been going for a little bit but we will keep an eye on it because that is entirely possible that that is the case. The only negative to it is that we see an accelerated results usually within the test environment within the real serps so this has been going [for a while] and we are right in the edge of a no in terms of how long it usually takes within a test environment.
Do Social Monkee Links Positively Impact Rank?
This is by Ryan Hough and our good friend Clint Butler. Clint, true friend of SIA and the time that he gives and has given to the group is phenomenal. This came about because some people were posting “hey, have you ever used this tool, have you ever used that tool?”. Last month, we looked at LinksRobots links and we actually got some really good results out of it. If you haven’t seen that, I recommend checking out last month’s dossier. It is surprisingly good, in my mind.
So this month, we are testing Social Monkey. Now keep in mind I have no affiliation with Social Monkey nor does Ryan or Clint, no endorsement or hate bomb is intended. It is just simply, has anybody used these things, let’s see what they do, is kind of what it is. LinksRobot, we had a lot of fun with it, again, no affiliation with anybody, just trying something out. I think actually Ben J from the group has been posting about these. So Ben, this one is probably for you, as well.
Five pages were set up for the fake keyword and then a hundred submissions from Social Monkee were dripped over nine days to the target test page. This is what it looks like in Social Monkee if you ever wonder what it looks like of the inside.
These are all the different settings where all the fake keyword goes, where the url goes, how it is all set up, the dripfeed, and number of submissions. Hilariously, it did drop. That is not uncommon when you fire links at something, to see some transitional rank where it will immediately drop. That’s normal and it is not a bad sign. However, it dropped to that spot and then it has come back only to the three spot. It came back to its original location. That is not the best sign.
Often when we see these things, we either see drop and then it shoots right to number one or we see some initial movement like to the two spot maybe one spot. You kind of see that instant boost and it sends up or we see a drop and then it charges right back up. In this case, we saw the drop, I think Google was sorting it up if it likes these links and then Google said nah, not really and just shrugged them off and put the target page back into its original spot.
Ryan and Clint checked out the links and of those initial links, you could see a bunch of them are dead and only a handful are still live shortly after checking them so, basically we did like a blast, dropped the site, a little time goes by and Google puts it back, Social Monkee kills these links because they need to use them again, released the pages again and there goes your ranking benefit. So, I would maybe avoid this one. If you need something, I would check out that LinksRobot first if you need something like these types of links. The Social Monkee one wasn’t nearly as fun as the LinksRobot.
If you are listening and you’ve got one of these things that you’d like us to give a go at, we’d be more than happy to. Post it in the Skype group, send it to me personally, and we would be happy to give any of these kinds of link program a run because they are a lot of fun to give a run through on.
Is an H1 in JavaScript Treated as a Normal H1?
JavaScript is a big thing, it is the bane of the existence of traditional SEOs because Google has had so many issues parsing JavaScript and displaying pages especially if you are relying on it for SEO purposes. We have been running a series of tests on JavaScript and we have seen that Google can see it – that you can put a keyword within JavaScript that is only existing in that JavaScript and Google will index the page with that keyword.
Now, we want to look at is it given the same weight? Can you rely on JavaScript for ranking purposes where you’ve got say htags or specific things that you know should move the needle? So, we’ve got these pages set up, the test page then received an h1, that is the only page our of all the test environment to receive an h1 and that h1 is generated by JavaScript and the rest of the pages are normal pages. The h1 that only exists on the test page is in JavaScript. If the h1 receives the impact that we expect it to get “normal” h1, this should pop to the number one spot quite quickly.
Initially what happened is the page dropped which was surprising and then I found out that my developer that was doing this, he put in an h1 but it didn’t contain the target phrase which is a bit of a biff on his part but that is fine, so the h1 exists, it gets crawled and drops, keep that in mind for putting JavaScript on your pages when all of your competitors don’t have any JavaScript. But we rectified that and put an h1 in that contained the target keyword and the real interesting thing, that is how the script looks like if you are a script nerd and you can see what that displays like.
Google can see it. It is there in the Search Console but it did not have a super great impact. It moved the page to the number four spot so one before its original but up from its drop but it didn’t pop it all the way to the top. This gave me some pause. It gave me a lot of pause. I was not expecting this. I was expecting that Google was treating the h1 like a normal h1 and Google was doing a great job at it and there were no issues.
I really expected this to move just right up and it didn’t. This gives me a lot of pause. If you are dealing with JavaScript elements to move SEO, I would be very wary with that. Just seeing that Google can see it, I do not think is enough. You need to make sure that Google is treating it as you expect it to be treating it.
In this case, I do not think Google is treating that h1 as an h1. We are going to run this again because I think it is a very valuable thing and a very important thing to be 100% sure on but if you are kind of in these point where you are relying on some JavaScript elements and you are not sure if you are getting the benefit that you should be getting, take a look at this. It is entirely possible, when you think about the rendered dom that Google is just seeing this as text, at this point.
The JavaScript renders, the page exists, and it is just rendering as text so in this situation, this page is just being seen as the keyword one more time on would have moved this page up a little bit but it didn’t zing it all the way up to the top as an h1 would because Google isn’t just seeing it as an h1, in this situation.
In a sense, it is seeing it as a larger text and we’ve been testing that over the last couple of months and the larger text didn’t do anything, it wasn’t treated as the htags, that might be what we are seeing here. Google is just seeing larger text and is not seeing the actual h1. So, keep that one in mind if you are in that situation where you are dealing with a lot of JavaScript, when you are doing JavaScript on the site, and there are some issues that you might be having. We are going to test this again but keep that in mind.
Will Google Index Numerically Sequenced URLs?
This is one of those things where one of the reasons why I recommend doing your own tests, you learn a lot even in the times that you make a mistake. You can just learn a lot with how Google works and what is going on. Normally, we do urls that look like this, we put the fake keyword as close to the .com as could be and then we do a dash and then we do a fake letter to represent a secondary word that is in the url.
The reason is, that gives a unique url and that gives as close as we can get it to eliminating variables and presumably, there is no ranking to nghi vs hbge, is the concept. So this is how a url should look like. I hired a new VA a few months ago and in one of her first sets, she didn’t put the fake letters behind, she simply let WordPress numerically sequence them.
The first one was the fake keyword and then the fakekeyword1, then the fakekeyword2, fakekeyword3, and fakekeyword4. The interesting thing was on the sets that she did and I think she did 3 sets, none of these 1, 2, 3, 4 urls indexed. The only one that would index was the one that was just the fake keyword. So I decided to set up a test to see if I can replicate that. So I created 5 pages, I just numbered them sequentially so that is the .com/fake keyword, then 1, 2, 3, and 4 then sure enough, the only one that indexed again was the one that just had the fake keyword.
So the 1, 2, 3, and 4 versions of this page will not index. I am wondering if this is one of Google’s very first ways to combat doorway pages or some sort of spam where people were mass creating pages and they were just numerically sequenced 1, 2, 3, 4, 5.
I pointed this out because you might be in a situation where you are doing this with landing pages and you might have a system where you are creating multiple landing pages and maybe you don’t care about if they index or not, if they rank or not. You might run in this situation where they might not rank simply because of this numerical sequencing. I am not sure and I do not think it affects pages.
If you look at some system, I think Wix used to do it in their system and they would just throw away a whole number sequence at the back, I do not think that is quite the issue but it is quite possible that is part of the issus because of you recall, three or four years ago, Wix was having extreme problems getting their pages indexed and ranked and they had that number sequence at the end, maybe that has something to do with it because I know they got rid of it. So keep that in mind if you are in a system that might do that where this kind of number sequence is going on with your urls, give that a look.
Another thing is, this might be a way to hide pages if you need to create pages and you kind of don’t want them to show up in the internet but you like to use them for link building purposes, this could be a way to do it where you could just simply sequence them numerically and perhaps they won’t index for you but you could do anything you need to do to pass links to them. Something worth testing, I didn’t test in this, this was just showing that was how they are treated in Google.
Does LSI Affect Rank?
We did this one before and I decided to do it again for two reasons, one is our test methodologies are better now that two years ago. We are simply running better tests, the environments are better, the results are more reckonable [?], you can see better patterns.
The whole methodology is better and it felt like a good time to run it especially, I think this was two months ago. Barry Schwartz tweeted that LSI isn’t a ranking factor and it really blew up and a flame within he SEO community. I was, this is perfect, we have better test methodologies and I like where we are headed and what we are doing and then Barry Schwartz does this, this SEO good vs that and then another SEO guru says something else and it’s the perfect time to retest this and so we did.
Real quick, we needed to define and that is where maybe things get off the rails for a lot of people on what LSI is. LSI is latent semantic indexing. It is important for you to understand the word latent and semantics, what they are. The classic example that I used is kitchen, if I am talking about kitchens, I am going to say things like pantry, stoves, sinks, cooking, tile, granite, farmer’s sink, those kind of terms.
Those are LSI which you’ll notice, those are the other words that come up in a conversation about something that gives context to meaning. That is what latent semantics mean. Latent means other or below and semantics means context or meaning. So these other terms that gives context and meaning to what you are talking about.
Keep in mind that none of those words are a synonym of the word kitchen or a variation of the word kitchen. I am not restating the word kitchen in a different way, I am saying other words that gives context to what I am talking about. If you are called the show 20,000-dollar pyramid, I can say sink, stove, cooking, pantry, in your house, and you would say “oh, a kitchen”.
That latent semantics, those other words gave meaning to what I am talking about and that is what LSI is and so, I think that maybe a lot get of the rails because people can’t figure out what LSI is in the first place and do not take time to look up what these things mean. Running with that definition, that is what LSI is. They are these other words that come up in a conversation. Now, are LSI a ranking factor? Are Google using these things? And that is what we can test.
I used the tool Ntopic for LSI. I like it, it is predictable, it is usable, and quite good, and cheap. Ntopic looks like this, if you are listening to this and we never talked about Ntopic before, you never heard about it, it gives you the score, don’t worry about the score, it is very gameable and not important. What is important is it pulls up all these other terms in here and these are often great candidates for LSI and these are the terms I try to get into my content.
I do not worry about the recommended percentages and the actual percentages and don’t try to shoehorn these in to my content either. I just make sure that I am pinging an appropriate number but this is where I get my LSI from and for this particular test, we are going to grab it from here, as well.
So what we did is, we indexed five pages. Because we are talking about LSI and we are talking about actual English, we had to use an absurd phrase for this and so in this particular example, we used “wonky ficus stand”, this then allows us to put it into a tool like ntopic and we can get LSI out. What we did was we put an LSI into the number three ranking page, just swapping out our lorem ipsum for these actual terms that pinged as LSI. The other pages get an equal amount of English words.
So for fig, we put in dog. The idea here is while fig pings as LSI, dog does not. It is a random word that is not LSI. Ficus, ball, tree, tire. The concept here is that the other pages are getting English words and they are getting the same amount of English words, they are not just getting LSI for our target phrase.
We relaunched all those bad boys and the LSI page goes to number one. We ran this another three times. So, back in the day, two years ago and some change, we declared that LSI is a ranking factor and I feel very good about saying that again. LSI is a ranking factor.
Now, Barry Schwartz is an extremely talented person and he is an SEO journalist and easily the best SEO journalist that is out there. He does not think LSI is a ranking factor because LSI is old. That is one of the arguments here. It is an old method, it existed before the internet, and therefore can’t be part of this.
Other leading SEO would tell you that they haven’t seen a Google patent for it, if there is no patent, it can’t exist and then these flame wars ensue and they go back and forth and they are very hateful and people have a lot of wild things to say but they never ever ever say that I have tested it and here are my test results for people to look at, let’s talk about these tests. It never happens. It is simply these arguments of anecdotal evidence, I used it and my pages have ranked or it is old technology or there is no patent or I’ve tried it and it didn’t rank my pages.
I love those flame wars because they distract people from actually doing their jobs and at the end of the day, we’ve run these tests and we know these things are working and this is what is going on. So, as I have gotten the last line here, when this happens in the SEO community, just sit back, smile, don’t even engage, let them go and run of the rails completely and entirely. Because if they are doing that, they are not spending time optimizing their pages, you are already getting an edge. Just grab your popcorn, work your LSI in your content, rank your pages, and just let them be.
Do Keyword Matches or “Variations” Affect Rank?
Keyword matches is a concept that comes directly from Ted Kubaitis and I can say without a doubt that this is the single largest movement forward in SEO thought for the last couple of years. Remember when we said just a minute ago that LSI are not variations of your keyword or not synonyms of your keyword? What we are talking about here are actual or potential synonyms of your keyword or variations of your keyword. That is what keyword matches are. They are a completely separate thing from LSI.
A good example is if your keyword was “DUI Lawyer”, keyword matches might be DUI and lawyer on their own as individual words. It might be DUI attorney, attorney and DWI. So these are words and phrases that are part of your exact keyword, they might be your exact keyword broken up into their individual parts, it might be a very close synonym.
In this case, lawyer for attorney and then, DWI, which is a different way to say DUI, depending on your jurisdiction. What is great and what Ted identified is Google shows you these things and shows you them in the serps and in the serps, Google will identify these terms and I think the right way to think of them are critical matches or critical variations of your target phrase.
The idea is, we can identify these things, are they a ranking factor? So, same with the LSI, we need to be able to see serps so we have to go with absurd phrases, in this case, in this example, sparkle lotion toenails is the absurd phrase and we get results out of that, nobody is overly optimized for that but we can then create an environment where we can get variations out of it and tune the pages accordingly. I used Page Optimizer Pro to get the keyword matches.
Page Optimizer calls them variations and that is what they look like and so this is what pops up, toe nails, toenail, glow, sparkly, sparkling, fingernails, sparkles, etc. and then, the recommendation from the tool was to use these variations 13 times in the paragraph text and that is what I did. For the target page, the number 3 ranking page, I put in real keyword matches 13 times.
For the other pages, they received also 13 words but they were not keyword matches. So we are here, we are talking about toenails and sparkling and glitter as keyword matches or variations. Here, they get butter, table, and elbow. Each time, the page receiving the variations shoot to number one.
This is a pretty phenomenal result here because it is identifying a new ranking factor and it is something that Ted has been talking about for a couple of years and something that I have personally bought into and I have been optimizing my pages for but here, we have some proof that this is a thing. For those of you scoring at home, this is the second ranking factor that we have identified this year that other people aren’t talking about. There is another edge that you are getting and it would be appreciated if you wouldn’t share that with anyone.
The important thing is, and often where this comes into play is in a long tail phrase. If you’ve got a three-word phrase or a four-word phrase, you might be really tempted to use that three or four-word phrase a lot. You probably put it in your metatitle or h1 and that probably is fine but then you throw it in the body 2-3 times and you will throw it into an h2, h3, and h4 and your img alt and you look at the page and it is kind of a couple of thousand words only and “ah, I really only got it there five times”, you may have it too many times.
It is entirely possible that is over optimized for that long tail phrase. How can you avoid that overoptimization? It is using that exact match phrase a few times and this would be determined by the serps and you have to do an analysis but then what else should you do because you still need to continue to optimize, that is where the variations come in.
You can take these variations, the single word usage in your phrase, the short phrase of your phrase, or those very close or critical synonyms of your target terms. Those are then what you want to optimize for. So you want to determine your variations in addition to your exact match and then you are going to want to do some analysis to make sure that you can optimize for your exact match and then also sprinkle these variations and as we know then, we got our LSI and that is a separate thing, those are also going in.
When you think about those three factors, most of the SEO community outside of the people on this webby and on the group, they are going to look at their exact match keyword and that is about it. What we have identified is that you need your LSI in there and you need your keyword variations in there because those are both also ranking factors and are extremely important to your onpage.
Take some time, look at it, if you’ve got pages that are stuck, why is this page not moving, I don’t quite get it, you might want to look at either your LSI or your variations or your keyword matches, as you will, as Ted calls them. That is where you might be missing. Your exact match usage might be 100% on pointe, you’re right in it, but those other factors which are ranking factors, might be a little off and that might give you something to look at in order to optimize your pages and get your edges.
You mentioned that this last test was second factor found by SIA, what was the first one again?
We found that unoptimized zones, htag zones are ranking factors. Let’s say you are in a situation where you have 10 h2s on a page, and even though they might not be optimized, if your competitors are in an average of 1, you are probably harming your page. That unoptimized tag zones, htags are a ranking factor, and we found that h1, h2, h3. Kind of going to a site structure thing.
I think it deals with what Google is anticipating, what it is expecting to see for a particular keyword. So, you are coming in with a page that has all kinds of h2s and h3s and the competitors don’t have that, you are harming your page.