We discuss the tests under the June Dossier.
Kicking it off with do Inserted Links in Canonical Pages Count? Is it a ninja sea hack that one can do to bump up rank? Followed by other on page tests such as, does Hidden Content on Mobile Negatively Affect Rank? Should you avoid hidden content when on mobile?
We continue with the iteration of our blockqoute tests with how does Google treat content in a blockquote, this time the test is in the inverse to see if the inverse rises to the top and if it is indeed a negative factor. The other iteration is, if high authority content in block quotes are treated the same as the first two tests or perhaps it shows differently?
Meta keywords was also tested and discussed if you should be putting your keywords in the meta keywords or is it something to avoid? Continuing on with our schema test series, will an internal link with Schema pass more juice than an external link without schema? Should you put your keyword when you embed a mymap to a page? Will this show up in the serps? And with that, we have the test “Showing Results for” Keywords or The Keyword Switcheroo.
Since we have been having issues with indexing, we test and discuss if the BulkAddUrl Indexing Service is something worth using to get your pages indexed. And with the indexing issues we’ve encountered, we also discovered a new way to test without the indexing issue being a nuisance and this was done in the final test for this dossier - does LSI proximity affect rank?
Great tests for this month and new ways of looking at things and testing methodologies.
Tests this month
Test 325: Do Inserted Links in Canonical Pages Count?
Test 326: Does Hidden Content on Mobile Negatively Affect Rank?
Test 327: Will an Internal Link With Schema Pass More Juice Than an External Link Without Schema?
Test 328: Testing Meta Keywords
Test 329: How Does Google Treat Content in a Blockquote? (Part 2 – Testing The Inverse)
Test 330: How Does Google Treat “High Authority” Content in a Blockquote?
Test 331: How Good is The BulkAddUrl Indexing Service?
Test 332: “Showing Results for” Keywords or The Keyword Switcheroo
Test 333: Will a Page With a Keyword in an Embedded MyMap Appear in The SERPs?
Test 334: Does LSI proximity Affect Rank?
June 2019 Dossier
Transcription Below:
Do Inserted Links in Canonical Pages Count?
Jo, one of our regular testers have been doing a lot of testing with iframes and this is an extension of something that he saw within the SIA Skype group where Bill was talking about some negative SEO with canonicals and he was wondering if this can be done in a positive way. If you were to place an extra link on a page that is canonicalized to another, will it boost a page?
What he was actually talking about was, you’ve got a target page which is the page you are trying to rank, what if you duplicated it and canonicalized it back to itself, would it get a boost? The idea is, what if you do this multiple times?
What he found was that, you can. Set two is the important set, moving up to the third up to that second set, but duplicating a page and canonicalizing it back to its original self, you can get a boost. You can take a money page and duplicate it for “innocent” purposes. That is the kind of thing that you could canonicalize to itself, you can link back to itself, and you can get a boost with that. So when you are looking for those opportunities for link building and boosting from your own site, this is actually a pretty interesting opportunity.
Did the other pages get indexed before adding the canonical?
Yes, they indexed as I recall from talking with Jo. It’s actually a cool thing, just taking a page that you already built, copying it, and canonicaling it because you are saying – hey, this was a copy – and then, you can get a link boost out of a page that you already built. You could do it in your pbns, you can have a whole hidden set, basically your entire site duplicated as you build it out.
I think this is what Weebly does. If you have a site, Weebly creates a subdomain that they do not show you in the same way, it’s almost the same thing and people talk about how Weebly sites are still good for that - that if you are building because it kind of creates a duplicate page but it does a canonical.
Let’s say you want to build links to your pbn, you build a canonicalized page and no one is ever finding those links and they won’t see how you were actually doing your link building.
Does Hidden Content on Mobile Negatively Affect Rank?
This was done by Marc and he has done this really cool thing where within mobile, he is hiding text – a certain percentage, on different pages. Some pages are 40% hidden and the test page is 100% hidden, it shows the ranking results between them – hidden or not hidden. What he found was that there wasn’t a lot of movement, he wasn’t 100% sure that the test set that he was using was being crawled by the mobile results.
That was an interesting thing because I actually had this with clients – where they are not sure if they are in or not in the mobile first index and these are large clients in huge niches. Since they are not sure that they are in the mobile first index, they are not sure that they are actually being crawled. In this case, it makes a great point. If you have clients and you want to see if you are in the mobile first index, do a hidden work that does not show up in the desktop view and would only show in mobile, and actually see if that keyword appears in the serps, then you know if you are in the mobile-first index.
At that point, you want to make sure that you have done some comparison between the site that you are editing because most people generally edit in desktop and if you are in mobile, then you want to take a look at the difference between the two to make sure that you are not missing anything from desktop to mobile. This test didn’t get us where we needed to but it did made the point of making sure that you are in the mobile first index, first.
Will an Internal Link With Schema Pass More Juice Than an External Link Without Schema?
Ryan has been doing a lot with schema, in particular pages that schema on them that have links on them – are those pages passing more juice? What is really cool is that yes – a page that has schema where Google can figure out what that page is about will pass more juice. Now, he is trying to compare what about an internal page with schema against an external page without schema.
He set it all up and it turned into an interesting problem. The test turned out to be a little bit of a bust because he didn’t realize he was using post and the post ended up on the homepage. That happens naturally in WordPress and didn’t realize that it would happen in the test site. So he used post and what he almost created was a mini silo because what ended up happening is the page, the homepage, the post page ended up as 1, 2, 3, beating everything else out.
Essentially, these little guys created their own internal silo and they beat everything else that he was doing with the test.
This goes back to what we talked about a bunch of times about your internal link structure and how powerful that can be. The one thing that you can control the most in SEO is your own site – you can control your own content. I have heard of and seen people that might have a site that they have somehow overcoat their links through and these are massive ecommerce sites but 98% of all sites out there, it’s really impossible to screw up your internal links.
You can’t bake yourself with your own internal links so if you have a structure for it and you have done your proper on page, you can create so much value within your own site and you worry less about the links that you have to build if you have that internal link engine going on within mini silos and this is just a fantastic example. It was a test that was busted but it turned out to be actually pretty cool.
Testing Meta Keywords
This test came from Ted Kubaitis. Ted has been hearing from a lot of people that meta keywords are harmful. Meta keywords is the most old school thing in SEO – you’ve got your meta title, meta description, meta keywords – and people are saying not to put meta keywords in because they are harming sites. These are one of the things that we’ve been talking about and I think that is silly. I think they are just ignored.
Ted run this through the system that he has for testing and what he said was that it is not a factor and if it’s harmful – it is really close. It resembles a smoking gun. It is a weird thing because I have often told people that I think putting meta keywords in is fine, it takes two seconds and it probably helps in Bing and you are probably not harming yourself.
This is one of the things that are, maybe not. What I think the lesson here is that I don’t really think there is a problem having one or two meta keywords, I think it is completely fine and I think it helps in Bing and Bing influences Yahoo, Verizon, AOL, they are all the same engine so it is probably not a bad thing to do, but don’t go crazy with it.
This has been the advice for some time, you shouldn’t go nuts on these kinds of things but it is also pretty interesting that it is possible that it could be bad, that it is really close so be super careful. I’d still tell my clients to do one but just be super careful with it and not get nuts because as Ted was seeing with his findings, they are right at the edge.
How Does Google Treat Content in a Blockquote? (Part 2 – Testing The Inverse)
This is part two, last month, we had the idea of showing iterive testing. Test number which was really weird was in the blockquote, the test page dropped – this was putting content in a blockquote and it drops. What is going on here, was the question, was Google negating that content so it drops because its word count went up? Is Google looking at this as a negative factor – you blockquote something and they demote you.
What I wanted to get people involved in was in what was going on with these things, to help brainstorm. This is part of the testing process. Part one was the blockquote that dropped so we want to run the negative immediately. When we see something that drops quickly, we want to run the test the other way – running in the negative – so that we can then see if we are really dealing with a negative factor at this point – running things the other way to see if it goes up.
Running in the inverse, we actually got it to move up a spot. This is kind of weird, something is going on here with blockquotes that is a potential thing. Again, keep in mind these are just one way tests. We haven’t run this multiple times, we aren’t getting to statistical significance, but now this is pretty interesting. We put blockquotes on the three and it drops, we put blockquotes on everybody else but not the three, and now it moves up.
When you ae talking about how tests work and how this whole system that we created works, now we are kind of – this is a bit interesting, it is weird. Again, we are also at the point where we do not know why. We can speculate - which is a part of the process - why this would happen, why would it not happen, but now we are saying that this is interesting if you’d take content and put it in a blockquote, it would drop. We still could be at the part where it is a word count issue, for sure. It’s a super strong thing, it could be worthwhile removing a keyword there.
Dori: Wouldn’t it consider that content in the blockquote? They have to be reading it, they read accordions and other stuff.
That’s the other thing here, maybe they don’t. Maybe they actually just negate it. You put something on a blockquote and they just wipe it out. That is part of this process that we are seeing. We have this first weird result where it dropped, so we ran it in the inverse and it goes up and that is super weird.
Dori: We have some input here – try putting the keywords in the blockquote to see if they are reading it and we could get something to rank, Tracey says blockquote citation or content from other sources maybe?
I think that was a wonderful softball tossed by Tracey, so the next thing we did was test 330.
How Does Google Treat “High Authority” Content in a Blockquote?
This came from other people that were listening and decided to get in on this. This was the point of this test – finding an odd result and walking through the iterative process of testing but also getting people’s feedback. The first thing was, what if you’ve got a blockquote from a high authority source, which was a great idea. But then we had to brainstorm, what is a high authority source?
This test concept came from Search Engine Journal and we thought, what if we took content from Search Engine Journal and blockquote it and then added it to a page. So now, we’re adding content to a page. What it looks like on the first two runs was that the content was just ignored completely, almost as though whatever you blockquoted is just taken out of the page. This was user generated.
My thought process was, if we get content from a page that is ranking for a particular keyword or some keywords and it comes from a site that has a lot of traffic, we will call that as high authority. That is the assumption that was made and again, you can disagree with that and that is totally fine. That is what the testing process is about.
We’re taking content from Search Engine Journal, we are saying that it is authoritative because this page ranks for its terms and it is a high traffic site. We are going to put it on a page, added a blockquote, and see what happens. What was fun in this example is that it moved right up to the number one. What we found here is that maybe Google isn’t ignoring it entirely because none of the other pages got this blockquote. This content existed and it moved right up to number one.
That is where we are at now in this series and this is something that can be continued. Adding keywords, removing keywords, are something that could be done. I hope that this spawns ideas for people, because we could keep this test concept going but this is where we are at with this process and what all the tests could see within sia, this is a process that we are going through, we run these different tests and a lot of tests do not get shown because they were part of this process of trying to figure things out and a lot of the tests that you see are the in result test.
What I am hoping is that, people have ideas about what might be going here, other variations, other things that we try, and then jump in. You could become one of the testers of the sia, you can join the group, and give this a go. That is where we are at now – we did a high authority blockquote, it moved the page up in spite of the negative factor that we are seeing the other way.
Is it possible that Google is seeing it as duplicate content? How many words did you use as a blockquote text?
On the page was 750 words and I think about 100 were put in into the blockquote. That is a great concept – that is the whole point of this, what could this be, what could we change, what is going on. Perhaps that is correct, that there is a duplicate content issue. I love that idea and that is actually something that I have not thought of. That is the point of this, I haven’t thought of that, that could be, let’s give it a go.
How Good is The BulkAddUrl Indexing Service?
=(I’m sure that you are aware of the extreme indexing issues the last couple of months. Google has rolled out several updates and every time Google rolls out an update, it just messes up everything. It is really hard to get pages indexed, it is a real pain and I am sure it hurts some of the links that you are building because you can’t get Google to run them, it is awful.
A service that is continually referenced as a good service for indexing is BulkAddUrl and Boryslav gave this a go to kind of get a baseline – what is a good indexing rate. He did five test keywords and five pages – 25 pages that he put in to get indexed. What he came out with was 76% index rate. If you played with indexing, that isn’t bad. It is a pretty decent number. The great thing about this test is we are in an era of weird indexation and people are looking for answers to how to index pages and you might be using services. This is an easy test to replicate.
You could create 25 test pages, put them into an indexer, it would only cost you a few bucks, and you can see what percentage indexes. This is an indexer that people likes and it is reputable and it does a really good job and it does about 76%, that is a good base line. The ideas is, if you run things in the indexed that you like and you are below 76%, that is probably is not getting the job done, it is time to move to another.
A lot of people ask questions about good indexers and here is how you can figure that out – run this kind of test through the indexers and get the number or percentages of how many were indexed. Seventy six from a good source is a pretty good number, if you are below that, that is not where you need to be, if you are above that, they you found a jump. What we wanted to get to the SIA group there is that is a decent number for services and test your own and feel free to post that within the Skype group.
“Showing Results for” Keywords or The Keyword Switcheroo
This is my favorite test in a long time. We tried to create fake keywords. That is what we are doing here because we want an environment that we can control. That is a big part of this concept. Every once in a while, you get a keyword that isn’t clean. So it is a fake keyword and it doesn’t exist and all of a sudden, Google is – did you mean this and they do the showing results for.
An example is, in the screenshot. This is the keyword switcheroo – my keyword is what I typed, Google switcheroos it to another and they show you results for that. What we realized when we were doing these tests were, they kind of messes up things if you are doing this with a rank tracker. We can’t get accurate results for these things.
I was – hold up, how is this actually happening in the real world? What if this was a brand? We did some deeper dive and so we did a fake keyword and it had the switcheroo and you can see that ProRankTracker which is a very good tool, it has the same page ranking for both but it isn’t. That is just the keyword switcheroo page – is that they are showing.
We realized that it is really problematic. It doesn’t happen a lot in the real world but let’s say you have a brand and it is a hipster cool brand that decided to misspell their word, if you are tracking things through a regular rank tracker, you might actually be a victim of the keyword switcheroo and you might not be getting accurate results for that actual term that you really want to follow. The other thing that we realized, a very clever way to do SEO sometimes, is to optimize for that misspelled word. That is a good tactic where you could make a lot of money from where there is a complex word to spell and they are often misspelled.
You want to rank track for that misspelled word and there is a chance that the rank tracker gets stuck with the keyword switcheroo and those are the results that they are going to show and they are not going to show accurately.
It is not a huge problem but we realized that it was enough of a problem to report.
Will a Page With a Keyword in an Embedded MyMap Appear in The SERPs?
I am constantly on the quest of finding factors. It is one of my favorite thing. Is that a factor? Could it be a factor? If you see something that could be a factor, the first thing that you’d want to look at as an indicator is will a page index with the keyword in that zone of in that factor area, whatever you’d like to call it.
What we are looking at right now is, if we create a mymap, if we have a keyword within a mymap and in this case, in the title, can we index a page with that. The idea is we’ve got an embedded map, is Google looking at that map at all?
What we did was we started a page and we indexed it for a fake keyword. But then, we created a mymap for another keyword. So we had a page that Google has crawled, it likes it, it exists, and then we are inserting something new to it. If you think you have a potential factor, this is step number one – create a page for one particular word and get it indexed, and when you can see it indexed for that keyword, now put your factor into play with a new keyword and once you’ve done that, see if the page will index for the new keyword.
That is what we are doing here, the page indexed for that keyword, we created a mymap and within that mymap title is the new keyword and what was pretty sweet is that it indexes. This is a cool find. That embedded mymap will be read by Google. At a minimum, we know that it will read the title and that is probably where you are putting your target keyword anyway, so that is where it exists, what is super fun is in a page that is devoid of any kind of real content, it will actually put it in the meta description.
That is a great way to get other words on the page. This is the kind of thing where you could do that, you can get a completely different keyword on the page that you want to rank for or perhaps maybe a really close secondary keyword or other keywords that you really like but you are already optimizing for one, mymap would do it for you. This is something that you could put into place.
One other thing that you could think about is if you were a fan of the concept of non-zero factors which is a concept that we talked about for a while. It is a little confusing. Non zero factors is the concept of it is better to have something in a factor than zero. It is better to be under than to have nothing within a zone. This was discovered this year by Lee Witcher, one of our members.
When you think about the match words – the bold terms that Ted Kubaitis found two years ago – I think the non-zero factors is the next most important thing in SEO. Those two finds, I think, are the two biggest finds that I have personally seen in the last two years. Ted’s match terms or variations – whatever term you like, the bold terms in the serps, and the non-zero factors that we talked about several times that Lee found.
If you like that idea – the idea of non-zero factors- this is the kind of thing worth drawing on, a map and putting a keyword on it may be a good idea because then, you would fulfill the non-zero factor concept.
Does LSI proximity Affect Rank?
The last one is also from Lee who does great work. Lee is a math analyst. He is the kind of guy that lives in math and we need people like him. I think my special skill when it comes to SEO is I have a way of thinking about things that might be a little bit different. When we look back to 2015 and I did the talk about SEORockstars about the weight of smoke, how do you measure the weight of smoke. I think that is my special skill, how to figure out a way to test Google’s algorithm.
I think that is my personal special skill. Lee is the kind of guy that can crunch through numbers like you wouldn’t believe. He’s the kind of guy that you can give a set of numbers and he can figure out where it is at. That is his special skill and I am happy that he is involved in SIA and gave us this.
This is a great one that I wanted to end with this dossier because it presents a new testing model and this is a model that I have never thought off and Lee came up with this concept and I think it is really great. What he is talking about here is LSI proximity. LSI are terms that come up naturally in a conversation.
The example that I always give is when we are talking about a kitchen, LSI isn’t a synonym for the word kitchen, they are other terms that come up when you are talking about a kitchen. So, sink, refrigerator, pantry, those aren’t variations of kitchens, but they are other terms that describe a kitchen or come up in a conversation.
What Lee is doing here is he wants to look at LSI proximity, can I jam LSI words close together to affect rank?
As we all know, we’ve had this crazy index issues for the last several months and what Lee figured out is that, that does not affect YouTube videos. YouTube videos and those videos that also show in video results are not affected by the same index issues that regular pages are being affected by. His thought process was, what if we created videos and see how they showed up within YouTube and within the video results and then extrapolate that data.
Here, he is creating videos and he is clumping LSI terms together, basically is what is going on here. You can see his results and the end result is, Google appears to be using LSI even within. These are videos that are launched but LSI is used in Google, within the Google videos but YouTube is not using it, is what he found.
Here is the fun thing which I wanted people to leave with, this is a great way to test things that you can do within paragraph text, if you want to do a quick test if something is a ranking factor or not, if something is involved within the paragraph text of a YouTube description, you can’t do a lot. You can’t do h1s, h2s, h3s, bold, or italic, but you can do text and there is a lot that you can do within text such as this and if you want to get into super advanced things, you can get into word to vec or human vectoring, other type of concepts with nlpt, those kind of things.
You can do some pretty advanced testing within the paragraph text of YouTube videos and you can consider the title as well as your h1 or your meta, at least you have that. You can use this as a quick test tool. Previously, I have used and I have spoken about using videos to keyword poke, to find soft serps, if you create a video, and then launch it, and see if it hits the serps on page one or page two, you know you have a soft serp and it is something you can go after but another way to use this concept is to do things within the paragraph for paragraph concepts within testing.
If anyone listening has done their own tests or of course done any work on the indexing issues that we’ve had, this is a great work around it. You can do some semi small test within the paragraph text using YouTube videos. It is a great way to launch things if you even want to do some testing to begin with and then you can extrapolate that data and move it into on page testing once the indexing slows down.
I loved this because really, it is a new way to test and I have never seen anybody ever talking about doing this type of test. This is a fantastic way to get at it and if you are new to testing and you want to give it a go, this is a great way to go because you can just create these videos with your content and you can rock and roll.
Just let me be clear, he is using YouTube videos instead of html pages?
That is right and the cool thing is that these things show in the Google Video results.
You could put a fake keyword in there and that would be the only results that would show up same as an html page?
That is right and they are not affected by the indexing issues that we have been having on regular pages. It is completely around that. You are limited a bit by what you can put in but you can start to run test on things for fake keywords or even real keywords within this environment and you can get past all those indexing and get very fast results.