For this month’s webinar, Dori discusses the tests from the July Dossier. She starts it off by an iteration of 2 previous tests to check if having a question mark in the title can affect rankings, followed by if there is a correlation between the number of pages per topic and rankings.
She also discussed if there is a preferred number of images on a page and if sentences per paragraph length matters. When it comes to links, which is better, topical or non topical links?
She also discusses the follow up results from the tag page with canonical test. Does it finally reach the top or did it drop? Another continuation test, how does the BulkAddURL indexing services fair well for bulk indexing of pages?
When building a site, what is the best domain extension to use? Will html pages rank better or does WordPress beat it to the top? Should you also add your site to Google Analytics? Is that a factor or not?
All these are answered in this month’s test and webinar. Jump aboard!
Tests this month
Test 335: Does Having a Question Mark in the Title Affect Rankings?
Test 336: Is there a Correlation Between the Number of Pages per Topic and Rankings?
Test 337: Is there a Preferred Number of Images?
Test 338: Topical vs Non-Topical Linking?
Test 339: Does Adding a Tag Page with Canonical Act Like an Internal Link? (Part 2)
Test 340: Do Sentences per Paragraph Length Matter?
Test 341: Google Analytics - Is it a factor?
Test 342: Will an HTML Page Rank Better than a WP Page?
Test 343: Continuation of BulkAddURL Indexing Service
Test 344: What is the Best Domain Extension To Use?
July 2019 Dossier
Transcription Below:
Hi everyone. Dori Friend here. How are you doing? Give me a one in the chat box if you can hear me and then a two in the chat box if you can see my screen. Hey Clint, I see you there. So, let's get started. Okay, we have some great tests coming up.
Test 335: Does having a question mark in the title affect rankings?
This one is, does having a question mark in the title affect rankings? Now, I have to be honest, this was originally inspired by Boryslav and he was inspired to do it by using Cora. He first did it months ago and I was looking at it and I was like these don't even have question marks on them.
So I said: let's do this test again. But basically, what was going on in his first test? I had misunderstood it. They were testing factors like who, what, where and if you had those kind of words in your title tag would it be a ranking factor?
Hypothesis is this question? Will a question mark in the title beat a page without a question mark in the title? I just wanted to know if putting a question mark at the end of a title, was going to give us a bump or not.
So first of all, I just I have Boryslav stuff in here and he had the word the in all of his tests and then changed out the word why. But they did have a question mark in everything and he did the same thing for who. They both shot up, but then they both landed in the fourth position and they're still there to this day. I checked them. That was kind of interesting. So I don't know what Ted thinks about that, but they didn't do anything worthwhile anyway.
So we decided this is gonna be a simple test. Let's just start off doing one, which is adding a question mark only to the title tags at the end of the title. So we did not use the who, why or anything like that. In these, I just wanted to start basic. You know, let's put a question mark at the end of one title and see what happens. And of course, do the reverse too. It bounced around and then it moved right to the top. But I'm like holy shit, man. This is Christmas. This is awesome.
You know moved to the top. First, you can see that there were only five indexed. I talked about an indexer but we're trying to go for eight to ten pages indexed and per keyword. But when we got five were like okay, let's start to test. So we ended up getting seven pages indexed on this and even with that, it shot to the top, but it bounced around a lot. It did come to the top again. But it bounce around, then went to the second place and then it finally disappeared.
That was a bummer but we were also checking the cache on that. Every time we would check the results, I would have my staff checked the cache because I want to see if Google was crawling it and if it was getting seen. Because a lot of times the movement is based on whether Google has crawled it and cached it.
So when we checked this - and we noted that before this page disappeared, the cache showed a 404 error on it. So, I didn't know what was going on with it - maybe somebody out there knows, but I didn't know.
Of course, we decided to do the test again and in the reverse. So this time, instead of putting a question mark on just one, we put it on seven and then left one without a question mark. The same thing kind of happened. This one started going down and doing the same exact thing; it just went down and then it disappeared. So I thought that was really interesting when we still had other pages showing, but we didn't have as many pages. There's seven but at the end there it's only four, but that page wasn't one of them.
Now on this one with the crawl date, the cache never changed. Iit had cached once and never never cashed again, while the others were caching. So I'm thinking, okay, they're caching, that means it's going down because the other ones are getting crawled. It never cached again and just disappeared.
So the results are inconclusive and what we need to do is get all the pages recrawled and cached again to get more of an insight. But what I got here is a 404 error. You know, it might be an indication of what Google's culling. They're culling out of the index of the search results, so that might be an unveiling of something to warn us of maybe one of our pages getting removed. I don't know. I also thought well, there's one thing different and that one thing is getting removed from the search results.
Maybe that has something to do with it. I don't know so. This, like I said is inconclusive. We’re gonna have to redo this again and see what happens and investigate a little bit further into the 404 error that showed up right before that page was no longer in the search results. So we don't know about a question mark - inconclusive.
Test 336: Is there a correlation between the number of pages per topic and rankings?
This one's by Marc. Thank you Marc. Test 336. Is there a correlation between the number of pages per topic and rankings? So his hypothesis is the more pages we have on a certain topic, the higher you rank in Google, right? Sounds good. My theory is, it would, of course.
So the background information: In local SEO, we can often see the websites that cover one topic from lots of different angles seem to rank better in Google. These are websites with 200 pages on the same topic. It could be an electrician who has a page for each suburb in their City and all relative services or related services.
Each page can be a source, it sends links and then it internally links back to the main ranking page. So it could be a home page that we try to rank or money sites. That's always kind of been case. This is really relevant too since we know that multiple inner links will help rank, even just on its own. So it makes sense.
He set up the test running and intitle and intext search operator for the five top search results in the SERPs compared to ranks 82 - 85 in a given search, specifically ABC electrician.com. So intitle would be, electrician results taking from Google pages. He is reporting that intitle gave much better results intext.com returns pages where the menu items have the keyword in it, which over inflates the number of pages and messes up with the data.
But the results are the strong correlation of negative 0.59 with rankings in numbers of page for sites:domain.com, which is intitle:electrician. Bottom line more pages you have a certain topic, the better your rankings are. Here is his graphs on that. Marc, I know you might be on here, so if you have anything that you want to add, you can just put it into the chat box, and I can let everybody know.
Correlation is not causation. When a local website has 300 pages such as an electrician plus suburb, this means it's in the hands of an SEO guy. You know it's been SEO when this happens. Therefore, you know there's other things that can be done to the website apart from mass building pages.
You can't really single out the factor of having many pages as the one that moves the needle. We can't say that clearly because there's other things going on. What's clear to say is ranking number one for suburb level query is fantastic. And if it also helps to power your home page, the money pages, then why wouldn’t you build more pages, right? Of course, so thank you Marc for that.
Test 337: Is there a preferred number of images?
So here's a test by Hunter. Number 337. Is there a preferred number of images? The hypothesis is? Google includes an image density/ratio and ranking content. Now this is something I can't imagine we haven't done before, but we hadn't done this before. We know Google ranks images, of course, and we know Google likes stock images images that it knows better than unique images. But we never tested to see if there was an image density in place.
You know much like keyword density, so this is something that I wanted to begin testing. We started off. We actually didn't put a thousand words but instead 990. I’ll tell you why later...
You have an article of a thousand words. How many images should you put in it? You know the theory would be at least one, of course, but is that enough? Is there threshold? If so, what is it? You know this is the stuff that I wanted to know. Just like the number of words or the number of keywords, I feel like this is pretty important. After this one, we're gonna do a number of videos or all kinds of different Media types,.
Here are the pages that we did. We used 990 words instead of one thousand because it didn't look as obvious as being an article ordered somewhere. Ten pages with close to a thousand words, so it's really easy to calculate image density. So this is what the pages look like: He put them in and used all the same images, and that could be something that we need to also test as well; different images. But this test is a beginning test into this concept. The images are all the same and the paragraphs are basically the same like in the same spacing and then the rest of the text was hanging down here.
Now, that's something else that could be tested. It could divide it up totally equal. Like, if there's three images on the page, it has to have equal text between. In any case, here are the sites, and here is how they are numbered. The page with 4 images climbed to the top right out of the gates, and then it shuffled a lot. The page with two images was there a lot and we thought we had a clear winner and then, when we were gonna call it, the page that had four images moved up and settled into that spot.
When we were writing up to test, this was the last image that we pulled and it was going to be like this looks pretty good. It's four. But that's not the case. We kept on looking at it and it's still shuffling around, but at the time it appeared that it was four and it was also coinciding with what was on Google Mobile too. So the desktop and mobile were look exactly the same, and that was really cool, but I got ta say... it's inconclusive.
Well, unfortunately, it's still bouncing around this. This column right here was when they called the test and when the tests were written up. Then these were where they were checked afterwards, so you were, we can say in the analysis, you know maybe two, four, six images every for every 1,000 words. Certainly not 7 or one because they were really down here, but I think this needs to settle a little bit more and we'll continue to update the leaderboard with this and do some other testing around images, because it's really something that I want to nail down as a best use like a sweet spot for keywords. I'd like to have a sweet spot for images and videos, and things like that, so we'll be continuing on on this one.
Test 338: Topical vs Non-Topical Linking?
Next, test 338: topical versus non topical linking. Now this is something I have wanted to test stuff like this for a while because I have a beef with it. So my background information is I've spent my career building links and linking non topical links and ranking them, like all day every day for years. So when I hear someone say, oh, it has to be a topical link to have any results. You know goes against all my experience that I have. I have to tell you, it boils me up a little bit. You know just because it does. This is the beginning of a deep dive into topical versus non topical. So this is a basic test that we started with.
We chose a set of four pages that we gotten keywords index for. We just went with Greek and lorem ipsum. I even wanted to have lorem ipsum text or content on the page; lower than Greek. I was giving Greek a handicap.
These pages have lorem ipsum, that's our standard. And then the five links going in from the five pages that were created on the same site so these are internal links. These five pages had Greek on them using the same keyword to link to them. These five pages had lorem ipsum on them, using the same keyword to link to it.
Not long right after the tests begin, Greek went right to the top. I'm like oh yeah! This is awesome, loving it, loving it, loving it! Lorem ipsum in second, just like I was hoping it was going to stay. But it didn't stay that way. Day 2, lorem ipsum was making a shift - and we got a ton of other pages that indexed. It's great for the indexer, but it was like dang it. The results are that lorem ipsum is on top and the Greek is second and it stayed there. M
The one thing it does show us is I'm going to deeper dive on the topical stuff and doing PBNs and things like that. This was straight on interlinking type of test with basic content. The only difference was the content. But you know what you can see, you can see that links make a difference, no matter what type of page they're on. So that's that made me happy still to be able to say that you know what links are still the bomb. In any case, what we'll dive deeper into that.
Test 339: Does adding a tag page with canonical act like an internal link? Okay, test 339 does adding a tag page with canonical act like an internal link? This is by Alex, and this is part two of a test, because he did this a while ago. He created a canonical and it started moving up. He used it as a tag; did a tag here and he used the exact match keyword.
He did one tag got the pages recrawled and it moved up to the second spot. This is after six weeks. It moved up one spot to the second spot. He sat on this for a while and eventually it moved up to the number one spot. So, yes, results are in after four months. It's still there. It seems like a tag page, a canonical to a page with the exact match keyword does act like an inner link and will move it to the top. So that is a really cool thing that we can use when we're doing different things. It’s very white hat as far as I consider it.
Test 340: Do sentences per paragraph length matter?
Test 340, do sentences per paragraph length matter? So here's another one that I thought we hadn't tested before. We've never tested like shorter paragraphs and we know that shorter paragraphs are easier to read online. It's different than when you have an ebook online. Bulky paragraphs are harder to read. You just don't want to sit there and read it for some reason.
In any case, this could be a secondary type of ranking factor, but I wanted to know if it was maybe a real ranking factor. So we created six pages and got them indexed. We cut one of the pages to 2-3 sentences per paragraph where the other five pages had 4-7 sentences per page. All pages had a total of 16 paragraphs.
Now I know what you're gonna say and what you’re thinking because it’s what I was thinking too, when I saw that. Hunter did this test and - and I was like oh wow - he pulled content out of it. So if we have any movement, that's going to be pretty cool because he pulled content out of it. All these pages have the same amount of content, which is, as we know, pretty big deal when you're doing testing.
It started moving. It moved up one spot and then it started going down and again it bounced up, but today it is still it's back in the second position, moving up just one spot. So the analysis is even though there's no huge bounces and rankings with any of the pages... and we believe that the word count mattered more than sentences per paragraph.
Which I think happens in a lot of tests that we do. So the other page had double the amount of words as the test page. So it's gonna be interesting. We're gonna redo this test and have all have the same number of words in the same amount of content. Just have the paragraphs being shorter and we'll see what happens. I'm kind of optimistic that that maybe this might not be just a secondary ranking factor - that it's easier to read and people stay on the page longer and click. Possibly Google likes it too. So we'll see about that.
Test 341: Google Analytics - Is it a ranking factor?
Test 341. Google Analytics - Is it a ranking factor? It's by Jo! Well, thanks to Ted, who did the did the tools. That's the hypothesis. Actually, the hypothesis is Google Analytics is not a ranking factor. I think we've actually tested this and found that it wasn't. So background information is having Google Analytics will add an extra connection and open your site up to Google. Will this add a trust factor? Perhaps slow down the site? Or perhaps just do nothing.
Google Analytics is growing with the internet, giving another source of data to Google. Of course, it does give us a source to Google. For a long time, us blackhatter's and people, especially like PBNs, weren't using Google Analytics at all. So we were afraid of Google Analytics. I haven't been too much recently because I don't see that as a correlation of de-indexing or anything of devalue.
In any case, he says, does it have any bearings on rankings? It really shouldn't. But it's best to run some exploratory tests and see. So the test went up and he's got some correlation stuff down here: 7 control pages with 3 test pages. Control with no Analytics tag. Test pages with analytics tags. They were then sent to the indexer and given time to settle and results recorded. It was repeated for a total of 10 times.
Test 2, for the second test, we've blocked out the control and test pages. That's interesting. Seven control with analytics tags and three tests pages with no analytics tags. They were then sent to the indexer given time to settle and results recorded. It was competed for a number of ten times, and here you can see his data and gives us the overall mean and the results.
He reports is inconclusive, even though you would think that you would get something like this. It wasn't conclusive. He says: there seems to be a very small difference here, but you would need to run this over a lot more sites to see it. I would say that there is nothing major here in this test. The fact that having analytics upped the mean in both cases is not worse than looking over a much larger data set at best it is a tiny factor.
You'd expect the test pages to be clearly sitting at the top or bottom to really give a good signal. And yes, if there were, then you know that there's something worth investigating. Google most probably has worked hard to keep things fast and impartial when it comes to using their service. All hell would break loose if sites using Google products got a boost.
Yeah. Perhaps repeat this with Google tag manager or with Google Webmaster tags. So that's that - and I do believe that we tested that before and didn't find that it had any ranking evidence at all and yet and it would be a crime if it did. An absolutely sin on Google's part. But they have sinned before and they will do it again.
Test 342: Will an HTML page rank better than a WP page?
Okay, a couple months ago we tested whether Joomla or wordpress was better. It ended up that right out-of-the-box joomla was a better platform than wordpress. We talked about would we go to joomla just for that slight ranking factor? Probably not for the ease of use of wordpress. So here we wanted to expand on that and we wanted to test HTML pages of against wordpress pages. So that is the background.
We set up six pages and then we got these indexed and created an HTML page along with that HTML page. When they first got indexed, it was way down here and it's stayed there for a while. It took two days, it went to another second, and then it went to the top and has stayed there ever since.
So does this mean change and go to HTML? No, it’s just the knowledge that we know that if we're doing something fast and dirty, you want to rank it maybe use HTML. Now this test was also done in reverse. I just don't have it in here and the same thing happened. The HTML went to the one spot. HTML pages are really quick to put up. I think I'd like to investigate it a little bit more. Do it in an environment that has more than four pages, although it started out with seven pages. But maybe do a twenty page test and see really if it goes all the way to the top. That would be pretty clear.
Something else to think about is making your PBN's HTML, especially, if you want to rank them. So Dori do a test - Testing, HTML PBNs vs. WordPress PBNs - great idea. We're going to be doing some more extensive tests on PBNs in the upcoming months, so that was a cool winner.
Test 343: Does BulkAddURL service work in bulk?
Here's test 343 does BulkAddURL indexing service work in bulk. So if you were here last month, the test was done and the background is, they did 25 pages and it indexed 76%, which is really great, especially since the indexing problems that everybody's been having lately. So what I wanted to do was see what if it would hold up when we submitted more pages. We ended up submitting 534 pages to this service.
We use the VIP service cost a penny more. So it's three cents per URL rather than two cents. It basically just cuts you in line. It doesn't do anything different. The regular people only get indexed after all the VIP people get indexed and that's it. We started submitting in mass now. We still did eight to ten pages, because we knew that at 76 percent, we were still going to need eight to ten pages to get that five page testing.
We were actually really happy to get a lot of ten pages indexed right off of that. So the results were on a third day we counted how many pages were indexed which was 374 out of 534, which made it a 70% index rate. So I would say that held pretty good. Definitely going to continue to use this now. That only cost us $16, and it makes a lot of sense for somebody who's not indexing thousands of pages to use an indexer like this.
Now next month, we're going to test Colinki, which is also a bulk indexer, but it costs $49 a month. Now you get to use you get to index 50,000 pages for that, but does it justify when you only need 500 pages indexed? Which now we have a lot of test sets already incubating, set up and ready to go because of this. And you can see after the third day, it's changed.
It's actually better because in the last results that I was showing you, we started with five and all sudden two more would get indexed, and that happened a lot. So we have a lot of test sets that have eight to ten pages. Which is really exciting. I think for me anyway.
Q&A
James says he has a client with WordPress site and added some HTML pages good and the HTML is ranking much higher. All right on James. I love it. One that gets confirmed.
Susie asks did you consider the speed in that HTML test? We didn't time it, but yes, we actually even wrote that speed might be a factor for that, but in any case it is a factor.
Let's see image density - does it depend on competition, not enough competition to Google settle for one number? That absolutely could be. In anything, if you want to do your competition analysis. I know a lot, these tools now will tell you how many images they're to put up based on your keyword and the competition in those top 10 or top 100. Do you have like a label in a generalization like we have? You know with pure density and stuff? I just kind of want to know what that generalization is without having to do a competitive analysis with a pool.
Test 344. What is the best domain extension to use?Test 344. What is the best domain extension to use? So, okay, the basic on page course I did this spring... I thought it was pretty interesting so I included it here. I’ve done this test a lot of times in my career because I’ve bought a lot of domains in my career.
In the beginning in 20001 for 10 years or something it seemed like the .org had always won. Just it always won. I was even an affiliate with Brad and Matt Callen and we were doing the Amazing Selling Machine. We bought a bunch of domains in the .org and we actually over ranked everybody using that. But we did a lot of different things. But that's still one.
A couple years ago, I did test it again and in a lot of different ways and I used five major extensions .com, .net, .info, org, .us. I totally got inconclusive results, which happens. It's kind of like the image thing. It does come up inconclusive, which was okay and actually .us was on top and .dot info was on top a lot of the time. But it wasn't there enough for me to say: Oh wow, that that is the extension I'm gonna use. I couldn't call it.
So the test set up… since there's so many different new domains out there I did go beyond the 5. I even got a .health. I did the keto diet beginners. This would be an affiliate. I really thought that this one (.health)here would would out beat everything. Because of the keyword in the URL, not the keyword, but the topic.. This is where I go into that topic stuff too; topical. Keto-diet-beginners. Keto diet is a health related. So that's where I'm like, okay, this is topical. This is going to win and all the results totally surprised me.
First off .tips and the .io were the first to index. Then the.org and the .guide indexed . The dot org was on top and I thought YEA! It is again. Here we go. It's just a rendition of years gone by with the .org. But then the .org disappeared. The .us appeared on top and then we got others index. The .net came, the .info came, but there was no sign at all of the .com, the .co or the .health, which is the one I thought was gonna win because it was health-related. We also lost the .io along the way. We never saw the .health, .co or the .com.
Here's another. You know the .us. .guide is in here a lot. The .guide held in there the entire time and it's here still today today. I checked it and it's still there. Now the .me is on top , but what does that mean? I mean here: here's like two pages worth the results, because we're using these sites as test sites.
Now so the .me is a test site. The .co is a test site. The .net. I don't know if we used it yet, but I accidentally clicked on the.net, so I don't know if that's why it's above. But I have to give credit to the .guide. I mean, if I'm doing a Keto guide, I think I might just buy the .guide. Because I’m doing the pipe ends in between, it's not a branding thing, this is obviously an affiliate play.
It sounds like the .guide might be the thing, especially because this would be a guide for the keto diet beginners, which I wouldn’t have known if I haven't tested this.
Does that mean I'm gonna go out and buy .guides for everything? Hell, no. I mean I'm still gonna go get a .com for my projects, or any of my projects that are branding. I want the .com version of that, but you know for an affiliate play or something really quick, use HTML and by the.guide. I'm even completely sad that my .org kind of went away now.
I want to also comment on these things are still indexed. The.com is indexed and the other ones that didn't show up for the keyword, they are indexed. If you do a site:domain search qualifier, they do come up. They just don't come up for the keyword that we were testing for. So what does that mean? I don't know, it's kind of strange. Would it happen every time? I don't know that either.
We'll have to test it again but that would not deter me from buying my .com or .org for my branded sites. I like the .orgs because of my previous ranking experience with them, and people seem to think that they are information and people go to the internet for information. So to me, it's like a secondary ranking factor more than a first. That is that on the domain stuff. That’s actually it on the 10 tests.
Info
We're probably gonna do these from now on on the first Thursday of the month. Instead of the last. We had that big webinar last week on Thursday with the NFG guys for SEORockstars. It ended up my wife and I got the flu.
Our next month, we're gonna continue doing the deep dive on the topical versus non topical linking. That is just that's me getting out my Yaya's. I really want to know plus you know in my PBN business, it’s just really important and I got an axe to grind on that.
We're also going to start out the test on press release services. Simon's gonna help us with that. He has press release services with a different couple ones so we'll be doing different things with that. We're gonna start that up we're gonna come back on the image density and start on also video density too. And see if there's anything there more on sentences per paragraph.
We’ve got tests in the popper, for you know, age domains versus vs. new domains and where to place your table of contents and we're going to do a deep dive into PBNs and linking in general, so that's all coming up in the next month. Oh, we're also testing Colinki to see how that works as an indexer.
Okay, that's it. We'll see you next month. We’ve got tons coming next month. All right, Dori out. Thanks, guys. Bye.