We have 19 tests for this month and it’s the 4th year of the SIA. For this webinar, Clint discusses the results for all the 19 tests and different applications for the results.
Our tester James Gorham comes up with 8 tests for this dossier. He tested if repeating the whole article In schema can boost rankings, if linking-out to a .Gov site and .Org can increase rank, if HTML comments can negatively affect. He also did a retest on keyword placement in the title. He tested if the amount of YouTube Subscribers/Views has an influence in rank when the video is embedded on page and if using Javascript on a page has an effect.
Another tester we have that came in with 8 tests is Boryslav. He tested if there is an effect on rank for text to html ratio. He also did a series of tests on indexing tools at to help us cope with Google’s indexing issues. He checked if HTML sitemaps are effective. Some of the other tools he tested are GSA Indexer, InstantLinkIndexer, and Speedlinks.
Which is the best to use? He also tested again which is better to use <strong> or <bold>? We’ve tested page speed and found out that it is not a ranking factor. In this iteration, he tests if loading time can affect rank. Lastly, he did a retest of keywords in h3 and h3 tags and their affect on rank.
Does Twitter Moments have backlink value? Should you be using them? We also had the continuation of the test on the optimum number of h2s on a page.
It’s been a few years since our Word Count test where we found 1300 words to be the sweet spot. Are there any changes in the data? We test that out again.
All those tests, we have in store for you for this month’s webinar.
Tests this month
Test 422: Does Repeating The Whole Article In Schema Boost Rankings?
Test 423: Does Linking-out to a .Gov Site Increase Rank?
Test 424: Do HTML Comments Affect Rank Negatively?
Test 425: Retest – Keyword Placement in the Beginning/Middle/End of the Title
Test 426: Does Linking-out to a .Org Site Increase Rank?
Test 427: Does the Amount of YouTube Subscribers/Views Influence Rank When Embedding on a Page?
Test 428: Will Embedding a YouTube Video in a Page Boost Rank?
Test 429: Will JavaScript on a Page Affect Rank?
Test 430: Twitter Moments Backlink Value
Test 431: Word Count Test 2020
Test 432: Text to HTML Ratio as a Ranking Factor
Test 433: Indexing With HTML Sitemap
Test 434: Which is Better, Stong or Bold?
Test 435: Does The GSA Indexer Work?
Test 436: Testing InstantLinkIndexer
Test 437: Retest: SpeedLinks
Test 438: Is Loading Time a Ranking Factor?
Test 439: Retest - Do Keywords in H2 and H3 Tags Affect Rank?
Test 440: Does the Number of H2s on a Page Affect Rank?
March 2020 Dossier
Transcription Below:
Hey, everyone! How are you doing? Welcome to the March 2020 dossier. From all of our tests, we got 19 tests. We did 20 tests but one of them I want to save because I think we can expand on that concept... The idea behind that test a whole lot more and turn that into its own unique thing that no one else is really doing or looking at. And then depending on the color of your hat, you can implement it and get a lot of stuff going on. That will be coupled with my schema tests.
I did a whole bunch of those. Got them all indexed and stuff and finally got the results. Couple that with the other tests that I just mentioned or talked about or hinted to will give you a lot more insight into the Google monster. Some things that you can and can't get away with and then give you some insights and why or why not you would want to apply this stuff. So I think I'll be really, really good.
I would be remiss in saying I hope you all are doing okay during the worldwide Covid lockdown as it were. If you want my own personal advice from my own stance with it, I would say that on a personal level don't be afraid of Covid. Fear is reasonable even I have a reasonable fear or a reasonable respect for the fear that comes along with the word pandemic. You're watching the news and everything, you see what's going on and the worst case scenario.
Those are always highlighted and stuff. At times, that will get you down. So if you have to shut off Facebook, shut off Twitter, get yourself a group of friends you acknowledge that we just want to talk about friends stuff and the rest of this madness... Let's just leave it alone for a little bit. That'll be good for your mental health.
And then on the fear side, don't be afraid to go out and do stuff. You gotta go to the store, you gotta get gas, you gotta get things fixed, you gotta do XYZ, just be responsible and follow the recommendations. I think everyone's going to get through this a lot better.
Business side, we did a poll on our show yesterday on SEO Fight Club… and I bring this up because it got a lot of negative reviews, which is weird. But I think from a marketing stance, marketers are shaming marketers for being marketers. And that's really sad right now. That's how we make our living. It just calls on us in times like this to be more socially responsible, respectful, and not to be ignorant of the situation.
But to just understand, "Hey, this is going on. I'm locked down right now. My business has to close right now, but it's not at the end of the world. And I have to think about what am I going to do if let's say the lockdown lasts longer than 15 days. Let's say the lockdown lasts a month, how am I going to save my business? How am I going to pivot to save my business?"
And if there's just no way to take customers right now, maybe there's a way that you can pivot and adapt your marketing so that when you can turn to customers they're already there. They're already in your appointment book. They're already ready to go and to pay you money.
Like I said, it's time to stop bashing marketers for being marketers. It's time to stop bashing each other. It's time to stop looking for the free stuff. There's plenty of people offering free SEO courses and etc. I can tell you, the ones that I've seen... Like 15 or 20 of them... They weren’t worth the money to be paid for nor were they worth the money to be watched for free.
Just be careful and let's start with thinking about ways to push ahead. Every instance of bad luck provides an opportunity for some good luck too. As long as you keep that in mind, you are going to be great.
Okay, let's go with getting on with the dossier. Like I said, we have 19 tests. Our testers really stepped it up this month. I appreciate them: Boryslav, James, Jo and Hunter knocked it out of the park. We got a nice set of tests. Hopefully, you get some value out of this stuff and value for your time spent on lockdown. Maybe you can do some testing of your own. We appreciate that and we’d more than welcome you, as members, to come in and offer your tests results to continue the fight.
Now, the first test is test number 422: Does repeating the whole article in the schema boost ranking? I don't know if I've ever talked about it in SIA with Kyle in a Tuesday Chats but the application that the schema article is you can use as articleBody section and put your whole article in there. The paragraph tags, all that stuff. It's essentially wrapping the entire piece of content inside of the schema.
Is it necessary or is it not? I'm not too sure but the test here says, what if I had the article and then I also put the article body inside of the brackets here, which is perfectly allowed within the rules of the implementation of schema. Is it going to increase my rankings? Jo knocked that out, he set up the ten test pages.
You'll see that a lot of us are transitioning over to this ten test pages/ten sites thing just to verify that our theories are right regardless of the tool that we're using. Jo uses his own tool. Boryslav has his tool. I use Ted’s tools. It's two thousand words with 3% keyword density across the test. And then the H1 tags on all of them and then each page he included the schema. When he said each page it is ten sites across the board for ten pages. Then some are controls and some are tests.
In here, you see that the pages with the schema where they ranked and relative to the controls if this test was a success like it was a definite factor everything would be all T’s up here in the 1-2 blocks. In this case, you can see it’s scattered around so his conclusion with that is that it's not a factor in the way that it's set up. It could be because the way he did it, he messed with the keyword density. Well, not so much messed with it but he essentially took a 2,000 word article at 3% and made it a 4,000 word article at 3%. So it didn't change the density at all and that could have somehow implicated a little change in the test.
The other thing that you could test with this is to ... What I talked about originally... Wrapping the entire article in the schema and just leaving it that way. Beyond this schema setup right here is the basics. This is the limit, then the minimum that you should be doing in this schema. In my test you'll see next month, proves that those things are pretty much useless.
The basic schema, the one that Google says to do in their Google documentation, the one that 99% of the plugins are doing right now is completely worthless. And I tested that and I proved it... We're going to show you how to change that up inside of a couple of different webinars, courses and stuff. So this test is really, really timely and I appreciate it. And then he comes up with another idea for testing there. So that's test 422.
Test 423: is Does linking out to a .gov site increase ranking? This is a James Goldsmith test. You've seen these tests before. We used various sites, various types of sites, various domains, etc. The test setup is the same. He's using ten test pages. I'm going to skip a lot of this, you can read it inside of the PDF yourself. But most of this is just images to show the progression of the dates and where the ranking went. You can see that it’s pretty much stayed the same.
So the linking out to a government test or a government domain in and of itself is pretty useless. Could there be some applications where you link out and then you put this in the schema as a sameAs or a mentions? I'm going to have to say as a hint that yes, that will help. You just need to play around with these and combine it up. But in and of itself, people tell you should link out to .govs and not .orgs or .orgs and not .govs but that’s not true. That's good to know, so thank you James for that test.
Test 424: Do HTML comments negatively affect ranking? This is also from James and there's another one in here. I believe Boryslav did one as well where he tested HTML, just like having HTML and of itself. If you don't know what HTML comment is you can comment out HTML. You see this here – this two brackets, exclamation points and the dashes, that essentially tells the browser "Ignore this part.
This is a comment". You can put it in a software. You can put it in a code. Well, software and code are the same thing. But pretty much any coding language you can use these comments sections and everything in between it won't be shown. That's the idea behind that. Then, at the end of it, there's another one which is essentially like that end tag for the comments.
Some really good examples, if you're looking at your Google Analytics when we implement that it says Google Analytics and you see that section there and it's a comment and then it closes it out and then, "Hey, this is where it ends." you can find that inside of your implementation. So, he's looked in and he wanted to know if I add those comments, would that going to increase rankings.
As you bounced through and looked through this you see that it goes up from two to three bounces back and forth. Really did not have a solid effect in the implementation the way that he did it. His conclusion is pretty much the same as mine. Just looking at test results is that it had minimal effect. It didn't hurt the website and it didn't help the website.
But what if we change that up a little bit and put some things like entities in there or maybe some LSI? So, we tinkered with the keyword density. I'm not sure if that would work and I think that would be a good follow-up test to look at in this case. I'd love to hear everyone else's thoughts on that.
Test 425. This is a retest. It's keyword placement in the beginning, middle and end of the title. This is from James and what he tested inside of here is the middle and the ends that can help you out. The beginning was tested naturally just because that's where it was inside of his setup. So the keyword is in the beginning and he just moved it around and apply the – and moves the keyword from here into the middle and there to the end. What happened was surprising.
The lead keyword in lead of the title tag is a correlating factor inside of a lot of terms and you could see that running a software like Cora or Surfer or POP lead the title off with the keyword or lead keyword matches (what it says in Cora) but in this case he got it bounced it up to one and two with the middle and putting it in the middle and the end. Maybe that's a transition that Google was going towards because in SEO, typically, he's going to put it in the front, right?
So, if I am trying to anti-SEO the algorithm putting it in the end, in the middle and giving it some added weight is certainly a possibility. What I would invite is this test to be run across ten sites now instead of just the standard one here and see if that's continuous across the board before we declare it's a thing. But you can take the recommendations of James here or just look at these for yourself and say that putting the keyword in the middle of your title tag is probably a good way to go. You do see a bounce here a little bit.
At the end of it they went right back up again. So, the keyword in the middle stays – it's pretty consistent. I would say, moving the keyword into the middle and testing that across ten sites and then we can probably call that a little bit of a trick that you can apply and see how it works especially if you're stuck. If you're like at three and you definitely want to get into one but you don't want to remove the keyword put it in the middle or somewhere and see how that goes out. That would be pretty cool to do.
James recommends that use it on the end too but the pattern here wasn't as consistent with the middle as it was with the end. It kind of dropped down then went back up. Dropped down and went back up. Whereas the middle was consistently the top three. So, I would go with the middle first until we can test across multiple sites to see if that's something that's going to continue on with the pattern.
Test 426. This one's from James also. Does linking out to a .org increase your ranking? I will go ahead and skip all the images and stuff. The setup is the same. All he did was put it a .org link.
I will note that the images say .gov but it's .org. I had to verify it for myself just to make sure that it was just a typo. No big deal.
You can see that the .org pretty much reacted the same way as the .gov did, provided no benefit whatsoever. I won't even say it even hurt the site. A one drop is just probably the randomness of ranking of the pages and Google sorting that out but in and of itself adding a .org will do nothing for you. I don't know if you were following around with these.
Hopefully, you were but those sites are designed to see if there is a benefit of linking to you one site versus another or one type of domain versus another to see if that helps you establish authority.
That was from a conversation where we're talking about E-A-T and everyone says this is what E-A-T is but Google says we don't have an algorithm for E-A-T. But on the flip side, they say, "But we have certain things that look for aspects that could be considered E-A-T".
So, these tests here – the reason you want to know this is because it helps you find the algorithms that are looking for those aspects of E-A-T. That's pretty interesting and I appreciate James for doing this.
Test 427. Another one, does the amount of YouTube subscribers influence the ranking on a page? 10 test pages and what he did is he picked two YouTube videos. One is the SEO Fight Club video and this one had a limited amount of subscribers. I think we've got just 3,520 subscribers. The video gets an average of a thousand views.
Then, we picked on one that was a high one. Like these guy is 5 million views. The subscribers ~ out in the screenshot, unfortunately, but I'm going to go with – if they get 5 million views they got way more subscribers, right? Then, he put those on two pages so there's lower sub and views and higher sub and views. They want to see what would happen.
In theory, if the hypothesis is correct the higher the subs and the higher reviews the more credit the page gets. These would all bounce up to the top. That didn't happen. They separated a little bit and then mingle - the lowers, highers, lowers, highers. Same thing here - lowers, highers. Kind of sorting itself out a little bit then it got all jumbled again.
You can see that across the whole pattern there's a higher right there then a lower, lower, lower, higher, higher, higher, higher, higher. It's really hard to say that having a higher sub – because it ended up this way. Technically, I could sit here and theoretically just continue to watch it over and over and over again until I get the result that I wanted but I think you see when it bounces around as much as it did it's hard to definitively say that higher views and subscriptions make that video more valuable to you ~ prove your rankings.
That's the same conclusion James that – I think, honestly, at the end of the day we probably wait all day long and these will continue to bounce one way or another and I don't ever suspect that they would sort out just as the way it is. Really interesting test. I like seeing the test with the YouTube stuff.
Test 428. Here's another one. Will embedding a YouTube video on a page boost rank? Just by having the video there. Pretty much the same setup: ten pages, drop the video on there and then he embedded it inside the test. As you see, it goes around, drops a little bit, goes back up. Just bounces around and didn't really do much.
He came to the same conclusion as I did. Just having it is not going to help you with rankings. However, there are some things you can do inside of the iframe that could potentially help with the ranking.
I'll look at maybe running a couple of tests on those and then I'll teach you that process of how you can do it. I think most of you might know it but you can put schema inside of that iframe. You can put your keyword inside of that iframe, etc. And help increase the relevancy of why that embed is there. We'll look at that and do a test for that. It should be pretty cool to knock out.
Test 429. This one is also from James. Will Javascript on the page effect rank? Remember the HTML one, this is HTML ~ . This is continuing into that with adding Javascript. He's adding Javascript. This is the example of the Javascript he added onto the test page and how he did it. He just used custom fields and applied it on there and then he added it to the page. You see there in the middle, it went up a little bit, it went up to the top and then went back down. Way down and it kind of went right back to where it started.
I think the conclusion would be that that extra code didn't help. This is actually backed up by the test that we'll see here a little bit from Boryslav that did pretty much the same thing. He just used different ~ things. So, having it on a page does, in and of itself, not affect rank.
In the format that he gave it can we do some other stuff? Can we add some keywords to it? Can we get some more relevance to it? We don't really know yet but that's another test to go on with this one. It'd be really good.
Test 430. Here's one that I did. One of the members is talking about Twitter moments and that they provide some backlink value. If you don't know, essentially what you do is you create a tweet and then you click on moments then you can create this moment and you can put a link in the description of it. Then, you can add your tweet to it and the links inside of the moment are actually dofollow. Plus, those moments indexed. As a matter of fact, they indexed a lot faster than my test pages did. The idea is because it's a dofollow link and it is from Twitter you should get a lot of rankings piece from it. I wanted to know and I tested it up.
I set it up, set it across ten sites. I did 10 tweets, 10 different individual moments etcetera and you can see that the average ranking kind of fell, stayed there at 6 as of the 20th of February. Then, you look down here on the 29th when I checked again. It went up a little bit more but at the end of the day they beat out all the control pages and even one of the pages fell out of the index.
Looking at this can be a little disconcerting because it's the average rank of 5.89 but when you look at it there's a 4, there's a 3, there's a 2 there but 5, 5, 5. So, in this case, especially with the way that this tool, you have to look at both those as ~. I would have liked to see more 2s in here and then probably could have said definitely even if it wasn't beating all the controls. 2 is not that bad so there's something. But with all the 5s in there and then 10s and then the one page fall out of the index I would have to go ahead and say that this link, in and of itself/by itself, there's nothing for you.
There are two more tests that I want to do. I'm going to test sending backlinks to that Twitter moment and see if I could pass them the juice through that and I'll do one that's clean. I'll do one using something like a GSA then just hammer it and see what happens. Then, another one is stacking those.
So, create a Twitter moment and then create another one and have that second Twitter moment link to the first one and just go along with that and then see if you can use the internal juice or Twitter to pass up power to the source page. That would also be a good one and knowing that why would anyone want to do that is if you're in a market where you just – you can only just get so many links. There's only so many people that want a link to a plumber. There's only so many people that want to link to an adult site or a gambling site or a Bitcoin site.
So, you gotta have some alternative methods for link building. If we know that we can create some Twitter moments those are clean tier 1s and then perhaps we can ~ to start hitting those with some GSA or something like that. That's an opportunity there that most people aren't doing and it's scalable so it doesn't really matter what your niche is. That's why I think this is really cool.
To sum, the link by itself is really not doing anything for you. It's just another good tier 1 link. It's not good, it's not bad. Just a tier 1 thing that has no power right now. But, hopefully, the other two tests, once I get those set up, will show us a way to turn those into something very useful.
Test 431. This is a word count test from Hunter. If you remember, SIA did a test a long way back and I think the number was 1300. It was the magic number for the amount of keywords that you should use and we wanted to look at that again. I had Hunter set this one up. We used 250, 500, 750, 1000 and continued all the way to 2500 and created those test pages. Back then, we're using bulk indexer to index and get that going.
At the start point here, you can see 1, 2, 3 and these numbers here and the image correlates with keyword number. It started out with the 250, was winning and then everything else was jumbled back and then it bounced around a lot more here. I think if you look at this, like the 250s, it went down to the bottom the lower numbers and then they get the 7s and the 10s.
Actually, I think these things are messed up. Yes, these are messed up. This image is the end, this is the result and then this is the starting point. So, something got messed up when they moved the PDF around and change the images. That's just how – Hunter writes a test and I read these.
So, this is the result. 250 and the 500s were 1 and 2 and then the rest were kinda right around behind them and then this was how it all started out. So, you look at it you're like, "Oh, the more content the better, right?" that's the way it looked but then Google sorted it out ~ and said, "Hey, the 250s are just fine."
Here's an implication that I wanted to get out of this one is that you can write a 250-word article or you can write a 5,000-word article as long as you're saying what Google wants inside of the algorithm for that piece of content to say. Typically, nowadays, I ignore the word count inside of Page Optimizer Pro, Cora and Surfer.
I don't pay attention to the word count anymore because I've proven over and over again in my own real-world application of SEO that I can rank a 500-word article just as well as I can rank a 3,000 especially if I'm saying pretty much the same thing in a more concise manner, which I think that Google appreciates anyway and certainly the reader does. So, if I could say something in 500 words why would I write 2,500 to say that, just because everyone else is writing 2500 words.
If you enter the application of lsis, entities and stuff you're going to increase your word count a little bit you're also increasing the relevancy of the piece of article making it even more super relevant for the keyword that you're targeting to the math. Right now there is no set word count. If you can say what you have to say in 250 words then go for it. There's no reason to go out there and write another thousand words just because one of these tools tells you to.
Test 432. All right. Let's get into the Boryslav test and this is the text to HTML ratio. If you've been using tools like SEMRush, their site auditor – I'm not sure if Ahrefs is adding it, but I've seen it in a few of them where HTML the text ratio is labeled as a concern.
In this test, what he did is he created these standard ten pages and he adds a lower amount of HTML, added extra to one and then a high amount to the other one. What he did is, essentially, he created some space. In this case, he did some white space and that was pretty much it.
You can see the results, the high to low and the high just stayed there in the middle across all over his tests. So, he came to the conclusion that he didn't see an extra boost from the code and he didn't see from not having it. He didn't get penalized. None of those test pages clearly did get penalized for having too much nor they get a whole lot of benefit out of seeing it or having too little.
He made the same ~ that I probably would've if you have too much extra HTML code inside of your pages. You could be hurting your load time which, in and of itself, might be a negative ranking factor. But at this point, we don't see that inside of the results so I think it's a pretty good thing to know.
If you got a tool and you're running or working with a WordPress theme or something like that, unless all that extra code is slowing your website down it is not hurting you. Those of you who like to beat up the page builders Elementor that adds a lot of extra code to your website and then typically doesn't slow much down. Visual builder is doing it.
Elementor has a lot of code and it loads inside of the head and at the leader your body pages for – pretty much all your pages. You look at it and you're like, "Oh, my god!" but as long as it's not slowing yourself down then you're okay. That's pretty much what boryslav's test here proves, and the ones combined with James as well.
Test 433. Another one is indexing with the HTML sitemap. We're in the SIA testing group. We have a Slack or a Skype group that a lot of the testers are in. If you want to join you're free to come in. The only rule of membership is to have at least one test submitted by the end of the month so we could publish it.
What we're having is an issue with indexing your test pages. Boryslav did a lot of tests this month with a lot of pages so he was pretty much leading the charge in some of these tests. I was doing some of my own stuff and boryslav was supplying them and checking with different services etc.
This one is the simple HTML sitemap. I'm going to assume that you know what an HTML sitemap is but I want to be a little bit careful of it. This is a standard sitemap. If you are looking at an HTML sitemap you get a really big difference. You got the standard code one and the XML and then you have the HTML one where you just listed out, lists all your URLs. There's a practical real world application of that as if you're doing...
Let me see if Lucy still got him on here. He's got some city pages. Let me see if I can find one. I just need to find him. He hides them really well. I can show you another example of how to apply this test. It's not going to POP. I have to run the whole sitemap to see it. But basically, what he's doing is he's got this page, say this is Phoenix SEO and then all the way down here he's using this same thing - a tab or whatever, and then list all those citations in there and all those citations are getting indexed.
Well, this is pretty much the same thing that you're doing with the HTML sitemap. You just say, "Hey!". A user can go in here and find it. Like, I'm going to and I want to find everything from Indiana for this site, for example, I'll click on that. So, you're doing the same thing with Googlebot.
Googlebot appears to be, for right now, favoring these HTML sitemaps over your XML sitemaps. I could be wrong but it might be an answer to those services that figured out how to leverage Google Search Console to do redirects and POP in other people's domains.
So, boryslav went through and did the submission. Five minutes after he submitted as HTML sitemap he used his logs and he saw the Googlebot was crawling it and 24 hours later all those pages were indexed, which is pretty much the same results that I was getting in applications with my testing with my own indexing stuff for test pages. So, this is really good information.
How do you leverage this? You have a Google Search Console and you also have the ability to use a domain to create an HTML sitemap. But in his one he used a domain and then he used a separate, different domain of the pages that he's indexing. So, the HTML sitemap had links to URLs on a different site which is key. In the example here where I used Salterra and it was saying you could put citations in here that's the application for doing it.
You can create an HTML sitemap page for all your citations. ~ now citations and there are your citations on here. Create another HTML sitemap ~ that tier 1 finest now whatever then just POP them all in and you'll boost up your indexing a lot better right now. Google's famous for messing with indexing stuff so prepare for that, for the change but right now HTML sitemap is working so you should implement this right away.
If you don't know how to do the HTML sitemaps get with me in Slack and either I or I'm sure Boryslav will jump in and help you or at least talk you through the process of creating that HTML sitemap. But that is the answer for the indexers that are not working right now. You're not always going to get this result. Sometimes I get three out of ten. Sometimes I'd get all 12 or 10 out of 10 but you can create those HTML sitemaps over and over again.
Another thing you can do is buy some domains and then create HTML sitemap pages on all those domains solely for the purposes – and add those domains to the search console solely for the purposes of creating your own indexing service. As long as that works why not. $15 for a domain, you buy ten of those and then turn those ten into your own indexing platform and you're off to the races.
Test 434. This one is testing whether a <bold> or <strong> works. I did another test, the HTML test and testing the tags. I'm pretty sure I shared it before and these are not even beating the controls anymore. It used to be the strong one all the time and it was something that you wanted to have it beat all the controls but now these aren't even beating controls anymore so they're low level.
But in boryslav test you can see here that the strong beat out the bold and there's probably one or two reasons for it. 1) strong is now the W3C default for HTML. They're phasing out bold. Then, 2) before ~ the algorithm, strong is like strong emphasis. This is the focus word of this page. This is what the page is about. So, when they looked at strong, bold and emphasis and together in that test it's strong one every time so they continue to use strong.
However, if you see a testing tool saying you need 10, 15 bolds anymore so far all of our testing – even this one is pretty much indicating that's crap. Strong beat-up old bold. But, at the end of the day, it's not really all that big of a deal. If you don't want to have like 10, 15 strong ~ words on your page don't even mess with it.
Test 435: Does GSA indexer work? This is a piece of software I bought it myself at the same time. Again, boryslav and a few others were fighting with indexing trying to figure the madness out and then we both picked up this software. I didn't want to see if it was going to work. One, because I'm not going to pay a bulk indexer or Omega to index GSA links. It is just way too expensive to pull that off so he had to find other solutions.
He ran the URLs from his other test pages inside of a demo version which is limited and then someone ran it inside of the full version. As you can see from his log it didn't trigger any Google bot visits but they did trigger other bots. At the end of the day, they didn't get any pages indexed running with that tool, which is not really surprising to me.
GSA makes some good tools but at the end of the day you have to figure out... And they do indexing through link building. Some of the targets are horrible and are not being crawled by Google anymore. That's not going to find the website.
If you remember there's another piece of software called G Indexer. They found some websites were in China that weren't always getting crawled. They are creating essentially HTML sitemaps like this and that was actualy working really well and triggered a lot of indexing. But that stopped because, my guess, the sites were bad and then just stuff that the software... I was paying for it for a while and it stopped working.
This is the same way here with this one. This is a low value software for them. It's a one and done sale so they're not really updating platforms and because the platform is not getting updated these HTML sitemaps that is creating or the links that is creating are not getting crawled thus nothing's getting indexed. If you are tempted to buy that tool I would say go ahead and pass.
Test 436. Next one is InstantLinkIndexer. He ran this one. He checked to see if Google bot was crawling the pages. Nothing crawling the pages and nothing happened for indexing so InstantLinkIndexer is also a bust. I think it was like $17 a month for a million links or some craziness. It sounds great but it was too good to be true and he just proved that. It's a waste of $17. Save that and go buy some Starbucks. If you're using that search you might cancel out right now.
Test 437. Then, finally he did the SpeedLinks. SpeedLinks is on the same lines~. SpeedLinks says, "Hey, we're going to index your ~ crawl it." ~ or SpeedLinks was just doing crawling too. He submitted all those links to that. He got the Google bot to go however nothing got indexed according to his results they're seeing but it did get crawled.
There's two things or ways of looking at this. Let's say you send all your tier 2s to SpeedLinks and it gets crawled. Is the juice getting passed up? Is Google seeing any credit? I say that's why I was such a fan of Colinkri but I'm not building my tier 2s right now so I stopped using Colinkri. But if I would and if I was I would probably just fire that back up and get them crawled or SpeedLinks, you can go and let that around as well.
I'm not really sure if I can recommend not to use SpeedLinks just based off on his results. If they're not actually indexing the pages then it's not providing the service that they said are doing. Boryslav's results weren't exactly clear because he said it was too hard to see if they are indexing or not.
My guess is because the volume of the pages is a little harder to check especially back then even I was having problems with the same tools that normally worked great. The checking of indexing was just horrible. Google changed the code inside of there and messed it up and even the automated checkers, schema ScrapeBox was running like crap back then.
This is definitely worth a good retest but I would look at it for yourselves. If you're running links to SpeedLinks and you're seeing a pattern where your stuffs are not getting indexed it might be time to quit them. I would also reach out to them as well if they're building backlinks to your links in order to get them indexed, if they're building backlinks right back to the HTML sitemap. If that's how they're doing it why don't you do it, right? Save yourself a little bit of money.
Test 438: Is loading time a factor? This is a continuation or probably retests of that ~ as ranking factor. ~ service and ~ faster than 3 seconds and you're pretty much all right in Google's eyes. Then after that you're getting into user stuff. But if you're over 3 seconds then that might be an issue.
He was looking at the load time here. Is that a ranking factor? Here's the test and here's the load time. He messed around with it and he used this code to change the overall load time of the pages. You can see it's all over the place. Here, all over the place. One second, one second. It really didn't matter. The two seconds they stayed down there at the bottom but that could be a UX thing just like Boryslav said.
Here's what I would do. I'll rerun this test and run in at 3, 5 and 9 seconds just to see if we can call it conclusive. If my theory of 3 seconds and faster is correct everything above 3 seconds should be here and then the 3 seconds would be down here, or the 9 and the 6. They should all be down here and then we can say definitively yes or no, Google was giving us a little bit of a lie on that one. It was a good test.
Test 439. Then, do H2s and H3s affect rank? John Mueller, in Twitter, was saying H2s and H1s and all that stuff doesn't matter but every time we test it, clearly, it indicates that. So, he's got H2s and some of his test pages got H3s. They always outrank the control. If it's outranking the controls – if there's smoke there's fire. It's there. So, I would say that disregard John Mueller and start putting your keywords inside H1s, H2s and H3s if you are not doing int and then clean those up and remove the irrelevant ones outside of there so that you stay on that outline format with your heavy text.
Test 440: Does the number of H2s on a page affect rank?
I think there was one more. Whereas Boryslav tested just having them ~ then Hunter tested, can you get too many? Can you have too much stuff? The H2s, there's 5, 3, 9 and there's the control pages and they pretty much stayed the same. So there was no number. If we have to call a number I would say don't do any more than 5. 9 seemed to hurt but 7 was down here and 4 was here and 3 was there.
So, it's really hard to say a number. I think ~ to definitively say don't go over 5, we'd have to do this on the 10-test site format and see if 5 stayed consistent across the board. If it did I think we can definitely say don't go over 5 but, at the end of the day, you can probably have as many H2s as you want.
The other one that beat there is 8, right? So, that didn't beat the control page which is surprising. They all started out beating control pages. It just didn't settle that way. So, maybe it's too many but maybe nine is over that threshold. Who knows what the math is thinking there but this is certainly a good retest. A combination of those two tests.
That's it. That's the end of the tests for March 2020. If you have any questions about that this would be a great time to jump in. I can answer questions on any test. I'll bring the table of contents back in here so you can see it and cite the number or something, to make it easier to answer.
Again, thank you all very much for showing up. I hope you get some value out of this one and remember we have the Slack group. So, if you're sitting ~ you're working you have questions, marketing questions and stuff like come in to Slack. Don't be shy, don't be scared. No one is going to call you dumb.
There are no dumb questions especially when everyone's looking for new solutions. It's good to bounce those solutions off each other. If you're in the mastermind groups make sure that you're in the mastermind groups and you're talking to each other and come in and problem solving. Don't sit there and lose all your clients before you say something. Don't let them all run off before you say something.
There are solutions that you can use right now to keep clients in business or to preposition in a way that when this is all over they turn on the fire hose and rebuild their income very quickly. So, if you need those ideas or whatever use the masterminds, use the Slack. Come in and talk to us. Do not not ask questions, be afraid and just start losing clients left and right. It's unacceptable.
Thank you very much for everyone for being here. It looks like we don't have any questions. Lauren, thank you for a shout out. I appreciate it and I will see you guys all in the Slack group.