In this webinar, we talk about the different tests and their results for the month of May.
Will Google consider two meta titles on a page? How about having two meta descriptions tags, will Google use both? What is indeed the impact of having multiple meta titles? Will it hurt you or will it help you? Does giant sized keywords affect rank or is it just a silly myth? Does Google treat content in quotes and brackets differently or are they just the same?
In the world of links, will sending links from one domain to two different pages on another domain pass juice to both pages? Can a number of links to a page with poor on page beat a highly optimized page? Which one should you focus on - on page or off page?
Lastly, in the world of indexting, does Google index the html source code or is it the rendered DOM? Can you index a page with html classes or ids?
Tests this month
Test 194: Will Google Consider Two Meta Titles on a Page?
Test 195: If You Have Two Meta Description Tags on a Page, Will Google Use Both of Them?
Test 196: What is the Impact of Having Multiple Meta Titles?
Test 197: Sending Links from One Domain to Two Different Pages on Another Domain
Test 198: Can Five TF10+ Links to a Page with Poor On-Page Beat Pages with Highly Optimized On-Page?
Test 199: Number of Unoptimized H2s on a Page (Part 2)
Test 200: Does Google Treat Content in Quotation Marks and Brackets Differently?
Test 201: Does a Giant-Sized Keyword Affect Rank?
Test 202: Can You Index a Page with HTML Classes or IDs?
Test 203: Does Google Index the HTML Source Code or the Rendered Document Object Model “DOM”?
May 2018 Dossier
Transcription Below:
Will Google Consider Two Meta Titles on a Page?
In this meta-title, tags is what we are talking about. This comes from a Skype discussion on some of the tests that we have run the last few times. We have found some cool CRO hacks or something that we could do with meta-titles. With meta-titles, we found that Google will index keywords that are beyond what is displayed. If you have a meta title that is too long, Google will do that “…” but they will still consider what is beyond and you can index a page with stuff beyond that. What we found is that Google won’t display anything beyond the “…” like with the normal character length is.
So even if you have written something that is relevant – so you have written two meta-descriptions and kind of squished them together, Google will not display the relevant part of the meta-title, it would just do the “…”. So then we started talking in the Skype group about what if you have two meta-title tags?
Normally and pretty much in every situation, a page only has one meta title tag. What happens if you put two? I actually did a quick search on this and I could not find anybody that looked at this. The consensus in the Skype group was that it would just be ignored. The second meta title tag would be ignored. But it seemed to be worth a go because if we can get this, then we can write as many meta titles as we like and get the one displayed for the service that we want.
Two separate meta-titles were put on the page, so two separate tags – tag one and tag number two. And then, we tried to index the pages with that. They keywords only exists in the meta-titles themselves. These keywords were not on the page. What we found is that Google indexed the first one but would not index the second one. So in that sense, at least for indexing purposes, Google is ignoring the second tag.
So then we gave it another go where we put the keywords on the page and then tried it with the second meta title and in this situation, while the page would index for the keyword – you can see the keyword that we are searching and there it is with text that is being pulled from the page – this is the first tag, not the second tag. So Google is completely ignoring the second meta-title tag in this situation.
It is not a cool CRO hack. You could not really get any extra benefit for having a second meta title. It would have been fun because obviously, you know that the meta-title is the most important place where you could put a keyword and it seems that Google has really locked down on manipulating extra meta-title tags. They would not index and they would not display.
If You Have Two Meta Description Tags on a Page,
Will Google Use Both of Them?
We know that Google would not use the meta-title tags, but what about meta-descriptions? The difference here is that when we ran on the CRO tricks, we know that we can write a meta-description that is longer than the character displayed. Google has been displaying over 300 characters, they may have gone back to 160 again but it doesn’t matter, whatever is too long.
Google is actually going to grab from that too long part and display the relevant part of the meta-description. The question is, what if we have two different meta-description tags then? Instead of squishing two into one tag, what if we just had two separate tags just like with the meta-titles. This one is more promising because we know that Google will display more – it won’t display more on the meta-title but they will display more in the meta description. Can we use two separate tags, is the idea?
This is how we set it up, we had two separate tags and you can see on the first tag it says “start of meta desciption one” and the second tag “Start of meta description two”. That is within the tag itself so then, if the tags display, we can see that we are getting what we put it. We had to put the keywords on the page for the page to index because we know from a previous test that we can’t index pages with keywords only in the meta-description.
These keywords were on the page and then for the first keyword which is in the first tag, you can see how it popped out, start of meta description one and there is the keyword, but for the second one, it ignored the tag and simply pulled the text from the page because we are not getting that start of meta description two here. In a similar situation, Google is ignoring this second set of meta-description tags.
It doesn’t change though the CRO trick that we have found. You can write multiple meta-descriptions, you can do that for different personas, or you can do that for other keywords that are important to you -just simply put them all into one tag, is the idea.
What is the Impact of Having Multiple Meta Titles?
We know that Google is ignoring the second meta-title but looking back and just squishing a bunch of meta titles into one, what is the effect of that? We know that we can do it, now we are looking at, should we do it? Is this a good thing for rank? Could we get a bonus? Could we write two meta titles and even though Google won’t display it, could we get some extra benefit or extra keywords, is kind of the concept.
This test was done by Ryan Hough. What he put together is a long meta-title that is targeting three different keywords, is the concept. Then the page was indexed. Initially, the page dropped for two of the keywords and went up for one. It was a drop, up, and a drop, which was a little heart stopping. In the end, how it ended up was that two of the keywords went back to the same spot and one went to the very top.
What we are getting from this is that there might not be any extra benefit to adding extra keywords to the meta-title. It does not seem like a bad thing to do, I do not think you are going to get any harm but it is not a super secret way to get some extra benefit. I would still do it in a situation where I want to get a misspelling in, where it is an obvious misspelling and people are using it but you obviously don’t want to put that in the copy, I would put that past the “…”.
Personally and I think that would be advantageous. But completely unrelated keywords that you might be going after the page, that is not going to be helpful. You could not rank the same page for three separate concepts, is kind of the deal here.
Sending Links from One Domain to Two Different Pages
 on Another Domain
Here, we are kind of emulating doing a PBN post or guest post or maybe even like a press release. We found last year that if we did three links from one page or post, we really could not get an extra benefit. I think we ended up running that one five or six times and one time, there was a slight boost but we really could not get an extra benefit by sending three links at once to one page.
What this is looking at is, what if we sent links to the same domain but two different pages on that domain? So we are doing a guest post, we are pointing it at our client’s sites, we have two links in there, what if we point it to two separate pages? Could we get two of those pages to rise?
In this particular fake post, there is a link to this domain going to that page and to the same domain, going to a separate page. The idea is, are we going to get a boost to both pages. What is pretty sweet is, we did. The one moved to number two and the other moved to number one.
Essentially, what this means is in terms of doing PBN posts, guest posts, or press releases, if you have the opportunity to put multiple links in, we know from the previous test that you probably won’t get a boost pointing all those links at the same page but you could point them all to the same domain and get a boost. Let’s say you have two target pages on a particular site, you could point it at both and you are going to get benefit for both pages eventhough they are both in the same domain. If you dabble into PBNs, this sort of thing is for you.
Did the links use different anchor texts?
These were just naked links, in this situation. Cause the thought process was actually from the initial test where the reason they did not get extra benefit is because they were all going to the same domain.
Dori: We used to use different keywords and anchor texts going to the same page from the same post and it would rank for both of those keywords. Have we tested that?
We haven’t taken it there yet but we could do that, we could check that out.
Dori: That was huge back in the day and I still think that it works really well but if we haven’t tested it, I’m not going to say for sure.
We can put that in the hopper, the next to this.
Can Five TF10+ Links to a Page with Poor On-Page Beat Pages with Highly Optimized On-Page?
This is a look at the classic case of links vs on-page. This was also brought to us by Ryan Hough. This was actually part of a different test that he was running and then he had it set up and it actually worked really well to then repurpose the environment that he created to use for this particular test. In this case, this is our target page, this page only has the keyword on it one time and that keyword is actually a link to this page and that is actually the original test.
This page is linking to that test, it was a PA1DA1 kind of situation, the lowest link that you can send and kind of looking at value there. To modify that, what we are going to do is we are going to send 5 TF10+ links to this page. It indexes for the keyword and the reason it indexes for the keyword is because it has anchor text one that page one time but that is it. It doesn’t have any other optimization. It is extremely poorly optimized for that particular keyword.
The rest of the pages in this environment – it is a twelve page environment, twelve different sites that are ranking for the keyword, they are all highly optimized or correctly optimized, if you will, might be a better way to say it. The idea is, will a page that has a keyword on it one time but is now getting decent links, will it beat pages that has no links but are more properly optimized, is the concept.
This right here is the page that we are pointing to, it is actually linking to that page. What is really cool to see in the results were, the page that was getting the link jumped, what had the links were driving links to that page actually got a nice boost and stayed there a little bit before actually dropping back down. So there is an immediate kind of stack boost.
What is interesting is that the page now that was getting those 5 links, the TF10s, it really did not move that much. It was in the 12-13 spot then it moved up to the number 9 so it is just a handful of spots. Basically the good on page pages beat the links, in this situation.
The idea here is that you really need to nail your onpage. You can beat some pages, clearly, and some of the pages in this environment was beaten and that can happen. But what it really comes down to it, when you want to do things the right way, you really need to nail your onpage to get the most boost and then start putting the links in, and then start doing the link stacking techniques cause it clearly works and there is a benefit in there. I think in a more real environment, it would have had a chance to stay if maybe it got some click through rate and stuff like that. Really the bottom line is, spend your money on on-page first then get the links.
Number of Unoptimized H2s on a Page (Part 2)
What we found is a negative situation and what we are talking about are unoptimized tags. These are h1s, h2s, h3s, are what we found that were unoptimized but were just the number of tags that were on the page. If you had more unoptimized tags than the average, your pages would drop. What we found is sort of a negative ranking factor where if you have a page that is just not what it is expected for that particular keyword, you are going to harm yourself by having more h tags even if they are unoptimized, is what we are shooting for.
Now we want to run it the other way. In this situation, we are only putting one unoptimized h2 on our test page and 5 unoptimized h2s on all the other pages. What we got from running this test multiple times, this is kind of indicative, in one situation, the test page dropped. In another, it had no movement, and in another, it moved up one spot – kind of all over the board. So, it is not as strong a signal in the inverse where you have fewer unoptimized tags than the competitors.
What is interesting is because there was a shift everytime we ran this, this is clearly something that Google is looking at. Google is looking at the htags whether they are optimized on unoptimized because each time we are getting a shift here indicates that it is something that impacts rank but when it comes to, are you going to get into a lot of trouble if you got fewer than the average, probably not.
Not as big a deal if you went the other way, if you went over stack your htags then you run into problems, if you are under, the results were all over the page, probably not something to be overly concerned about.
Does Google Treat Content in Quotation Marks and Brackets Differently?
This comes from an SEORoundtable. I enjoy these tremendously when they make these kinds of statements. It was tweeted at John Mueller and Doctor Pete from MOZ and the answer was that there was no difference. John Mueller tweeted his agreement – That matches my “understanding”. I say that he is trying to be funny with those quotes but really, it doesn’t make me feel any better.
If you read the article, they also talk about parenthesis. We are looking at quotes, parenthesis, and brackets. We are going to run this both ways and I started with parenthesis because it seemed like a more natural thing. It seems pretty unusual for me that you would put something in brackets. More likely that you have something in parenthesis. So we are looking at parenthesis vs quotes in this test to see if they are treated differently.
Here you can see that we have keywords added in quotes for our 1, 2, 4, and 5, and keyword added in parenthesis in 3. In the end, when just running the test to see if they are treated differently, there was no change at all. Then running the search with parenthesis and with quotes to see if there were changes, what I found really interesting there was there wasn’t any change. Even though these pages have parenthesis on them, the bottom 4 or 5 spot did not jump to 3.
Same with the quotes, number 3 with the quotes did not jump anybody even though that search matched the keywords that were on the page. Pretty safe to say then that they were not lying on this one if that was the situation of quotes and parenthesis being treated the same.
Now, brackets can mean different things and quoting languages so we are going to run brackets as well just to see if it is possible that search modifiers, you know there are a lot of weird stuff in search modifiers, there is a possibility that there is some residue hanging back with something from back in the day when search modifiers work, so we are running this test now with the brackets. But as it stands, when you are talking about quotes vs parenthesis, they are treated exactly the same and it looks like they are completely ignored.
Does a Giant-Sized Keyword Affect Rank?
I think I have easily received 8 emails about this or people pinging me on Skype or something like that talking about, in particular with htags, do you really need to have an H1 and an H2 or is it really just the size of the text on the page because the idea here is that Google is smart and they know that people can manipulate htags so they are not really looking at the tag, they are looking at the size of the text.
In a sense, you can swap out an h1 for just text that is the same size. That is running right now, this first test though is like – oh well, we need to look at first before we go to that is if large text do anything. Is that something that is even being considered just on its own, is the idea. Kind of as we unpack these concepts, can I swap this for that or does this or that work, we often lay a foundation first and kind of see if it works on its own. That is what we are doing here.
We have a giant sized keyword, much larger than the text – significantly larger. That is not an H1, it is just simply a large word that is on the page compared to the other pages – they have normal sized text. In the end, we got no change. There was no impact to the larger text.
Obviously, we wanted to run it the other way so now this one has small and these ones had large and again, there was no change. My initial thought in the htags vs size is that no, it needs to be an htag. That is the whole idea. Google is looking at the hierarchy, the semantic order of your tags. This leads me to think that we are going to get the outcome as well. But the initial thought is no, having a giant text like your page is yelling at Google for a particular keyword is not going to bring any extra.
Can You Index a Page with HTML Classes or IDs?
Basically, these are just naming conventions. You are naming some things so that you can say that this does that basically and you name it so that you can use it again. The thing is, the name is completely arbitrary. It is completely descriptive of what the result will be but it is a completely arbitrary thing. You can name something anything that you want to name it. That means you can name your target keywords. So, what we want to see is, can we index a page by naming classes and IDs with keywords because that means Google is naming it and they are considering it and if we can index for it then we can start hiding keywords.
What this looks like in here is this is a div, id, fake keyword, class, then a fake keyword. The idea is, will these pages index for that fake keyword. I think we ran this 10 times, I think. In the end, nothing indexed. I have seen this as an SEO ninja hack where you can hide keywords, ids, and classes and it is going to help you out, at least maybe hide words that you do not want on the page. You can hide them in IDs and classes and it looks like that is not the case at all.
We are testing now if these things affect rank, if we can actually get movement with them but when it comes to indexing a page just on this, it is not going to work. I can see that it would affect rank if you actually also have them on the pages optimized for that because we see that a lot for correlational data like from Cora for example where tags seem to play a part in it. I can see this affecting rank but if you say you want to completely hide a whole subset of keywords that you want the page to rank for an you didn’t want them displayed, you are not going to get it done doing this.
Does Google Index the HTML Source Code or the Rendered Document Object Model “DOM”?
Ryan Hough had the lead on this in terms of implementing and running it. This was Ted’s concept and Ted also supplied the code for this particular test.
The concept here is if you look at a webpage and you right click and inspect the page source, does Google look at that when it is indexing pages, when it is reviewing the page, or is it actually rendering the page with the DOM, like the actual structure of the page. Is that what it is looking at, is that what it is indexing.
A couple of implications here, one is Google says they use the render dom and a lot of SEO theorists base it on that. If they don’t and if they are just looking at page source and are not nearly as advanced as how people make them out to be, they are basically Netscape or Lycos and simply are looking at the source code for everything.
The other thing that we are looking at is that this could progress our knowledge when we know that they are actually using rendered dom, we could actually play a lot of things in understanding how Google is interpreting a page. For example, we could find perhaps where a page load time ends, we could look at when they stop listening to a page, as you know this goes into Cloaking issues. So, step one is we need to figure out are they looking at the source code or are they actually rendering the page?
How this works is that we have two different keywords here. The first keyword is only in the html. The second is actually issued through JavaScript in the DOM, the idea is, the initial html has this keyword as soon as it rendered, that keyword gets swapped out for this one. If this first one gets indexed, then we know that Google is just looking at html. If the second one is indexed, then we know that it is looking at the rendered dom.
This is what they look like within fetch and render and this is the page indexed for the separate keywords and you can see that Google actually indexed them and they exist. Here we go, this is the html, not indexed. This is the rendered dom, and it is indexed. So we know conclusively that Google is using the rendered dom.
The thing is, html and rendered dom are probably like 95% the same but there is a lot that could be different within them and a lot of tools just used, the source code when they are evaluating things are not using the rendered dom. That is a way that you could actually trick tools that your competitors are using. But then also, we can look at ways to exploit now how Google is using the rendered dom. How long are they “listening” to it and how are they “listening” to it? Once we have those answers, we can really expound upon these concepts of how to manipulate pages, if manipulating pages is your thing.
Dori: Yeah, I like manipulating pages. We started talking about cloaking and now I had actually gotten into this business in 2001-2002, we started cloaking and really doing well because everything is based on onpage mechanics. Coming back to on page like we are talking a while ago is really exciting for me to perhaps get back into cloaking where onpage is all it takes. You can really do a lot of different things when you are cloaking and sending traffic to a single source. It is really exciting for me. I think about going back to cloaking anyway but if we can get this and tell when Google is coming and not coming, it gives us that more compass to go after.
It is a huge edge when it comes to how you put things together. Your timing on putting them together, when it is built, when it renders, that sort of thing, that is a lot that can be done here. We are exploring this in the next couple of months.
Dori: You said something about tools – people’s tools are probably using the source code, is that correct?
Yes, this is the next thing to look into – Screaming Frog which is a tool a lot of people use. I think they are just using the source code, I don’t know that they are actually rendering the page. They might be now. They might actually have advanced features but there is a good chance that they are just looking at the source code and if so, you can actually give your competitors that are scraping your pages or looking at how you are doing things, you might be able to give those tools different information than what you are actually giving Google.
Dori: Can you explain what rendered dom is and why it is different from the html source code?
It actually is just the page as it is. It is what you see in fetch and render, as they render the page, that is what it is. You are actually looking at the page as it is rather than when you right click and inspect the source and you see all the different tags and codes.
Dori: But it is the source code that creates that rendered page?
Some of it but then that is when you can get into things like Javascript where you can swap things out so you can have a script that can swap out pieces of code so that for example you would show one thing then swap it out for another but that will never show in the source code.