For this webinar, we talk further into detail on the tests for this month's dossier and how you can apply it to your site or your client's sites to help you move rank.
Test 1: OnPage Word Count
Test 2: Mythbusting H1
Test 3; Critical On-Page Factors
Test 4: Do Google Docs Pass Link Juice
Test 5: What are the Strongest Web 2.0's?
April 2016 Tests
Notes Below
The purpose of the monthly SIA webinar is to go through the recent tests we have done and answer any questions there are on them and talk about ways to use the information we discovered.
Webinar Agenda
*What is the SIA Testing Club?
*Why is testing important? How it helps you?
*Rules of Engagement
*Testing vs Case Studies
*How we do tests?
*This month’s test
*What we have slated for next month?
*Q&A
What is the SIA?
SIA stands for the SEO Intelligence Agency. It is a community of SEO Professionals that was initially brought on with a presentation in SEORockstar that inspired me to start testing again. We have the FB and Skype group where the community comes together. Skype is where you could find somebody if you need a question to be answered right away.
The Facebook group is also a community where you could ask SEO questions. We have the webinars where your questions will be answered live and also chats on Fridays. We will also bring in special guests for time to time.
SEO test results are delivered to your doorstep monthly. We have already started doing group tests and some individuals have also been showing their tests. We are pooling all these and we are going to be reporting on them in case you didn’t see them. We will be organizing this in a way where it would be really easy for you to pick up your handbook and see the answers to your questions.
The big picture is:
- All of the testing will reverse-engineer Google. That is the primary reason of the group: to reverse-engineer what is working on Google.
- Create the most exclusive “Google Ranking Guide” based on scientific findings of singular algorithmic variables (no theory). We are breaking it down into a single thing and finding out if this is a yes or a no, if it is working or not, what is better? We are going to start testing each one of those variables and create this massive Google ranking guide which is going to be based on scientific findings and not on theory.
- Build a platform to keep it all organized so the test results will be easily accessible to you. That is the big picture – offline with the binder and then online.
How does this help you?
- Time is big money. Our test results will save you the time and money from doing them yourselves.
- It also saves you time with your processes. When you do tests and get the results, you get to do what is working. You can spend more time on what works instead of those which do not.
- It gives you a safe space to ask advanced level SEO questions. Even if you are a beginner, feel free to ask anything. That is what we are here for. That is how the advanced level got to where they are. Sharing the material allows us all to gain more knowledge. That is what is really important in an SEO community like this.
- When you know the results and you are not guessing, it gives you the confidence to know what is working.
Housekeeping rules
We will post the rules for the FB group and they will apply to Skype as well.
- Simple Rules
1. No Bashing of any kind - Don’t put anybody down. Be nice.
2. No Pea-cocking – Keep your ego in check. Keep it on the down low.
3. No spamming other members - No hitting up for JVS.
4. No antagonists or haters allowed
5. No stealing of content – This content is for YOU and YOU only!
- Please remember when posting:
1. This is a place for open, positive scientific SEO conversations.
2. This is not a place to promote yourself, trash talk of any kind or any programs (including this one), or its members. We want to keep this a great, positive place.
3. This is not the place to post questions that can and should be answered by our support staff. For feedback that is directed to the SIA staff – which we love, please keep to good form and present to us privately, thus not show-boating for agreement.
· Posts that do not adhere to the above quality will be deleted.
Suggest a Test
Anytime that you would like to suggest a test, go to http://suggestions.thetestingclub.com
Testing vs Case Studies
Testing - Breaks it down to a single variable so the answer is a definite yes or no.
Case Studies – Are mostly done on live sites and because of multiple variables, it can only deliver a subjective answer.
Are great and valuable. That is what we do and have done in SEO. Now, we’re just breaking it down to smaller pieces so that we could get a more definite answer on each one of the aspects to the algorithm and how they work together.
How to set up test
1. Choose a keyword which doesn’t exist in Google’s database – http://soybomb.com/tricks/word
2. Come up with a good lorem-ipsum text – http://generator.lorem-ipsum.info/
Reasons why I like the generator:
- You can start with not using lorem ipsum.
- It can generate random stuff for you.
- It gives you the plain text option, so you aren’t adding extra code when you are copying and pasting on web pages.
- It also lets you choose word count which is very important when running test especially if you are testing keyword density.
- It has the ability to edit.
3. Set up pages to eliminate variables
I often use WordPress and I make it look like an actual page – add in an H1, more texts, h2, more text, etc. Then, I will replicate the content and put it in another page. The only difference in the page would be that one thing that we are testing.
One other small point that you would probably run into is in creating a unique URL. I have gotten past that with the use of the keyword in the URL and with putting two fake words behind it with matching amount of letters.
The idea is to have the URLs with the fake keyword closest to the .com and two words that have equal number of letters after the keyword. That is the only thing that would be different between the two pages. Everything else would be exactly the same. That is how close you can get to eliminating variables down to as little as possible.
Types of Tests
a) 2 page test – Simply two pages where we are looking at 1 particular variable.
b) Larger field (5 pages) – We launch 5 identical pages, have them indexed, and see what order they are indexing in. Then, we will fire something out, or change one to attempt to move one in the larger field. This is often done with links test. The idea is to see which one is #2 and then fire links at it and see if we can move that to #1 or if it drops down. We can then manipulate things from there.
c) Full SERPs (10 pages) - We take up the whole page in search results. An example would be the on page test where we are going to run ten pages so we can fill up the entire SERPs.
d) Case Study – Case studies are done once you feel good with the results you are getting in the test environment. You can take it up a live page before you take it up in a client’s site.
4. Observe and record – I use Google sheets and visualping.io
This month’s test
1. On-Page Word Count
We lay down the foundation for our word count test. In this test, we set up 10 pages with each page having a specific amount of words on the page. In this case, we started it up to 100 and took it up to 1000. So, page 1 has 100 words, page 2 has 200 words, page 3 has 300 words, etc. They all maintained the same keyword density and the target keyword was all set to 2x per 100. We had a 2 percent density.
The concept here is all things being equal, the keyword density being equal – is the number of words in a page a ranking factor? The results was yes, it is a factor and what is really amazing is, they ranked in order from 1000 going down to a hundred. We will be going into a lot of detail into word count in the next set of tests.
Keyword density vs optimization rate
Keyword density is how many times the keyword is placed in relation to the content. It is kind of an antiquated idea because it is too simplistic. When looking at the optimization rate, what you look for is the keyword density in the different places where you could get a keyword. You can put a keyword in the URL, body, metatags, metatitle, metakeywords, metadescription, etc.
A more advanced way of looking at it is by looking at a page holistically. It is not necessarily on the density of one thing but it is the density of all those different places and how they interact with each other. When you are looking at your own optimization, it is an important to keep in mind where the keyword is going.
All the places where it can go and how you have it, in relation to the top 10 pages in the SERP results or the top 30 pages in the SERP results because that will give you an idea on what optimization rate you need to have on your page. Look at the pages that Google is clearly rewarding and you could target an optimization rate based on the target term weighted between those different sections.
2. Myth busting
In this particular test, an SEO professional friend of mine called me up to say he has been called in by a company to consult. The company has an in-house SEO guy that mentioned a ninja hack and was surprised that he didn’t know about it. The hack is that you can wrap an image in an h1 and it becomes a super h1 or super image. My friend asked if I heard anything about that and I didn’t but we could test it. We created the test page, grabbed an image, wrapped it in h1, and tested it.
The result was no, it isn’t any kind of super special hack. We run it a couple of different times just to be sure. For round 1, we had two pages where we pitted them together. For round 2, we did a multi-field test where we got one super h1 against a whole bunch of other ones and the idea was to see how it fares. It lost again. It is simply not a factor.
What is great about testing is that, you can come back and say, “No, this doesn’t work. I run a test on it and I can show it.”
On the last page of the h1 myth, you can see a spreadsheet where I recorded the data. Recording data – dates, times, notes – is important when testing.
3. On-Page Factors
I have the critical on-page factors test running all the time. I also recommend this as a great kind of test to start with if you haven’t started any testing yet or are looking into starting a new test. You can create the pages and get a feel of how it all works. If you have an on-page factor that you use or heard of a factor, throw it in the mix.
There were two big takers in the test:
a) Get the keyword in the URL, in the body, in the metatitle, and the h1. Intuitively, as an SEO, you probably know that these are the best places where you could put in a keyword. The test result backs this up.
b) The metatitle that I used for this particular test was the location so that it would be easy to identify when we print out the results, and what was interesting was that in those two results, Google completely ignored the metatitle that I gave the page. That was for the h1 and the h2. I am sure a frustration that you have had before is Google uses a different metatitle from what you wanted.
The test gave some insights as to what Google might be using and clearly it’s looking for indicators for its title probably in the metatitle that you put there but also if your alt also matches your h1 and h2 - are they close, and does it make sense?
I’ve run this test many times but in running it again, I saw this for the first time. Now, you can see the places where Google is getting the information from to throw in the title. I imagine there are probably different factors there but those are the two big ones.
If you are in a situation where you really want a particular title and you really want Google to use it, throw it in your h1 and something similar in your h2. At least, if you can’t get it in your h1, then get it in your h2. Coincide that with your metatitle and you’ve probably got a big chance of getting that display.
4. Do Google docs pass link juice?
The short answer is yes, they do. This test came from an observation that I made when I saw all these Google docs and sheets that were ranking when people published them. When they rank, that means Google is reading and indexing them.
If Google are reading those, how about the links that sit in there? They pass juice. This opens up a ton of possibilities. Apparently using Google docs for links is a part of a program that someone is selling for almost $1200 – you’re getting access to this for a lot less.
This technique costs nothing, other than the time to put it up. You have complete control. You can change your content and links anytime, even remove them. You also have the opportunity to have all your links on top of the page (links on the top page are more powerful than those at the bottom).
Plus, you can also create tiers with it. You can have Google docs that support other Google docs. You can point these to your money sites or elsewhere. The possibility of building entire link empires off of Google’s own property is fantastic. Don’t limit yourself to just one mode of thinking.
Example of how to use it is for local SEO. If you are based in Austin and you want to make sure that Google understands that but you don’t want to put unnecessary information on your sites such as things to do in Austin because that doesn’t make sense for a plumbing site, you can start building out local SEO mini sites within these Google docs and point them in.
The idea is, not only are you getting a link coming in, you are also getting the association of the content around that link. If you have extra content, this is another place where you could publish those content then, link to your site. You could also use it to build links using LSI (latent semantic indexing) terms.
5. Strongest Web 2.0s
In this first round of web 2.0s, the real strong ones are Blogger.com, Tumblr.com especially on mobile, Google sites, WordPress, and Joomla. Google sites is sites.google.com/yourpage. Blogger is blogspot.com, which is also owned by Google.
This is in conjunction with the Google docs ideas where you can build lots of amazing links that are inexpensive and in properties that you can control. These are strong sites to use for web 2.0 building and you can integrate them with your Google doc links. It allows you to create tons and tons of tiers of link building that you have easy access to.
One thing you have to be careful about is to not fire links you can’t control to your money sites. By setting up this kind of link building, you have the ability to control the links -- where they go, what the content is, etc.
One of our members, Brian, has had a lot of good luck with about.me. He builds out very nice looking sites on them. If you’ve got the resources to build out really nice web 2.0s, you could also do well. Even with the ones that didn’t do well – like about.me for example. All these properties would work pretty well if you are going to build out really high quality stuff.
If you have less time and resources and need a place to start and focus on, the top ones are the ones to start with. If you are looking at the strength, which ones are really going to pop out and help me in providing the best quality link, go with the ones that are listed.
QUESTIONS?
I still see Google changing my metatitle without it being in my h1 or h2. Does anchor text have an impact on what Google will make your title tag as well?
Google is trying to put out something that is definitely relevant to the actual search result. The idea would be, if the search was about red bicycles and you are putting red bicycles in your meta title and in the h1, h2, that is when you have the best opportunity to have the title that you want to be displayed. Is it going to happen every time? No, because I have seen a few occasions where Google flat out refuses. Even on what looks to me to match up exactly.
Could you talk about the anchor text for the Google docs?
I create them like what I would do to a regular webpage. When you create the docs, you can tag it as something like an h1, with htags. I write them like an article and I almost always use the anchor text because it is a valuable link.
You tested for 2% keyword density, can a test be done for 1% keyword density?
Yes, that is not a problem at all and if you are interested in running the test, I will be happy to run the test with you to see how that works.
Would be nice to see the impact of having a virtual office with address and phone to setup a corresponding Google+ page.
On the idea of a virtual office, something that is interesting is (something that I had not tested), I was on MozCon Local a couple of months ago and Darren of Whitespark showed test results on fake businesses which he run through citation programs to see results and he found that if you are using a big virtual office like Regus, Google shut that down because it is easy for them to get Regus on the list. But if you used a local company, those got through. Having a unique phone is also important.
Can you discuss how you use: http://visualping.io/ ?
The program checks if there is a change to the webpage you have asked it to monitor. I use it in tests because it is nice not to have to check your test every second of everyday. Visualping will do a check once a day for you and it keeps a record of it.
I am very interested in the impact of local schema. Will you be doing tests on that? Are there any best practices you can share now?
There would be a lot of tests in schema. I have found that in November or December, Google released a new manual action in webmaster tools or the search console that there is spammy schema markup. I thought that was really significant because schema has always been the next big thing for quite some time. I see that as a giant indicator that it is coming or it is here because Google would not take the time and would not have an actual person to manually look at your schema if it isn’t a factor. I would definitely put schema in your radar.
When I spoke in SEORockstar, I have mentioned I have done some schema test and a lot of them are inconclusive. Some things seem to work then they just seem to go away. With the introduction of this new manual action, that tells me this is where it is going to be.
There is the schema tool that you can use and that everybody here has access to. I recommend using it. What I do with it is, I mark the organization and the local business. Many of the large services are covered within schema. I use that to markup the homepage. The next thing I’ll do is to mark up products and services. What is nice about that is, if you use product in conjunction with offer and rating, that is how you could get stuff to pop out in SERPs. If they are doing a service, I mark up their service pages. Schema is an additional package I offer when I do optimization.
Schema or JSON Difference, which one is better than the other?
JSON is a form of schema. The two big forms of schema are RDF, which is rich data format, and JSON. There are a couple of others, but those are the big ones that most people are using. RDF is an html tag and if you ever look at a page source, it is the one that says itemprop. It is tagging something in the html. JSON is a script and it tags everything together in one script and you put that entire script up at the top of the page that you are marking. RDF goes inside individual things. JSON goes inside but in one big script. I like JSON because it is a lot easier to do.
On Schema – would you recommend site wide or different for each page?
I do different on each page. I don’t do anything site wide. I put the organization on the homepage, marking what kind of business it is, then I mark up the products, services, etc.
Schema displays, JSON-LD does not.
Schema is the whole system. If you go to schema.org, you can see different types of schema that exists. Often, if you are using a WordPress plugin, because WordPress has issues with tagging, if you use that plug in and put it in, it will only display it because it has to because of that plugin limitation. It is usually not desirable because of how they are displaying it for you is duplicating what you have on the page.
So schema is the over arching thing. JSON-LD is one form of it. RDF is another form. Microdata is another form.
Thoughts on Project Supremacy plugin?
I heard a little about it. Just that, I think, it has schema as one of its things, but I don’t have an opinion because I haven’t used it.
What about backlinks that are hidden on site by visitors, but still picked up by crawlers. Just as effective, less effective?
Dori: I remember somebody doing a huge presentation on this back in Traffic Conversion and showing people that you could hide links this way especially with CSS because Google crawls it just like anything but the way it is set up, users can’t see it. It would, more than often, pass a manual review.
My hypothesis is, it would just be as effective.
I think it is as well. If the crawlers is seeing it just as the same as any other link and you think it is going to be credited the same way. I get nervous though about doing cloaking. If they have the ability to see something you’ve cloaked, that seems likely.
What is interesting is I’ve done some testing if CSS is crawled and what you can and can’t put in there and what I have found is, if you can highlight it on the page, it can be read. Even if it is just in the CSS. If you can highlight it, it can be crawled even if it is something that does not technically exist on the page and is dynamically inserted in the CSS, you can get it to show.
Is duplicate content a penalty?
The idea there is if there are any penalties being applied to both. Duplicate content, in this kind of situation, isn’t going to be the penalty that will get it deindexed because you won’t get there until you have a manual spam action. What you will get will be a lowering or negative impact. So, if there is a penalty put in on the pages, it is going to be put on both of them. So, it is not an issue in the test because essentially, even with the variable, they are both being treated the same way.
The idea is when we launch this pages, they are both launched at the same time and they have both been submitted to Google at approximately the same time. I don’t think there is any way for Google to know which came first.
If you have two identical pages, how much do you have to change them to make them “different”?
In the tests that I have run, I could literally tell you it is one thing. Just changing an htag makes it different, changing an image makes it different.