For the September webinar, the tests in the 6th dossier were discussed further. Questions that came up regarding the tests were also answered.
Test 51: Which Link Shortener Is “Better” To Send To Ranking A Page From A Google Doc
Test 52: Which Link Shortener Is “Better” To Send To Ranking A Page From YouTube
Test 53: Do PDF’s Pass Link Juice?
*Determine Whether A Link In A PDF Impacts Rank
*Does a URL in a PDF Need To Be Hyperlinked to Pass Juice?
*Determine Whether A PDF Ranks Better Or Worse Than A Web Page
*Determine Whether Adding Metadata To A PDF Improves Rank
Test 54: Search Volume Test The Keyword Planner vs Reality
Test 55: Blackhat Special: Do .gov Redirect Links Affect Rank?
Test 56: Determine The Relative Strength Of An Internal Link
Test 57: Is Latent Semantic Indexing (LSI) A Ranking Factor?
Test 58: User Metrics: Is Time On Site A Ranking Factor?
Test 59: User Metrics: Is a Google Analytics Conversion A Ranking Factor?
September 2016 Dossier
Transcripts Below:
Which Link Shortener is “Better” to Send to Ranking a Page from a Google Doc?
This is Carolyn’s brainchild. She took different shorteners – Google, Bitly, Capsulink, etc. – and threw them into Majestic and realized that there are different metrics with these shorteners.That got her to thinking if using different shorteners may impact things differently. That is where these two tests came from. In this version, we are looking into putting it on a Google doc.
The reason why we are doing that is it’s something that we have shown a few times and the things we have tested with are passing juice through a Google doc. So we thought, what if we use a shortener? There are a lot of advantages there - maybe disguising what you are doing better? Some shorteners, if you pay for the premium version, allows you to change the destination of the URL so it is possible that maybe you could get extra value from it.
In these tests results, we made sure that we used different Google drives. The reason is that in previous testing, we realized that if you use the same drive over and over again with the same Gmail address that it is attached to, in certain things you get diminishing returns or you may not get any value at all.
For each tests, to keep things completely clean, we used different Gmails and Google accounts. Then in all these, we threw things into the fourth spot. The reason is that we know that we are going to get positive movement because the Google docs do pass juice.We already know that it is a positive ranking factor. The idea is that it might give us some range as to how much juice we are passing.
In the end, the bit.ly link went from the fourth spot to the third spot – one move up. Capsulink went from 4 to 3 and the Google shortener went from 4 to 2. It was pretty exciting that the doc continued to pass juice as it normally would and in the Google shortener scenario, it moved two spots. Now when we use a regular link, we usually get it to the two or the one spot with Google doc so the bit.ly and the capsulink might have diminished it slightly. The Google was probably right where things are normally are.
Which Link Shortener is “Better” to Send to Ranking a Page from YouTube
The next test looked at the exact same thing – shorteners but this time for YouTube. Again, we talked about YouTube stacks and the things that you can do to build links for your clients and for your own websites. Then, looking at combining those things that we already explored and discussed, I thought, what if we use the shorteners that are showing different values?
What was great there and what was really exciting was that with the naked URL went from 5 to 3, which gave us a baseline. Bitly also went from 5 to 3 but Google and Capsulink both went to number 1, outperforming the naked link. We usually see the naked link in the 3 or 2 spot which is not a super strong thing but the Google shortener and the Capsulink shortener seem to jazz it up a bit which is pretty exciting. Google shorteners are free so you can use them as much as you like. For Capsulink, you have to pay but what is nice about it is that even at their base level,you can change their destination URL.
Based on these two things, if you are interested in using a shortener in these situations, I would definitely go for the Capsulink or Google shortener. I personally have started using Capsulink in my own when I am pointing different stacks to different things. Now, we are testing to see if we can prove that you can get link juice passing to more than one destination when you change the destination URL in Capsulink.
We are working on seeing if there is value there, in the sense that you can reuse the same link that you already have. That has huge implications when you think about putting a link in a press release or on a real site. Why is this important? Because we have seen in a previous test that you are going to get the echo and that echo will lasts for months and months. The goal here is to see that we can use the same link multiple times. This is how these kinds of tests are going.
From the test results, it looks like using a shortener actually adds juice, right? Not only do you get the benefit of being able to change the destination URL when you have to but it is adding juice?
Yes, we are seeing a little extra bump which is pretty exciting.
Dori: That is really exciting because my software that I have a patent on doesn’t have shorteners. But to be able to change the anchor text and the URL, and here, using a shortener is able to do that automatically. You have the control here plus it adds juice. This is super exciting, and I just got some ideas.
It is amazing and easy to use. It is not anything tricky or technical and it is simply just shortening the URL.
Do PDFs Pass Link Juice?
I really like this test because back in my early days of SEO, I remember a client asking me specifically if we should put links on PDF. I did what everybody that has no idea does in that situation. I gave my opinion and I said yes, but really I had no idea. I assumed it would work but I had no concept of it. That is why I enjoyed this test because I get to answer that.
This was Ryan Hough’s test, who is one of our regular contributors to our testing and he did a great job at this test. The basic concept here is that, does it pass juice? Yes, it does. A link in a PDF will pass juice. What was really fun was when we went to the next steps on this.
We sought to answer: Does a URL need to be hyperlinked to pass juice or will a normal URL in the PDF that is not hyperlinked pass juice? The answer is no, those will not pass juice. In order to get value, you would need to take that extra step and hyperlink the URL.
The next thing that we looked at is can a PDF beat a webpage? All things considered, if you have an equally optimized webpage vs a PDF, is it possible that the PDF is going to outrank the page? The answer there is no. The PDF dropped and stayed at the bottom on that test.
Another thing we looked at is what about if we added a metadata, because within the PDF you can add specific metadata – you can add a page title and other things similar to exif data and you can put on an image. Maybe that would give a boost or juice. In this test, no, it didn’t. It stayed at the bottom and locked in at the bottom.
What we have learned through these series of great tests is that:
- a PDF will pass juice but the link has to be hypertext. It can’t just be a naked or an unclickable link.
- the PDF on its own will not beat an equally optimized webpage.
- adding in metadata to the PDF does not give it an extra boost against any webpage.
Those are all great things because if you are planning anything, it would save a lot of time in what you are doing and it will also give you some direction on some best practices involved.
Dori: I have always considered a PDF to be a channel like how I view Google ranking channels - products, videos, press releases, a Soundcloud, or something like that. So, I believe and do experience that. I see a PDF come up into a play when there are a lot of other webpages, when they are doing channels. I want to throw that out that sometimes, in my experience, when Google is trying to grab from every different kind of media that the PDF will show up.
I think when you have the website that you are working on for yourself or for a client and you are looking for opportunities to build tier 1 links, good links that are coming in from decent or DAed properties that are going to create a nice circle around that site and will pass good safe juice, PDF is one of them.
Dori: I think you can also rank that PDF as part of those channels. I am always about taking over the page one with media that Google is looking at. I should probably do a test on that.
I totally agree on that because when you look at how Google really loves YouTube videos, they rank easily. I think a PDF with a little love and attention can probably be the same thing.
Search Volume Test: The Keyword Planner vs Reality
What we were looking at here was what the keyword planner in AdWords give you and what actually exists. The keyword planner is pretty much what every tool uses to extrapolate search volume data from. So, when you put in a particular keyword into any tool, I would guess that every tool is using the AdWords planner API in order to predict its estimates as to search volume.
I noticed a couple of years ago that in AdWords data, what they give you as estimates are really not close to reality. They just completely go off the rails from time to time. In my own practice, I had some client accounts that were very aged and we could go back and get a year’s worth of data on some specific keywords and that allowed us on this test to really take a nice look at real data for a year, what is AdWords telling us in regards to estimates?
I could see AdWords cop out when you do a month’s worth of AdWords campaign and you get a couple of different data and they say that it is such a small sample set, that is only one month and you need to extrapolate for a year. What is nice about this is being able to extrapolate for years andbe able to see some differences.
In the end, what we are looking at was that there was a dividing line and it seems when you are above 200, the planner is not bad now, at this very moment, and what I am estimating right now with my own clients is a 20% below or a 20% above. The planner tells me a thousand, I am telling client’s right now that the search volume is somewhere between 800 and 1200.
The real important point is that when it is 200 and below, it is a complete lie. I do not think Google has any concept or real way to know what those numbers are when they get below 200. If they tell you a hundred, it could be a hundred and it could also be zero. It could be even more that 200.There is a lot of variance there.
If you want to test this on your own, one thing I would encourage you to do is to check out if you put in a particular keyword in search and then scroll to the bottom and see suggested searches at the bottom, things related to your search. Grab one or a bunch of those and throw them in the keyword planner.
What I found with those is that they often come out with 0 results. That is actually impossible because in order for those things to show up as related to your search, somebody should have searched for them before. Google isn’t creating brand new keywords. What they are doing is they are aggregating things that have already been done so when you look at suggested reports in the SERPs and you put them into the planner and you see zero, you know that there is a disconnect.
The data that we can get from AdWords on the lower end or longer tail keywords are going to be very off. In your own practice when you are looking at long tail keywords for example, it might not be a bad investment to do a smaller AdWords campaign to see if you could get better results as to what you are actually playing with.
For example, you’ve got 15 long tail terms and the planner is telling you they are all 200, it might be worth your time to do a months worth of research on them within an AdWords campaign to get better search volume so that you could get better informed decisions on whichpages to build out for your organic search.
Dori: I like this one because it gave me more inspiration to go after longer tail keywords that are 200 and less knowing that there are maybe more or maybe less too but it is true. It can’t be 0 if they are related search terms so there might be something goofy going on in Google there.
I think they use different methodologies and as a result, the AdWords planner and assessment,the sampling is too small that it couldn’t give you a nice accurate prediction until it gets to that 200 mark and then the subset is so much larger it can give you a better idea of what is going on.
Blackhat Special: Do .gov Redirect Links Affect Rank?
This test popped up as a result of a conversation in the SIA Skype group. If you are not part of the Skype group and you do not have a Skype account, I recommend getting yourself one because there are a lot of fun things that is discussed there from time to time and this is one of them. I remember this technique from a while ago where you can go a .gov link by hacking the redirect page on a very old .gov site.
This is a security flaw that when you click on old websites to leave and you go to redirect page for a minute and it would give you a countdown page from 5 to 0 and it says “Hey, you are leaving our site and going here, 5, 4, 3, 2, 1” and then you are redirected. You can create that redirect page simply by knowing the parameters in the URL string.
Essentially, you would create any link to any page you would like from this redirect page.This technique from a while ago would give you bunches and bunches of .gov links going to your sites and the presumption being that .gov links are as good as you can get. The .gov links do not arbitrarily link out, there isn’t any spam .gov sites, if you get any .gov link, presumably that is the best. Maybe .edu after that and the rest from there. We know this technique but does it actually do anything?
Part of my concern is that these pages are temporary. They do not actually show up until someone goes to them. My thought was that they would never pass link juice because they do not necessarily exist but I ran them through a link indexer so that Google would actually go through them. I did the same thing with those links in Majestic. Out of those 5 links, 4 actually showed up. However, they did not move any kind of rankings. It didn’t move anything at all.
In a follow-up, we actually blasted links through those pages with the hope that maybe they would pass juice and affect rank. That follow up will be on the next dossier. But on their own,they probably won’t move rank. It would look pretty cool and you’d feel like a ninja. Unfortunately, it would not do that much for you.
Determine the Relative Strength of an Internal Link
Silo-ing is a real pet project of mine. I am a silo advocate. If you are building a site architecture, I recommend silo-ing. The thing to keep in mind in silo-ing is say you have yoursite.com/targetpage and by target page is what you are trying to rank for a specific keyword and you want that page to show in organic results. How you silo that is most often through a blog but it could also be done through pages but in this scenario it would be through a blog.
You would have mysite.com/targetpage/nameofblogpost so that URL structure for that blogpost contains the target page and then it contains a link on that blog post going to that target page. That is called a virtual silo and it does not need a child-parent relationship to do that and you could handle it through your blog. There is a little bit of tagging expertise to it but it is not too terribly tricky. That is what we are talking about here when it comes to silo-ing.
The yoursite.com/targetpage should contain your keyword. So, if you are selling red bicycles it would be yoursite.com/redbicycles and then you’ve got your content that supports that page and you might have yoursite.com/redbicycles/5tipstofindthebestredbicycles and that is the post that you have written about red bicycles and in the URL structure, you could see that the red bicycles page is in that URL and there is also a link going from that page to the top level page.
What is really awesome in this test is we were testing just a Google doc link. We were pitting internal links against a Google doc because we know that passes juice and it is probably in the lower end if we have a baseline of which passes juice. It would probably be the Google doc just on its own.
You haven’t blasted any links to it and you haven’t done anything else to it. It is just something that we know passes juice. Google doc is something at the bottom of the scale in terms of power. We also know that the DA (pick your metric) is very high but we know it is a baseline link.
We learned that two silo’d links beat the baseline link. So, that is two posts properly silo’d and linking back to a target page will beat one baseline link. This is exciting because when you think about building even a Google stack, there is more effort that goes into that than building a couple of blog posts. Your own site or your client’s site probably have a lot of unutilized blog posts that aren’t properly silo’d and they aren’t properly linking somewhere. You can immediately get an equivalent link juice pass just by reorganizing your site.
Last year, I talked about this in SEORockstar where I silo’d a site with three years worth of blog posts that were really well-written posts. I silo’d them all and pointed them to the proper pages and did some on page optimization and I got first page rankings out of it. That is when I realized that silo-ing is a very powerful thing.
How many Google Docs do you build from one Gmail account?
If I am pointing them to my money site, I only do one. I use one Google doc from one Gmail account and I do not use it again.
Can you set up a silo on an old site without changing the URLs?
From a WordPress perspective, you have to change the slug – that little piece past the .com.You could still point from a random page in but that random page, if it doesn’t have the target URL page in it, it is unsilo’d at that point. In this test, Ryan did that and it is just not effective.
What if someone wants to look at your posts and your categories?
What I do when I set this up is my next and previous stay within one category. Say, when they click next, usually in that scenario, you would go to the next most recent post based on time. Inthis, I would send them to the next most recent post in this category.
Dori: You do this manually right? It automatically goes by date in WordPress
Yes, that is correct and WordPress does not automatically categorize by pages. What I did was build a plugin for this. You can then categorize the pages and match the slug so that when you click in the post, this category goes to this category and it matches up and points correctly to your target page.
The only URLs that I would change would be supporting content pages. I wouldn’t go in and change a top level page that is currently ranking for something. Where you would look at changing the URL structure is in your blog section or your supporting content pages.
The reason is when most people write a blog, they are writing great content about the things that they do or the things that they sell and you have to kind of cross your fingers that Google makes the connection between this content and that page. What you would do in a silo is remove the mystery for Google and let it know that this page is supported by this piece of content.
That is what the silo structure does so that when Google crawls through those folders, it would crawl through your target page and then get to the supporting content and then there is a link from that content going to your target page. It creates a silo which is a contained environment.
The idea is that Google would just crawl around that particular category seeing all those content.For Google then to go to the next target page in that site, it has to crawl up and out of the silo.
On an e-com site, is it best to leave the category URLs on the product pages?
I actually discussed this a little bit on e-com structure. If the category is part of the search, I leave it. If it is something generic like “products” I leave it and change it to something that is a target term.
Most out-of-the-box things use mysite.com/product/product-1 and the idea is that all your products are under the product category. Nobody searches for “product red bicycle” so that is a disadvantage but if this category is something that is actually part of the search, then it would be advantageous to leave it in. A direct example would be “services” – sometimes the word service is something that somebody would search for. “plumbing services”, for example, wouldbe a good one to leave.
Is Latent Semantic Indexing (LSI) a Ranking Factor?
LSI, the way I define it, are terms that naturally come up in a conversation about a particular thing. So, if you are having a conversation about kitchens, you might say things like sink,cabinets, pantry, etc. – all of those terms are LSI. Most people define it as synonyms or variations of a term but I think that is incorrect. In the kitchen example, it would not be variations or synonyms for the term kitchen. What we are looking at is if LSI terms are a ranking factor. If you do a search on that, you get a wide array of opinions. We want to narrow downand see if it is actually a ranking factor.
We used the tool ntopic.org which is a tool that spits out LSI possibilities and it gives a ranking of how LSI-d your content is. Whether or not the tool is accurate is not the point but we want to be able to show that we have two pieces of content equally optimized but one piece has LSI terms on it and the other doesn’t. On the first one, the tool was showing us that it had 80.93% LSI.
Page 1 was at 80.9 while page 2 was at 36%. What we did in the situation is that we had the exact same number of words in the lorem ipsum. The optimization was exactly the same. We changed out some words in the lorem ipsum for both. Page 1 had LSI terms while page 2 just had random words that aren’t LSI terms and that has nothing to do with the target page.
I ran this multiple times and in the end, the page with the LSI terms wins every single time against a page which was equally optimized for a random keyword. The LSI terms were just in the body content. The number of words between the two pages are exactly the same and even the spots where the LSI terms were and the random words were added, were exactly the same.We made sure that the random words that were chosen were not part of the recommended LSI terms in Ntopic.
User Metrics: Is time on site a ranking factor?
This is a follow up on the bounce rate test that we did in the previous month. The idea here is that we are looking at user metrics. The assumption is that these things are part of the algorithm. It is a broad assumption across the board, that there are things that Google clearly see in Google Analytics and the idea is that if Google can see that in GA, they can then use it and extrapolate that data to qualify how good your content is.
If people are engaging with it, if they are spending time on your site, if they are clicking through and seeing things, those would indicate that you might have better content than somebody else and that would be a ranking factor. The purpose of these user metrics are very obvious. Are these things actually a ranking factor?
We found that bounce rate clearly is. So in this next two, we were looking at time on site – the dwell time or the time that somebody sites in page. Then the other one is conversions and how that is done is you can set conversions in Google Analytics and it is something that we have recorded. This is a great theory and I liked it a lot because the concept is that you told Google what is most important to your site and this is what I am trying to track.
The idea then would be if you succeed in that, you have proven that your data is appropriate for that search –would Google reward that and can you move up in rank?
With the time on site, we can’t say that it is a ranking factor at this time. There were a lot of problems controlling people that were doing the actual tests. We did get positive movement with the time on site which we were happy about but we cannot conclusively say that time on site is a ranking factor in the same way that we can with bounce rate. What I am telling clients right now is that it is probably part of the algorithm and we work on ways to make people stay longer. In terms of testing we could not say a 100% that it is a ranking factor.
The obvious thing is that different sites have different content so that they are going to have different bounce rates and time on sites. I would imagine there would be a kind of interplay and we see in the result that came up here. The page that moved up to number one and stayed there has a much better bounce rate than the experiment page and has the second most visits.The experiment page had the most visits and the best time on site. There are some interplay with all those concepts – an equation where in bounce rate may be more important and maybe x your time on site x your visit.
I would encourage you to do something on your site to make people stay longer. Bounce rate may be more important.
User Metrics: Is a Google Analytics Conversion a Ranking Factor?
When it comes to conversions, unfortunately, we can’t say that it is a ranking factor at all. In the end, our test page had 60 conversions compared to no conversions on all the other pages and inexplicably, the test page dropped to the bottom and I have no idea why that happened but that salts in that Google is not taking an extra step to see how conversions are. They are not individually adding that metric into the algorithm at all. They are not doing a per site evaluation, e.g on this site has these important thing and they are killing it so their content is better – they are not doing that and it makes sense.
Dori: I think it is kind of stupid on Google’s part not to do that. If somebody puts Analytics on their site and they know they are converting, it seems to me that it is a good site and people like it for that keyword.
number 1, outperforming the naked link. We usually see the naked link in the 3 or 2 spot which is not a super strong thing but the Google shortener and the Capsulink shortener seem to jazz it up a bit which is pretty exciting. Google shorteners are free so you can use them as much as you like. For Capsulink, you have to pay but what is nice about it is that even at their base level, you can change their destination URL.
Based on these two things, if you are interested in using a shortener in these situations, I would definitely go for the Capsulink or Google shortener. I personally have started using Capsulink in my own when I am pointing different stacks to different things. Now, we are testing to see if we can prove that you can get link juice passing to more than one destination when you change the destination URL in Capsulink.
We are working on seeing if there is value there, in the sense that you can reuse the same link that you already have. That has huge implications when you think about putting a link in a press release or on a real site. Why is this important? Because we have seen in a previous test that you are going to get the echo and that echo will lasts for months and months. The goal here is to see that we can use the same link multiple times. This is how these kinds of tests are going.
From the test results, it looks like using a shortener actually adds juice, right? Not only do you get the benefit of being able to change the destination URL when you have to but it is adding juice?
Yes, we are seeing a little extra bump which is pretty exciting.
Dori: That is really exciting because my software that I have a patent on doesn’t have shorteners. But to be able to change the anchor text and the URL, and here, using a shortener is able to do that automatically. You have the control here plus it adds juice. This is super exciting, and I just got some ideas.
It is amazing and easy to use. It is not anything tricky or technical and it is simply just shortening the URL.
Do PDFs Pass Link Juice?
I really like this test because back in my early days of SEO, I remember a client asking me specifically if we should put links on PDF. I did what everybody that has no idea does in that situation. I gave my opinion and I said yes, but really I had no idea. I assumed it would work but I had no concept of it. That is why I enjoyed this test because I get to answer that.
This was Ryan Hough’s test, who is one of our regular contributors to our testing and he did a great job at this test. The basic concept here is that, does it pass juice? Yes, it does. A link in a PDF will pass juice. What was really fun was when we went to the next steps on this.
We sought to answer: Does a URL need to be hyperlinked to pass juice or will a normal URL in the PDF that is not hyperlinked pass juice? The answer is no, those will not pass juice. In order to get value, you would need to take that extra step and hyperlink the URL.
The next thing that we looked at is can a PDF beat a webpage? All things considered, if you have an equally optimized webpage vs a PDF, is it possible that the PDF is going to outrank the page? The answer there is no. The PDF dropped and stayed at the bottom on that test.
Another thing we looked at is what about if we added a metadata, because within the PDF you can add specific metadata – you can add a page title and other things similar to exif data and you can put on an image. Maybe that would give a boost or juice. In this test, no, it didn’t. It stayed at the bottom and locked in at the bottom.
What we have learned through these series of great tests is that:
- a PDF will pass juice but the link has to be hypertext. It can’t just be a naked or an unclickable link.
- the PDF on its own will not beat an equally optimized webpage.
- adding in metadata to the PDF does not give it an extra boost against any webpage.
Those are all great things because if you are planning anything, it would save a lot of time in what you are doing and it will also give you some direction on some best practices involved.
Dori: I have always considered a PDF to be a channel like how I view Google ranking channels - products, videos, press releases, a Soundcloud, or something like that. So, I believe and do experience that. I see a PDF come up into a play when there are a lot of other webpages, when they are doing channels. I want to throw that out that sometimes, in my experience, when Google is trying to grab from every different kind of media that the PDF will show up.
I think when you have the website that you are working on for yourself or for a client and you are looking for opportunities to build tier 1 links, good links that are coming in from decent or DAed properties that are going to create a nice circle around that site and will pass good safe juice, PDF is one of them.
Dori: I think you can also rank that PDF as part of those channels. I am always about taking over the page one with media that Google is looking at. I should probably do a test on that. I totally agree on that because when you look at how Google really loves YouTube videos, they rank easily. I think a PDF with a little love and attention can probably be the same thing.
Search Volume Test: The Keyword Planner vs Reality
What we were looking at here was what the keyword planner in AdWords give you and what actually exists. The keyword planner is pretty much what every tool uses to extrapolate search volume data from. So, when you put in a particular keyword into any tool, I would guess that every tool is using the AdWords planner API in order to predict its estimates as to search volume.
I noticed a couple of years ago that in AdWords data, what they give you as estimates are really not close to reality. They just completely go off the rails from time to time. In my own practice, I had some client accounts that were very aged and we could go back and get a year’s worth of data on some specific keywords and that allowed us on this test to really take a nicelook at real data for a year, what is AdWords telling us in regards to estimates?
I could see AdWords cop out when you do a month’s worth of AdWords campaign and you get a couple of different data and they say that it is such a small sample set, that is only one month and you need to extrapolate for a year. What is nice about this is being able to extrapolate for years andbe able to see some differences.
In the end, what we are looking at was that there was a dividing line and it seems when you are above 200, the planner is not bad now, at this very moment, and what I am estimating right now with my own clients is a 20% below or a 20% above. The planner tells me a thousand, I am telling client’s right now that the search volume is somewhere between 800 and 1200.
The real important point is that when it is 200 and below, it is a complete lie. I do not think Google has any concept or real way to know what those numbers are when they get below 200. If they tell you a hundred, it could be a hundred and it could also be zero. It could be even more that 200.
There is a lot of variance there. If you want to test this on your own, one thing I would encourage you to do is to check out if you put in a particular keyword in search and then scroll to the bottom and see suggested searches at the bottom, things related to your search. Grab one or a bunch of those and throw them in the keyword planner.
What I found with those is that they often come out with 0 results. That is actually impossible because in order for those things to show up as related to your search, somebody should have searched for them before. Google isn’t creating brand new keywords. What they are doing is they are aggregating things that have already been done so when you look at suggested reports in the SERPs and you put them into the planner and you see zero, you know that there is a disconnect.
The data that we can get from AdWords on the lower end or longer tail keywords are going to be very off. In your own practice when you are looking at long tail keywords for example, it might not be a bad investment to do a smaller AdWords campaign to see if you could get better results as to what you are actually playing with.
For example, you’ve got 15 long tail terms and the planner is telling you they are all 200, it might be worth your time to do a months worth of research on them within an AdWords campaign to get better search volume so that you could get better informed decisions on which pages to build out for your organic search.
Dori: I like this one because it gave me more inspiration to go after longer tail keywords that are 200 and less knowing that there are maybe more or maybe less too but it is true. It can’t be 0 if they are related search terms so there might be something goofy going on in Google there.
I think they use different methodologies and as a result, the AdWords planner and assessment, the sampling is too small that it couldn’t give you a nice accurate prediction until it gets to that 200 mark and then the subset is so much larger it can give you a better idea of what is going on.
Blackhat Special: Do .gov Redirect Links Affect Rank?
This test popped up as a result of a conversation in the SIA Skype group. If you are not part of the Skype group and you do not have a Skype account, I recommend getting yourself one because there are a lot of fun things that is discussed there from time to time and this is one of them. I remember this technique from a while ago where you can go a .gov link by hacking the redirect page on a very old .gov site.
This is a security flaw that when you click on old websites to leave and you go to redirect page for a minute and it would give you a countdown page from 5 to 0 and it says “Hey, you are leaving our site and going here, 5, 4, 3, 2, 1” and then you are redirected. You can create that redirect page simply by knowing the parameters in the URL string. Essentially, you would create any link to any page you would like from this redirect page.
This technique from a while ago would give you bunches and bunches of .gov links going to your sites and the presumption being that .gov links are as good as you can get. The .gov links do not arbitrarily link out, there isn’t any spam .gov sites, if you get any .gov link, presumably that is the best. Maybe .edu after that and the rest from there. We know this technique but does it actually do anything?
Part of my concern is that these pages are temporary. They do not actually show up until someone goes to them. My thought was that they would never pass link juice because they do not necessarily exist but I ran them through a link indexer so that Google would actually go through them. I did the same thing with those links in Majestic. Out of those 5 links, 4 actually showed up.
However, they did not move any kind of rankings. It didn’t move anything at all. In a follow-up, we actually blasted links through those pages with the hope that maybe they would pass juice and affect rank. That follow up will be on the next dossier. But on their own, they probably won’t move rank. It would look pretty cool and you’d feel like a ninja. Unfortunately, it would not do that much for you.
Determine the Relative Strength of an Internal Link
Silo-ing is a real pet project of mine. I am a silo advocate. If you are building a site architecture, I recommend silo-ing. The thing to keep in mind in silo-ing is say you have yoursite.com/targetpage and by target page is what you are trying to rank for a specific keyword and you want that page to show in organic results. How you silo that is most often through a blog but it could also be done through pages but in this scenario it would be through a blog.
You would have mysite.com/targetpage/nameofblogpost so that URL structure for that blogpost contains the target page and then it contains a link on that blog post going to that target page. That is called a virtual silo and it does not need a child-parent relationship to do that and you could handle it through your blog. There is a little bit of tagging expertise to it but it is not too terribly tricky. That is what we are talking about here when it comes to silo-ing.
The yoursite.com/targetpage should contain your keyword. So, if you are selling red bicycles it would be yoursite.com/redbicycles and then you’ve got your content that supports that page and you might have yoursite.com/redbicycles/5tipstofindthebestredbicycles and that is the post that you have written about red bicycles and in the URL structure, you could see that the red bicycles page is in that URL and there is also a link going from that page to the top level page.
What is really awesome in this test is we were testing just a Google doc link. We were pitting internal links against a Google doc because we know that passes juice and it is probably in the lower end if we have a baseline of which passes juice. It would probably be the Google doc just on its own.
You haven’t blasted any links to it and you haven’t done anything else to it. It is just something that we know passes juice. Google doc is something at the bottom of the scale in terms of power. We also know that the DA (pick your metric) is very high but we know it is a baseline link.
We learned that two silo’d links beat the baseline link. So, that is two posts properly silo’d and linking back to a target page will beat one baseline link. This is exciting because when you think about building even a Google stack, there is more effort that goes into that than building a couple of blog posts.
Your own site or your client’s site probably have a lot of unutilized blog posts that aren’t properly silo’d and they aren’t properly linking somewhere. You can immediately get an equivalent link juice pass just by reorganizing your site.
Last year, I talked about this in SEORockstar where I silo’d a site with three years worth of blog posts that were really well-written posts. I silo’d them all and pointed them to the proper pages and did some on page optimization and I got first page rankings out of it. That is when I realized that silo-ing is a very powerful thing.
How many Google Docs do you build from one Gmail account?
If I am pointing them to my money site, I only do one. I use one Google doc from one Gmail account and I do not use it again.
Can you set up a silo on an old site without changing the URLs?
From a WordPress perspective, you have to change the slug – that little piece past the .com. You could still point from a random page in but that random page, if it doesn’t have the target URL page in it, it is unsilo’d at that point. In this test, Ryan did that and it is just not effective.
What if someone wants to look at your posts and your categories?
What I do when I set this up is my next and previous stay within one category. Say, when they click next, usually in that scenario, you would go to the next most recent post based on time. In this, I would send them to the next most recent post in this category.
Dori: You do this manually right? It automatically goes by date in WordPress
Yes, that is correct and WordPress does not automatically categorize by pages. What I did was build a plugin for this. You can then categorize the pages and match the slug so that when you click in the post, this category goes to this category and it matches up and points correctly to your target page.
The only URLs that I would change would be supporting content pages. I wouldn’t go in and change a top level page that is currently ranking for something. Where you would look at changing the URL structure is in your blog section or your supporting content pages. The reason is when most people write a blog, they are writing great content about the things that they do or the things that they sell and you have to kind of cross your fingers that Google makes the connection between this content and that page.
What you would do in a silo is remove the mystery for Google and let it know that this page is supported by this piece of content. That is what the silo structure does so that when Google crawls through those folders, it would crawl through your target page and then get to the supporting content and then there is a link from that content going to your target page. It creates a silo which is a contained environment. The idea is that Google would just crawl around that particular category seeing all those content.
For Google then to go to the next target page in that site, it has to crawl up and out of the silo.
On an e-com site, is it best to leave the category URLs on the product pages?
I actually discussed this a little bit on e-com structure. If the category is part of the search, I leave it. If it is something generic like “products” I leave it and change it to something that is a target term.
Most out-of-the-box things use mysite.com/product/product-1 and the idea is that all your products are under the product category. Nobody searches for “product red bicycle” so that is a disadvantage but if this category is something that is actually part of the search, then it would be advantageous to leave it in. A direct example would be “services” – sometimes the word service is something that somebody would search for. “plumbing services”, for example, would be a good one to leave.
Is Latent Semantic Indexing (LSI) a Ranking Factor?
LSI, the way I define it, are terms that naturally come up in a conversation about a particular thing. So, if you are having a conversation about kitchens, you might say things like sink, cabinets, pantry, etc. – all of those terms are LSI.
Most people define it as synonyms or variations of a term but I think that is incorrect. In the kitchen example, it would not be variations or synonyms for the term kitchen. What we are looking at is if LSI terms are a ranking factor. If you do a search on that, you get a wide array of opinions. We want to narrow down and see if it is actually a ranking factor.
We used the tool ntopic.org which is a tool that spits out LSI possibilities and it gives a ranking of how LSI-d your content is. Whether or not the tool is accurate is not the point but we want to be able to show that we have two pieces of content equally optimized but one piece has LSI terms on it and the other doesn’t. On the first one, the tool was showing us that it had 80.93%
LSI. Page 1 was at 80.9 while page 2 was at 36%. What we did in the situation is that we had theexact same number of words in the lorem ipsum. The optimization was exactly the same. Wechanged out some words in the lorem ipsum for both. Page 1 had LSI terms while page 2 justhad random words that aren’t LSI terms and that has nothing to do with the target page.
I ran this multiple times and in the end, the page with the LSI terms wins every single time against a page which was equally optimized for a random keyword. The LSI terms were just in the body content. The number of words between the two pages are exactly the same and even the spots where the LSI terms were and the random words were added, were exactly the same.
We made sure that the random words that were chosen were not part of the recommended LSI terms in Ntopic.
User Metrics: Is time on site a ranking factor?
This is a follow up on the bounce rate test that we did in the previous month. The idea here is that we are looking at user metrics. The assumption is that these things are part of the algorithm. It is a broad assumption across the board, that there are things that Google clearly see in Google Analytics and the idea is that if Google can see that in GA, they can then use it and extrapolate that data to qualify how good your content is.
If people are engaging with it, if they are spending time on your site, if they are clicking through and seeing things, those would indicate that you might have better content than somebody else and that would be a ranking factor. The purpose of these user metrics are very obvious. Are these things actually a ranking factor? We found that bounce rate clearly is.
In this next two, we were looking at time on site – the dwell time or the time that somebody sites in page. Then the other one is conversions and how that is done is you can set conversions in Google Analytics and it is something that we have recorded.
This is a great theory and I liked it a lot because the concept is that you told Google what is most important to your site and this is what I am trying to track. The idea then would be if you succeed in that, you have proven that your data is appropriate for that search – would Google reward that and can you move up in rank?
With the time on site, we can’t say that it is a ranking factor at this time. There were a lot of problems controlling people that were doing the actual tests. We did get positive movement with the time on site which we were happy about but we cannot conclusively say that time on site is a ranking factor in the same way that we can with bounce rate. What I am telling clients right now is that it is probably part of the algorithm and we work on ways to make people stay longer. In terms of testing we could not say a 100% that it is a ranking factor.
The obvious thing is that different sites have different content so that they are going to have different bounce rates and time on sites. I would imagine there would be a kind of interplay and we see in the result that came up here. The page that moved up to number one and stayed there has a much better bounce rate than the experiment page and has the second most visits.
The experiment page had the most visits and the best time on site. There are some interplay with all those concepts – an equation where in bounce rate may be more important and maybe x your time on site x your visit.
I would encourage you to do something on your site to make people stay longer. Bounce rate may be more important.
User Metrics: Is a Google Analytics Conversion a Ranking Factor?
When it comes to conversions, unfortunately, we can’t say that it is a ranking factor at all. In the end, our test page had 60 conversions compared to no conversions on all the other pages and inexplicably, the test page dropped to the bottom and I have no idea why that happened but that salts in that Google is not taking an extra step to see how conversions are.
They are not individually adding that metric into the algorithm at all. They are not doing a per site evaluation, e.g on this site has these important thing and they are killing it so their content is better – they are not doing that and it makes sense.
Dori: I think it is kind of stupid on Google’s part not to do that. If somebody puts Analytics on their site and they know they are converting, it seems to me that it is a good site and people like it for that keyword.