Link Building Evaluation
Link Building | SEO Analytics | Whiteboard Friday
SEOs can use metrics like Domain Authority (DA) and Page Authority (PA) to assess the effectiveness of their initiatives (PA). But how should they be used? We’ll demonstrate how to think about these metrics as part of a holistic approach to link building analysis in today’s Whiteboard Friday.
a whiteboard with measuring recommendations for link building techniques
Open a high-resolution version of the whiteboard image above in a new tab by clicking on it!
Transcription of Video
Good morning, and today’s Whiteboard Friday topic is link building measurement. So, obviously, this is a large and old topic in SEO, and it’s one that SEO Visionary Experts, as a firm, is highly concerned about, right? Two of them are Domain Authority and Page Authority our most popular products that are frequently used for this purpose.
But this isn’t going to be an advertisement. I could stand here and say things like, “Obviously, these are the top measures in the world,” and so on. That is not my purpose in being here. I’m here to provide some nuance on how to utilize these metrics, how to think about them, and how to combine them with other metrics, rather than just having one tool and claiming it solves all problems, which isn’t always the case.
PageRank from Google
To do so, I’m going to start from the beginning, with Google’s PageRank model from 1998. Now I realize that a lot has happened in the world and with Google since 1998. But this was Google’s initial way of thinking about links, and it’s still the best we have to go on in a lot of respects. Many modern SEO best practices and beliefs are still based on this original idea, with the exception of a few things we’ve picked up along the way that don’t really have a basis in anything Google has said or done, which is why I’d want to point them out.
So PageRank began as a method of leveraging links to estimate the likelihood that a user will visit a website, which is already intriguing because it indicates that this is a popular model. As a result, when we discuss this now, we frequently discuss concepts like trust and authority. I’m sure those are relevant, but it’s worth remembering that this was originally merely a method of gauging a page’s popularity.
It’s worth noting what I said about the page, not the domain. So, imagine a world where the internet only has one page, which I’ve dubbed Page A. Now, if there is only one page on the internet, estimating the likelihood that a random browser will visit that page is not difficult. They’re almost certainly on that page. It’s still not that difficult if we introduce a second page, and we just assume it’ll be 50-50 and so on.
That’s the sort of chance we have to deal with as a starting point. But then, when one website links to another, we may go off on a tangent or add a little spice to the scenario, which is obviously what we’re interested in. So, if A links to this second page and there are only two pages on the internet right now, ignore these other boxes; they’ll appear later; there are only two pages on the internet right now, and A links to the second page.
We claim that this chance is passed on 0.85 times. Now, 0.85 is a somewhat arbitrary constant. It’s from a Google document from a long time ago. It’s probably not that exact figure, but it’ll suffice as an example, and it’s the best we’ve got.
So, why did we say 0.85 in this situation, by the way? Why didn’t we say that everyone on this page clicks through? That’s because we expect some of them to go off and do their own thing, to stop surfing the web and do something else. This dampening factor turns out to be critical in a world where sites link to each other in a large web rather than simply one link in one way.
So that’s all good and well, right? What if we added a third page to the internet by adding a second link? As a result, this is still a pretty basic model. We have three sites and two links on the internet, and the links only go in one direction.
This is quite straightforward. However, we can’t have both of these pages acquire the entire probability in this scenario. No, the users aren’t going to both sites. They’re on their way to one of them. As a result, 0.85A is divided in half. But then there’s this one.
In a more complicated model, we might say, “Oh, one of these links is more likely to be clicked on, thus it gets greater probability,” or something along those lines. However, in this simplified version, we’re implying that it’s split in two. Now, we’ve already learned something significant in this example, because by adding another link, we’ve diminished the value of the existing links, which is something we rarely consider in the context of link building.
But that’s what we’re thinking about when we talk about not having too many links on the top nav and such in technical SEO conversations. We’re attempting to concentrate our strength where we need it the most. Finally, I vow that the [indecipherable] will come to an end shortly. Finally, what if this system experienced another jump? This 0.85, this dampening, happens again in this scenario.
So 0.85 multiplied by 0.85 equals 0.72, which is less. So it’s now 0.85 times this page above it, and it’s dropped even further. This is why, as technical SEOs, we get caught up in things like chain redirects and the like, and why we believe they are vital.
That’s where ideology like this originates from. So I’m not going to elaborate on this simple PageRank explanation any further. I’m attempting to bring a couple of things to your attention here. One is that the value of these connections is heavily influenced by the characteristics of a page, such as the number of links that the page sent outward and what linked to the specific page.
Note that I made no mention of domains in this section. This might happen on four separate domains. It could be contained within a single domain. Here, we simply discussed page specifics. In terms of how they think about pages vs domains, Google has been a little confusing over time. However, they claim to be more concerned with pages than with domains. So that’s fascinating, since these may all be on the same domain, yet this page could be significantly weaker and pass on significantly less strength than this one.
Metrics for developing links
That’s intriguing, and it’s something we don’t generally consider when it comes to link building. So, if we go back to what I said I was going to talk about, actual link-building metrics, there are a few characteristics we’re looking for.
These first two are what I haven’t mentioned yet. Fast metrics are what we want. We want it as soon as possible so that we can report to our client or our employer or whatever, and also because we’re busy individuals. We do not wish to squander our time.
We want metrics that are widely understood, so when I tell my boss, “Oh, I got you a link with DA 90,” he or she or they are likely to grasp what that means. They’ll ask, “What are you talking about?” if I mention it had a Tom Capper score of 38B. As a result, I’ll need to utilize a metric that’s quite commonly understood.
Specifications for the page and links
Then there’s the page and link level specifics I mentioned earlier. So, if I consider a metric like Domain Authority, it performs admirably on the first two and mediocrely on the third, because it is trained to some extent on rankings, which is what this is determining.
So there’s an advantage there. It takes some of this into account, but it’s ultimately a domain-level metric. As a result, by definition, it must handle all pages on a single domain similarly. This has both advantages and disadvantages.
Using metrics in tandem
So here’s what I’d like to do: metrics on a chart like this and suggest how you may utilize them in tandem.
As the vertical axis, I’ve chosen actually. So the closer it’s to what we’re truly attempting to measure, which is Google’s opinion of the link’s value, the higher it will rank. But then there’s this convenience meter of fast vs. slow. So a statistic like Domain Authority is almost certainly present. It’s quite quick.
It’s really common. However, because it is a domain-level statistic, it is missing some of this nuance and is addressing a somewhat different question. The goal of DA is to answer the question, “How likely is it that a page on this domain would rank well, all other factors being equal?” That is a different matter than how valuable the link is.
Yes, that is an excellent candidate. So, like most page-level metrics in the industry, including Google’s and our own, Page Authority is initially influenced by both domain and page level characteristics. We’ve conducted correlation studies and similar research.
Because it’s a more accurate statistic that captures some of this nuance, it’s a lot closer to gauging the worth and ranking potential of a single page than the Domain Authority is. However, you can take it a bit farther.
We try our hardest, but it isn’t always possible. However, if you’re ready to wait even longer, a statistic like referral traffic could be used. Please accept my apologies for my dreadful writing.
So, with referral traffic, we’re interested in how many people really visit my site after clicking on the link I created. That’s intriguing because that’s exactly what Google was attempting to figure out in the first place. So if we can measure it, we’re coming close to what they wanted.
We’re sort of capturing the nuance of whatever sophistication they’ve built in. That has several obvious disadvantages. One is that many link-building campaigns perform poorly on this statistic, and you can draw your own conclusions. The other is that you’ll have to wait a long time for this data to become available, and even then, there could be issues with the client’s analytics or other issues. That is, in any case, what I wanted to share with you today.
Essentially, I recommend that you use all of these measures, as well as some others that you could add to this chart. So I’m curious as to what metrics you’d use and where you’d draw them on a chart like this. I drew these green lines in as a guide since I believe you could undertake prospecting and basic reporting to the client in this first section, before even building the link.