Using link data, although it can never be perfectly accurate, allows you to take a more scientific approach to your SEO strategies. How can we leverage link data for actionable insights? I recently wrote a post breaking down the Mormon SEO strategy, which is a very well thought out plan of action.. There seemed to be a lot of interest in the tools I used and the insights they provided. It's impossible to cover all scenarios, but my hope is to show how this data can be used.
Tools Used
- Excel
- SEOmoz API / Open Site Explorer
- Free Version of Majestic SEO
- Ninja Skills (prerequisite of working at Distilled)
For most of this post, I’m going to look at the SERPS for "nashville real estate", a competitive result from my hometown.
The Data Set
To start, I pulled a wide range of link based metrics from the Top 10.
Caveats: Your may want to remove large domains and exact match domains from your analysis. Large domains with solid brands may just add noise to your analysis. Exact match domains may be getting a substantial boost and may skew your analysis as well. I have removed both Trulia and Yahoo Real Estate from this set.
Data
- PA – Page Authority, metric from SEOmoz
- P LRD – Number of Linking Root Domains to the PAGE
- DA – Domain Authority, metric from SEOmoz
- R LRD - Number of Linking Root Domains to the DOMAIN
- Links – Raw number of links (SEOmoz data)
- Deep Links - % of root domains that link to content that isn’t home page
- mR – mozRank, SEOmoz’s metric similar to PageRank
- DmR – Domain mozRank, similar to mR, but on a DOMAIN level
- DmT – Domain mozTrust, SEOmoz metric to measure TrustRank
- EA LRD – Number of Linking Root Domains using EXACT match anchor text
- EA % - Percentage of anchors with EXACT match anchor
All of this information was acquired from SEOmoz’s data set. Items like deep links %, exact match linking root domains, and percent of anchors with exact match anchor were all obtained with some basic analysis. Now that we have our first data pull, we can dig in and talk about what these mean.
Breakdown the Top 10
If my goal is to break into the top 10, I’ll line my site’s metrics up against the trending of the top 10 and their averages. This will identify areas where I’m over- or under-performing. In addition to comparing against the average, I can look at the distribution of metrics across the top 10. I’ll throw out a few examples in this niche that jump out. The factors I look into may depend on the data pulled for a niche.
Link Root Domains
As we know, the number of linking root domains correlates well with rankings. This why in my link builder tips post I suggested getting a link, or just a few, and moving on from a domain. The majority of the sites in this niche sit around 100 to 200 unique linking domains. This graph gives you a gauge for breaking into the top 10 results. However, the top two results have pulled away from the rest with around 800 unique linking domains, so ranking number one may be difficult.
Exact Match Anchor Linking Root Domains
This metric was measured by looking at the Anchor Text Distribution tab on Open Site Explorer. I was a bit surprised by the results. The number one site has managed to receive links from 544 domains with the exact anchor text “Nashville Real Estate.” Regardless of how they managed to do so, this makes it a highly competitive term to rank number one for. However, it also shows that the majority of the Top 10 do not have a large number of domains linking with well-optimized links for this phrase. This graph, and the one before, shows opportunity exists for this keyword between result 3 and 10. This is why it’s important to note the distribution as well as the average.
Percent of Anchors with Exact Match Domains
Looking at the percentage of anchors using this exact match phrase can provide some interesting insights. First, the top ranking sites are excessively optimizing their link’s anchor text. Second, Google is allowing this significant optimization. Even the third result has nearly 30% of their links using this anchor text. Your link building strategy here could be aggressive without sticking out. Additionally, these sites are overly focused on one golden phrase. I’d have to dig deeper, but this may be indicative of several open opportunities in mid to longtail phrases. I’d likely take advantage of this short-sighted targeting strategy by first focusing on picking up the terms they’re ignoring.
Strategy
To rank well in this niche, you’re going to need a large number of links from a diverse set of domains. Google seems to be forgiving of over optimization and has allowed aggressive link building. The first ranked site appears to sufficiently answer the query, so I don't think there is a reason to not allow it. Most of these sites do not have significant content strategies pulling links to subpages and most do not engage heavily in social media. Although an aggressive strategy would not stick out in this market, I'd caution again the aggressiveness of having almost 70% of your link profile with the same anchor text. This type of optimization is dangerous, limiting and ignores many other phrases.
Link Profiling
Dr. Pete wrote a great post on link profiling, which allows you to see the distribution of inbound links to a domain. This will give you insights on the quality of the links a site has acquired. I also use this when reporting the links I’ve built to my clients to visually demonstrate the distribution of links I've acquired. I’m going to briefly show an example of this profiling using Page Authority, but I repeat this with metrics such as Domain Authority, mR, DmR, and DmT. Disparities in these distributions can draw your attention to items you might not have considered.
Radar Graphs
Not a new tool, but I like the radar graphs in SEOmoz Labs. These are great for demonstrating link analysis or providing a quick audit on a phone call or email. Let’s say the #12 ranked site nashvillebuyers.com called up and quickly wanted to know what it’d take to break into the Top 5. This quickly shows they're falling short on several significant ranking factors. This could also show that they’re beating other sites in all metrics. If this is the case, I might start looking at other factors like anchor text, anchor text distribution, link quality, and on-site targeting.
Top Content
After looking at a site’s links, it’s important to evaluate what’s the top content drawing in links. One of the best guides I’ve seen about visualizing this analysis is the post What Are My Most Linked to Subfolders? written by SEOgadget. This report can be pulled from Open Site Explorer’s Top Pages. In my Mormon SEO strategy post, I used the SEOmoz API and created a pivot table to find the most linked to content in their massive link profile.
Keyword Tag Cloud
A tag cloud of anchor text is a quick way to visualize the distribution of keyword anchors and pull out the terms being targeted. The most interesting anchor cloud I’ve seen so far is that of the Mormon LDS church.
To get this, I pulled the max number of links from the SEOmoz API, saved the list of anchors to a text file, and uploaded them to Tagxedo. Doing this quickly highlights major targeted terms and visually demonstrates the distribution of anchor text.
Broken Links
Another quick check that can bring you big wins is checking sites for 404 pages by looking at Open Site Explorer’s Top Pages tab. (Hopefully this isn't new to you guys, but worth mentioning.)
Reasons to Find Broken Links
- Great way to start a relationship with a webmaster: let them know, people like to reciprocate.
- Contact the people who are linking to them and get that link.
- Generate content ideas based off that type of content that has acquired links before.
- Recreate this content and get websites to switch the link.
Link Growth / Velocity
One more check I like to make is link growth rate. Aaron Wall wrote an article a while back about link velocity and the role link growth rates has on your link profile. MajesticSEO provides these graphs up for free and they can provide some interesting feedback. They provide link discovery and cumulative link graphs.
What we might be seeing here are a few strong pushes early on, followed by very little promotion in early 2007. At the start of 2008, the link building ramped up and remained constant. This is indicative of regular on-going link building, which is no surprise considering the optimization of these efforts. It’s also important to note that it ramped up even more during the last half of 2010. If this was your competitor’s profile, this is valuable information to have. This link growth rate makes this site even more competitive.
Now a profile like this is a bit more expected. They seem to have had a strong push at one time, but the link acquisition has faded over time. They have periods of spiky growth and have had minimal link growth in the last part of 2010.. This is a good sign if you’re looking to outrank this site.
Chopping Up The Link Export
One of the last few questions left is how and where are they getting these links. How can you chop up their link export to get actionable insights and link prospects? Let me just cover a few quick ideas.
Export CSV from Open Site Explorer, open it in Excel.
Filters & Sorts
- Filter Sites with Exact Match AND Phrase Match Anchor
- Filter by Branded Anchors and People's Names
- Sort by PA to get strongest PAGES
- Sort by DA to get strongest DOMAINS
Questions to Ask
- What content is attracting these links?
- What types of sites are these links from?
- Are they manually building these links? (directories, articles, guest blogging)
- Do they appear to be paid?
- What's the IP address and WHOIS information on anything fishy?
- What strategy is working here? Or is there even a strategy?
- Where are they getting the branded and name links?
- What communities are they participating in?
Perform a search against page title and URL
- How many links come from pages about same keyword?
- How many links include mentions like directory, links, resource, article, forum, etc?
Export into Google Custom Search
Yahoo use to allow you to perform search queries against a link profile's content. However, with the death of the Yahoo! linkdomain, this is not longer possible. Luckily, you can use Open Site Explorer and GCSE to do the same thing.
Search Ideas
- Sponsorships queries
- Paid link footprints
- Blogging footprints
- Person’s name
- Brand name
- Forum footprints
- Embed footprints
There a lot of different queries you can run, these may depend on what information you’re looking to get.
Although this is not an exhaustive look at analysis techniques, I hope it helps give insights into how link data can be used to create a strategy or understand your competition's. Feel free to find me on Twitter if you ever want to chat more about link analysis.
Learn More:
If you’d like to learn more about link building, we’re hosting two Linkbuilding Seminars in London and New Orleans this year.
You know it's a good post when an hour later I'm still reading it and it's related sources...
haha, thanks. :)
Great post! Too bad it proves that agressive heavily optimized anchor text linkbuilding is a win :(
Nice post I like how you have broken down each section into an easy to understand guide. Competitive link analysis is a big area of SEO. I think OSE is one of the best tools for conducting the research and also throwing in the data from Majestic SEO too. It is good that SOE moz has developed these tools as with out them the link research area would be alot harder with all the restrictions that have been placed on Yahoo link data for example.
Kepp up the great posts and good work seo moz !!
Thanks!
And I agree, the value of OSE / Linkscape just keeps going up as link data gets hard to get from search engines.
Seems they want to make it even harder, I mean we still have Google Webmaster tools and Bing webmaster tools to pull some data but the good thing about OSE is that it is inderpendant and it provides some real kick ass data. But yeah Majestic SEO is also a great player in the market I see alot of companies are feeding into their data.
Thanks for the MajesticSEO comparison, we are constantly looking at improving our link acquisition process and this blog has given us a new angle.
Hi Justin,
Appreciate your great effort for composing this practical post, I think it deserves further promoting, now it has Chinese version for 1.3 billion people to enjoy and share, hopefully you like it anyway.
Thanks for the translation! :)
It's a pleasure, Justin :) Looking forward to more masterpieces from your end.
This was extremely helpful. I am new to SEOMOZ and learning about the link building tools is very helpful.
Is there a general limit to how many links to add in day? I did about 10 in 90 minutes. As I get faster at this, is there a rough number or does it depend? my site is about 46 pages, slowly growing, and about 2 years old.
Thank you. :)
There is never a limit to the number of links you acquire a day. The issues to consider are: 1) what's the link growth in my niche like and will I stand out 2) what % growth of my profile is this 3) why am I gaining links, and is it a legit reason.A piece of viral content could gain hundreds of links in a day, but there is nothing back about this. But I wouldn't build 500 directory links in one day.
If you have a content / marketing backed campaign, I wouldn't be concerned. These are all natural links. If you're "building" links, you want to be more considerate of how you get them.
Ah thank you. Those were all 'building' links. So its not just the links as a number its how and where they are acquired. Got it. Thank you very much.
Hi Justin,
Great post and I've had it bookmarked for ages waiting for a time when I could do a full competition analysis. I'm a little unsure of how you went about getting the metrics in your very first table. Did you enter all of this manually?
I have the same question, and also in particular how is measured the % of deep links?
This article is really great and usefull, thanks a lot.
+1 on this question
As I'm working in a more competitive field than your Nashville example, it often feels like an uphill struggle. So actionable posts like these can really spur me on when I feel like I'm fighting a losing battle with link building. Great post, thanks.
Filtering tips for Excel:
=COUNTIF(SHEETNAME!C:C, "main keyword ") - Exact match anchors
=COUNTIF(SHEETNAME!C:C, "*keyword*") - Keyword in phrase
SHEETNAME - what you have named the tab where the OSE link report is
C:C - the column where the keywords are in your spreadsheet
I think Mike will be publishing something cool soon on Excel, with all kinds of cool Excel tips and hacks.
Good stuff, I'm seriously in need of an SEO's ultimate guide!
That will be cool. Mike, keep us posted when you get the article up.
One way you can filter your data out after you've exported the CSV is to create a pivot table. It gives it a clean look, while tallying up certain pages/root domains/anchor text used on links coming to your pages. I've always found this to be quire useful when you're looking at the overall site's inbound links.
Why no "love" for the Mormon Church?! You reference them all over this post and yet only link to your post on Distilled...twice. Why wouldn't you link to their website. In going to the site, I see the name of the Church is "The Church of Jesus Christ of Latter-day Saints" and Open Site Explorer shows that name or some variation of it as the top anchor texts used. This provides some insight around the tag cloud. All of the words in that tag cloud could be considered "branded" terms which they should naturally be doing well for with or without a linkbuilding strategy. Right??
Thumbs up for the rest of the article though. This is a solid piece of work and will probably be one of the more "linked to" posts on SEOmoz. Congrats on that!
The not linking wasn't intentional. I was focused on citing resources for learning more. I just didn't think about it. I did link to LDS sites 4 times in my article on Distilled.There is lots of great comments on my Distilled post from members of the church regaurding their branded terms. I'd recommend jumping over there to check them out. It might be a better place for that discussion.And thanks for the comment! :) - hope it was valuable for people.
Lots of info here. Thx for the analysis :)
Gary
Great post, I love your analytical approach to the subject. It gets me wondering about about automating the workflow using SEOMOZ api. I just wish I had more time :-)
Yeah, I was chatting with my developer over the weekend about how it'd be cool to automate some of these processes.
Hey Justin, what programming language is your developer using to build out the api?
Most of the hacky tools he builds are in Python
Definitifely on my to do list!
"Ninja Skills (prerequisite of working at Distilled)" is really what the cherry on the icing was for me! Brilliant Post and excellent analysis.
Wow Justin,
you did really a great job with this so actionable Guide...
I especially appreciate:
Finally, as here you talk a lot of OSE, I would like to recall a post by Fabio Ricotta that was exactly about how to use Open Site Explorer as a competitive analysis tool: An Inside Look at Competitors Backlinks with Open Site Explorer.
Thanks for pointing to Fabio's post, I'll check it out.
What a fantastic post Justin.
I have two questions for you:
Love the questions at the end and your search ideas.
Thanks!
To answer:
1) I mostly use SEOmoz, and tend to only peak at Majestic for the velocity graphs. I'm just a fan of SEOmoz and like their link information. This is just my preference.
2) I'd say no, but I'm a bit unclear on the question. I think link metrics are improving, but none of them are perfect yet. However, they're the best available right now.
(If I completely missed that question, let me know)
Maybe I didn't investigate well enough, but do you know how SEOMoz calculate mR? This is what I found: https://www.seomoz.org/learn-seo/mozrank I guess it doesn't really provide a great level of clarity for me, perhaps because I am not a mathematician or a programmer. So basically what I was saying is that the metrics provided seem a bit hazy to me and I would like to understand them all better before trusting and making decisions. I guess my only way of knowing is correlation and testing.
The SEOmoz staff might have a better answer on this. You could maybe ping Aaron in customer service to see if there is an answer. My guess is that the exact calculation is proprietary information. It's modeled off the concept of PR, but I don't know the math they're using myself.
From what I know, they put testing into metrics like PA and DA, and have published several posts regarding those metrics. I'm not sure exactly how they're calculated though.
Great article Justin. I did want to mention that your top site in that analysis nashville.mls.com is the Multiple Listing Service for that area, which is where all real estate persons will upload their properties and link to for clients to do searches on. That is why it is so much further ahead from the other sites.
It's Nashvillesmls.com, which isn't, to my knowledge, an official MLS, but is owned by someone, maybe an agent. I may be wrong though. I didn't look into why they had those links.
Nice intro to link building analysis. Just reading through, I realize its definitely a lot of work, prepping and planning as well as deep understanding beyond just reaching out to a random number of blogs.
The strategy insights you share is definitely a great starting point. Time to open up my excel and learn to be a Ninja.
A good post broken down and explained. Back link analysis is important for research purposes and link building ideas. This post explains more in depth how to analyse and choose good quality links. I have always been told that keeping our back link profile high quality and industry specific is more important than just promoting the site with links to anywhere.
It is nice just to read a more advanced post on analysis and I will definitely be looking to take on board these methods.
Thanks for the free tool from SEOMOZ, I generally go for vendors or from some online tools.Thanks
How do you work the deep links % in your table?
how do we extract the data in the spreadsheet you showed us here more easily in seomoz an OSE?
wow.. Fantastic steps.. I really like it. keep up :-)
Hey Justin awesome post.
I totally agree with you about removing large domains or solid brands. It just messes up the data and can be misleading at times.
Thanks for sharing. Cheers!
It is a nice post Justin your ideas are natural and helpful for SEO also by breaking the whole topic you make it easy to understand the meaning what you are going to tell the audience. It is a top content according to my opinion, your presentation makes the topic too easy and understandable.
I have a concern. What is more important to Google, content or links?. Well I've seen sweb without quality content very well positioned,
thanks
Justin, great post. Any change you can share the API call you used to pull the back-links for the TagXedo.com keyword cloud?
Thx
UPDATE: Can get the anchor texts from OCE...still be nice to see the API call -:)
Great post. I did something similar to this last week but not as indepth... Difinetly helps determine the focus on where to start on the back link campaign to be effective against competition...
The data shows how to works it. and how to increase the backlinks of the websites.
Very cool post. I love the use of Tagxedo.com with the Linkscape API too. What a cool way to visualize the anchor text usage of inbound links for a site!
Thanks for sharing your ideas about link building. For beginners like me, this will truly be a big help in order to further improve my skills.
great post. only thing else i would add is exporting the top 10 data into excel sheet and compare it to the website you are trying to rank. i usually include and count the estimated total anchor text of each site and add that number into the data which helps you know how many quality link you need to build with that anchor text to be able to compete with your competitors.
SEO is all about exploiting your competitor's weaknesses and overcoming your own. Analysis techniques like this are vital to doing that. Knowing what other sites are doing is a huge part of making a winning strategy. Identifying a weakness like intense over-optimization is just one great thing about good analysis.
Justin,
You have provided some great data and insights into why competitors own certain keywords. I think you call out the importance of having a variety of linking domains and offer some good strategies. It's hard for Google to really flag a site for link spam for aggressively building links if they are coming from a lot of different domain sources. It looks way too natural and hard to separate from those social/viral links.
Justin, great work here. You really nailed down the methodology I use too. OSE has so many uses, you just need to know how to slice and dice the data. Thanks for giving us such a great visual.
Great tips here, though at least for some of these metrics a program like Market Samurai or SEO Spyglass would save a LOT of time
I've used Market Samurai, but not SEO Spyglass. These days I work mostly with SEOmoz data, but for massive projects I'll ask Ben to build something out to use the API.
Mmm... that would be a blasting tool, if Ben himself would work on it.
He has built some really cool tools. I'm impressed regularly by stuff he puts together. He helps save me a lot of time.
What do you do when there's huge discrepancies between the data from OSE and SEO Spyglass? I've just started using SEO Spyglass, and see some pretty big differences between, say, the number of links metrics between the two tools, which is confusing to me.
Ian
I've never used SEO Spyglass, so I can't speak directly to its data, but it sounds like the same issue as comparing MajesticSEO data to Open Site Explorer data. The numbers just aren't anywhere near close.The advice I've given in the past is: don't try to compare cross platform data, but just data period to period (or site to site). Look for trends and changes, or differences within the same data set.These tools collect, store, and process data in very different ways. I have my personal opinion on the ones I like, but some SEO's like other data sources more.I'd either collect from one source, or from multiple sources, and report on just the one or all. None of them are completely accurate. The differences are a whole post itself and I've seen some threads and posts out there on it.I'm a bit bias on the tools I like though.
thanks for share bt sir my site da nd pa very high bt not rank why ??
site punsong.com
Hi there amlijatts, thanks for commenting! Because this blog post is older, you're more likely to receive a helpful response by posting your question in our Q&A forum. It's a great place to crowdsource answers to questions like this with experts in the community :) Hope that helps!
Thanks for providing a thorough guide to backlink analysis. However, my intuition is that page segmentation a hugely important factor in understanding link authority, albeit a factor that is really challenging to get a handle upon. Page segmentation seems to be a factor in Adwords Quality scores, so I assume it also is an organic algorithm factor.
When working for a consultant. we created a page that included a snippet about each of the consultants, listed in order of seniority. I bought the names of each of the top rain makers via Adwords and linked to the page with the consultant bios. There was a direct correlation between how high on the page the consultants' snippet was and the quality score of their name as a keyword.
Complete, easy to understand and keep open the door to diferents points of view.
This is the kind of work that is easy to understand but take a long time to get at this point uh?
Congrats!!!
I really enjoyed the post mate, thanks for taking the time to break it down like that! Stumb and Tweet from me sir!
Dude, what happened to all the images? I was just showing my boss how awesome it would be to emulate your competitor analysis reporting.
Same here @DarinH. I just hit him up on twitter @justinrbriggs, hopefully he can get the images back.
Justin, great post. Perfect timeing as I'm looking at various backlink visualisation tools. Speaking of which I came across (and lost) a tweet about an interesting new backlink visualisation tool that created similar graphics that were created by Touchgraph. I don;'t suppost you or any of the other readers here may have seen it?
Cheers
Ed