SEO

What 70+ Hours Of Backlink Checking And Research Shows…

What are your direct competitors doing in the way of getting more backlinks? My research in over 30 verticals sheds some light on the question.
This is part 1 of 2. In this part, I’ll address three issues:

  1. Practical uses for this competitive intelligence,
  2. My methodology in conducting this research,
  3. Some fun data points for niche research,

In part 2, “Backlink Research Shows How To Compete Elite (Part 2)” I cover:

  • Screenshots of the data,
  • My analysis and conclusions,
  • Some caveats.
  • I’d love to hear comments here and on my own site from people of all experience levels. There are no dumb questions: a desire to learn is about the greatest sign of intelligence!

    Just before I begin, I’d like to thank Alex Chudnovsky and his excellent team at Majestic SEO. They created the historical backlink comparison tool that made this all possible. They also collaborated with me on this and gave me exact numbers from their data, as well as lightning-fast support. (And thanks to Ann Smarty of Search & Social for making me aware of Majestic and other fantastic tools.)

    Also, I think it’s important to point out that Majestic SEO has the highest link counts by far of any tool out there. It’s the most comprehensive backlink research tool, which is valuable, but you should know that you can’t compare the numbers to Yahoo Site Explorer – it’s apples to oranges.

    Uses For This Competitive Research

    1.Spy+VS+Spy What 70+ Hours Of Backlink Checking And Research Shows... 1 – Compare yourself to the competition quantitatively, counting: i) links, ii) link value on a cost-per-link basis; and estimating i) man hours required to build equivalent links and ii) time value on a cost per hour basis.

    2 – Find out if an exact match domain is worth the price. Can you rank with fewer links from a narrower group of sources?

    3 – Find out if competitors have either strong brands, strong relationships (brands are relationships, anyways) or big link buying budgets, by looking at how many sitewide links they have.

    4 – Assess the breadth of competitors’ relationships.

    5 – Answer the question, “How fast and how many links do I need to build?” This question has never been addressed, afaik.

    6 – Make the business case for SEO by showing what the competition is doing.

    In other words, you can:

    • Find out what you’re up against
    • Determine what ingredients you need to rank
    • Project costs (which is part of forecasting SEO ROI)
    • Present the case to the boss with all the numbers

    My Competitive Intelligence Research Methodology

    Category Selection

    1. I began by identifying industries and verticals based on the Yahoo Directory’s main categories – those linked to off the directory’s homepage.

    2. From those, I picked 15 single word ‘keyphrases’ and typed them into Google. I also picked 15 two-three word keyphrases, aiming to find mid-tier keywords (judged by volume and competitive interest).

    So in total, I looked at 30 different keywords. That may not seem like a lot, but this actually took hours to compile, screenshot, and analyze.

    Also, I felt it was enough since I canvassed the vast majority of the Yahoo directory’s main categories, as defined above.

    I wanted to get a broad cross-section of the web, without going overboard and drowning in the data. I think that my approach achieved that.

    Search Results Data Collection

    joost devalk What 70+ Hours Of Backlink Checking And Research Shows... 3. I typed the selected keywords into Firefox’s Google search bar. I used Joost’s personalized search disabler to get clean search results.

    Domain Name Selection

    4. I selected 5 of the top ranking domains, with the following criteria:

    • I gave preference to higher ranking domains.
    • I excluded a domain if it was only present by virtue of:

      • Universal Search – Eg Domains present in the results by virtue of their images, videos, Google news status etc
      • Non-Industry-Specific – Eg Wikipedia, CNN, random colleges. These exclusions also covered sites which had pages of theirs ranking, but on the whole were not about the given keyword.
      • Subdomains – Unfortunately, Majestic SEO’s data only goes to the root domain. This forced the exclusion of numerous university results, standalone blogs etc.

    Competitive Intelligence Data

    4. The selected domains were plugged into Majestic SEO’s excellent Historical Backlink Tool. Plugged in, by Alex Chudnovsky and his team at Majestic SEO, who gave me the exact numbers in their data. I don’t think that’s publicly available, but maybe they’ll make it available for a price.

    As an aside, I encourage you to set the visual display in the Historical Backlink Tool to Cumulative (via the dropdown on the left hand side of the tool). This view shows

    • If a competitor is catching up to another
    • Competitors building leads in link numbers, sustaining leads, or resting on their laurels.

    Metrics

    1) Industry medians vs averages vs individual sites’ number:

    Medians:

    Because many categories have a competitor who is disproportionately stronger than all the others, taking the industry average for these numbers gives a skewed view (upwards). To be more realistic as to what it takes to rank, I therefore chose to look at industry medians.

    A median is the “middle of the pack” number: half of the other numbers in the set are greater, half are smaller.

    Practically speaking, this makes a big difference for most sites. Rather than being way behind, they’re closer to the middle of the pack.

    And for those who were leading, this typically shows a greater lead. This is more appropriate than using an average. Because an average number would show leaders competing against themselves in part, whereas a median more accurately reflects competition against others. It would be odd to say, “I have a 5000 link lead over the group comprised of myself and others.”

    And some sites that would otherwise be laggards (Buy.com and USA.gov in referring domains) turn out to be the middle of the pack itself.

    Averages:

    I initially thought looking at averages would be valuable, but because the numbers are normalized as a percentage, they doesn’t mean much. 200% growth is meaningless if you started with 100 links…

    However, I left the data in there because if you compare apples to apples, it can be interesting to see who’s closing in or who’s building their lead. In other words, look at two competitors who both had about the same amount of links last year. Who grew faster during the last 12 months?

    Individual sites’ numbers:

    You may be able to catch up to the pack (the median), but can you catch and outpace the leaders? If ranking #3 or #4 isn’t good enough for you, then you need to look at the average numbers.

    2) Year over year growth numbers and percentages:

    What is the cost of delaying the start of an SEO program?

    As a consultant, I know as well as anyone that the long sales process is a pain in the butt. But what I didn’t know was whether the often months-long drawn out process could be shown to hurt clients. It can.

    When I worked with Ice.com, I thought that the site had so many links that picking onpage optimization low hanging fruit would maximize their ROI. While that was true, I realize now that once the low hanging fruit had been picked, it would still be necessary to keep building links. (I eventually hired a rockstar SEO for Ice, when I returned to school full time. You can find his blog at SEO Skeptic.)

    Some fun and actionable data points for niche researchers:

    The hardest verticals to compete in (imho) are those whose players have the greatest speed of link building, the deepest relationships, and the most links.

     What 70+ Hours Of Backlink Checking And Research Shows...

    I should also give an honorable mention to “Domain names,” whose ranking leaders GoDaddy and NetSol had 46 million new links since last year!

     What 70+ Hours Of Backlink Checking And Research Shows...

    The ratio of “backlinks per referring domain,” which I’ll call “Sitewide Strength” can show several things:

    • How big the sites are that competitors are getting sitewide links from. EG TechCrunch with 1000s of pages or a brand new blog with 5? Whose blogroll are you on?
    • Competitors buying sitewide links.
    • Competitors having big brands and thus being cited really frequently in the press. EG Google.
    • Competitors with strong relationships; each relationship generates numerous links.

     What 70+ Hours Of Backlink Checking And Research Shows...

     What 70+ Hours Of Backlink Checking And Research Shows...

    Raw Backlink Numbers:

    • Only domain names, at 80% industry average, had a year over year backlink growth rate below 95%. But this needs to be contrasted with the niche’s ridiculously high average year/year growth numbers, which lead all categories. Average year/year growth in this niche was 32 million links… (Not median growth, but average, which includes mega leaders like GoDaddy and NetSol).
    • Godaddy apparently has 600 million – yes, million – links. Though most of those come from parked domains, I would imagine.
    • Network Solutions apparently has 112 million links, most of which I imagine are also of the parked-page-linking-to-the-registrar type.

    This is part 1 of 2. Part 2 can be found on my own search marketing blog. If you liked this post, add my rss feed to your reader to download the spreadsheet and screenshots with all the raw data so you can run your own analysis, and of course get my latest posts.

    Comments are closed.

    23 thoughts on “What 70+ Hours Of Backlink Checking And Research Shows…

    1. Thank you for this. Articles like this help people see the behind-the-scenes complexity a well planned and orchestrated link building campaign requires. The days of Googling “Submit Link” [keyword] ENTER are long gone.

    2. Thanks for the post, however I was perturbed by going to your website and having to enter some kind of password that I don’t have, therefore disabling me from reading the rest of the post. I was definitely well on my way to adding your blog to my RSS but this has definitely eschewed me from that process.

    3. Great article Gab!

      I was very happy to provide some numbers – I’ll just second Eric (hi there!) that the days of simple life are well and truly over, competition is high so planning is essential and planning is very difficult (if possible) without good intelligence.

    4. Eric – I’m flattered to receive a comment from the likes of you! Glad you liked it :)

      @Ross – Sorry about that, it was just a question of missed coordination, as I didn’t want to publish part 2 without part 1 being up, and I wasn’t sure when exactly SEJ would post part 1. I hope you’ll reconsider and add my blog to your reader :).

    5. Hey Gab,

      Excellent post. Could you elaborate a little on the below steps?

      “Category Selection

      1. I began by identifying industries and verticals based on the Yahoo Directory’s main categories – those linked to off the directory’s homepage.

      2. From those, I picked 15 single word ‘keyphrases’ and typed them into Google. I also picked 15 two-three word keyphrases, aiming to find mid-tier keywords (judged by volume and competitive interest).”

      How would this work for doing research on say…the “Widget” industry?

      1. Hey Patrick,

        Thanks for the kind words.

        Category selection basically amounted to me browsing the Yahoo directory for major categories and [arbitrarily] picking some keywords. If memory serves, the keywords I picked were basically category or subcategory names, or else just words that I instinctively associated with the category. I’ll be the first to admit that my process, while documented, was not scientifically rigorous or what not.

        For the widget business, I’d go to the Yahoo directory and see what the main keywords were by reading top level and sublevel category names were, and/or use my instinct and/or read the major competitors’ homepage title tags ;).

    6. Gab,

      I always appreciate when someone takes the time to do the legwork you’ve done. Just reading end result stats and making claims doesn’t cut it for me.

      I just wonder how the whole back-link report thing is going to shake out when Bartzland becomes bingified… Personally I think it’s going to be a sad day in our industry.

      Kind of like when Yangville took over and decimated Overture.

      1. I really hope she lets Y! Site Explorer stay up! Much as I like Majestic, I have a better feel for the numbers and competitiveness they indicate when the numbers are from Yahoo’s.

    7. First any data that is just the links to the home page is not a very good recordset especially since the deep links are by far the most important. It may be that Google doesn’t even count all of the home page links at some point so I have to question the choice of data source. Secondly real competitive analysis has little to do with links it should be more about how they got the links, links are such a small part of it, though I do admit backlink is part of that discovery but it shouldn’t be the main reason for the competitive analysis. Link analysis helps discover the most effective content doing something about that in the end will go further than just chasing a competitors links.

      1. The data isn’t just the links to the homepage, and I never said that. Try out Majestic’s tools and see for yourself.

        Second, I don’t see why Google wouldn’t count all links to a homepage.

        “Secondly real competitive analysis has little to do with links it should be more about how they got the links, links are such a small part of it, though I do admit backlink is part of that discovery but it shouldn’t be the main reason for the competitive analysis”

        I think you’re suggesting that I should have told you how to rank in 30 different niches. With all due respect, I think you can either afford to pay for that information, or spend the time to find it yourself.

    8. Hi Gab,

      Thanks for the informative article. Also read the 2nd part posted on your own blog. Really appreciate research like this. However I think you’re missing a few crucial elements in this case. What you did was calculating the number of backlinks and the number of different domains linking. However what I really missed in this research were: a) relevancy of the backlinks and b) the anchor URLs or the pages the backlinks pointed to.

      In the 2nd part of your article you mention the rankings of BLue Nile for “jewelry” and the results for “baseball”. I believe these are perfect examples of why relevancy of the backlinks for the targeted keywords forms an essential part of any backlink research. Are the backlinks of Blue Nile relevant for the keyword “jewelry”?

      Again great to see you taking up this research, but I think it’s not complete enough to draw major conclusions from it.

      1. Sebastiaan, let me first thank you for the comment.

        Second, if you want relevancy of backlinks on this scale, for 30 niches … you’ve got to do that yourself. I want to help, but if you want to do SEO at an advanced level, you can’t expect to have others do all the work for you.

        Re: Blue Nile – I suspect their lower rankings are due to poor anchor text since their name doesn’t contain the word jewelry.

        As to it not being complete enough to draw major conclusions, I think you’re missing the point about how volume is still a major player.

        1. Hi Gab,

          No need to do research for me. Point I’m trying to make is that volume, in other words the sheer number of links, isn’t the only factor that determines whether you rank in the top results or not. Of course if in a specific market the top ranking site has a million backlinks, it tells you something. But like your own research shows there are examples of sites with less backlinks ranking higher. And that’s where relevancy becomes a big factor.

          So I think the conclusion is: if in a market all top tanked sites have a huge volume of backlinks that is also growing, competition is fierce and it won’t be easy to enter as a new party. But can you beat a site that has more backlinks than you have? Yes you can! (using someone else’s famous phrase here…)

    9. Thank you for a very informative and well researched post. I did log on to your suggested “personalized search disabler” and was not able to get the pug-in. I wonder if anyone else had this problem?

    10. Sebastiaan Elsholz: you are right saying that simple number of backlinks (or referring domains) is not sufficient to rank, relevancy plays important roles and also quality of backlinks.

      It is however right (in my view) to look at typical backlink counts by industry (or some other fairly broad category) – clearly there are substantial differences between them as Gab’s research indicates. Some categories are simply much more competitive than others, common knowledge, but this article puts some hard numbers to put this competition into perspective.

    11. This a very good post!

      This is a good example of Information Overload. I cannot wait for your next article. because this good one makes me see the goodness of the next.

      Thanks for sharing this with us!

    12. Yes great article Gab. May have to delve into Majestic…been spending too much time on the other one. Like the idea of the ‘sidewide strength’ view – these are supposed to be discounted by Google, from my experience this is generally untrue.