Interview with SEO Evangelist Rand Fishkin

by Jon Cooper

There are few people that have helped shape the SEO industry more than Rand Fishkin, the CEO of SEOmoz. Not only does he give out a ton of awesome, actionable tactics & strategies on the SEOmoz blog, but he’s also one of the bright minds behind Open Site Explorer, one of the most popular link building tools out there.

That’s why I’m excited to share with you a few answers he gave me to some questions I had about link building and a few other SEO related subjects. 

1. At last year’s LinkLove conference, you said, “The best link building is when you hit the Publish button”. While this has been a huge success for SEOmoz, a lot of small businesses have struggled with this idea of “attracting” links. What advice would you give them?

Building a content marketing platform – a blog, a web content series, a highly-subscribed-to article resource, etc. is a huge, huge potential win. However, if you can’t invest in content for your own site, another option is to use the power of already-powerful publishing sources. Guest authoring amazing work can get you in front of tens of thousands of readers at popular blogs, news sites, etc. but you’ll have to be willing to do exceptional work. Attracting links naturally is not easy, but it is incredibly worthwhile.


2. In a recent study, Open Site Explorer has been shown to provide as low as 3% of a website’s total number of links. Do you see this as a problem? Explain. If it is, are you making steps to solve it?

Hmm… Yeah, I’m seriously doubting the methodology there. We have access to a lot of link data from sources like Google and Bing Webmaster Tools, and will see as low as 40% (particularly for newer sites or those who link build from far-off, less well-traveled parts of the web), and as high as 90%. I’d be particularly wary of comparisons using Majestic’s historic indices, as these have link data that may not have existed on the web for many, many months, and given the web’s natural rate of churn, this will make “all links” look far, far inflated to an actual count an engine might see.

All that said, we do have plans to launch an index in the next 3-4 weeks that’s ~3X the size of our largest index to date, and if our members are happy with that expanded data (and don’t find these far corners of the web too spammy and thin in content to be useful), we’ll maintain those sizes going forward.


3. Currently Open Site Explorer’s index updates once every month. Will SEOmoz update it more frequently in the future? Why or why not?

We’d love to 🙂 The challenge is with processing. Other link indices, like Majestic, don’t need to run link-graph calculations like MozRank (our version of Google’s PageRank), MozTrust, Domain Authority, Page Authority, etc. Thus, they can continually update a static database. This is a great tactic to keep costs low and to have a huge index that can stay relatively fresh. However, it means you can’t do things like predictive metrics against Google’s rankings using machine-learning-based algos, or mimic other search-engine calculations (like PageRank or TrustRank).

Sometime this year, we hope to reach 3-week or possibly 2-week freshness, but it’s a massive data-crunching and cost challenge. Running processing and serving of Linkscape on Amazon’s EC2 in January alone cost us ~$300,000 and took almost the entire month. Doing it twice as fast would cost more than twice as much given our current code and infrastructure, so we need to work out some of those issues before we can cut down the processing time.


4. Do you have any suspicion as to which link evaluation factor is regarded in Google’s recent algorithm update? If so, why?

I thought Bill Slawski’s post covered some potential items in expert fashion: http://www.seobythesea.com/2012/03/12-google-link-analysis-methods/. Unfortunately, I’ve neither seen nor “felt” any changes in the SERPs that would justify a particular hypothesis yet.


5. What set of skills does a modern day SEO need in order to succeed?

I wrote a post on this last summer: http://www.seomoz.org/blog/the-responsibilities-of-seo-have-been-upgraded. My feeling is that a modern SEO can be a specialist and succeed, but it will likely be in a narrow and/or tactical fashion. That’s not to say specialist tacticians aren’t fantastic sometimes, just that a modern marketing department needs to approach SEO holistically and that means having areas of strength around design, development, content, social, outreach, analytics, CRO and community along with the classics of keyword research, on-page optimization and link building.


6. What’s the greatest challenge you’re currently facing with your new project Inbound.org, and how are you going about solving it?

Right now it’s really participation on the “incoming” page. There aren’t quite enough great submissions coming in yet (though thankfully, people do seem to only be submitting pretty good stuff and very little spam/crap), nor are there enough folks voting up the quality links. That said, it’s been cruising along organically without a lot of input or direction from Dharmesh and I, which is great. I think that means the marketing community was ready for something like this and has embraced/made it their own.


7. In the past 6-12 months, terms like “inbound marketer” and “digital marketer” have started replacing “SEO”. As a result, what does this tell you about the future of our industry, and how will you & SEOmoz adapt to this change?

I’m not sure if I totally agree with that statement. I’ve seen many folks who say things like “I’m responsible for a lot more than just SEO. This title doesn’t fit me fully anymore,” and that’s different than saying “I’m replacing my ‘SEO’ title with ‘director of inbound marketing’ tomorrow.” My opinion is more that as marketers engaging in SEO have evolved to encompass a wider array of channels, their titles are changing. SEO is still SEO – still means what it always did and still has a critical place in the marketing mix for every company on the web.

As far as SEOmoz goes, we rolled out some light social media monitoring tools last December but are actually planning to focus intensely on our Linkscape index – growing the depth, breadth and freshness as I noted above, as well as some new SEO/search-focused features in the software. In the longer-term, you’ll see us re-invest in social, content and broader marketing analytics to help keep up with the needs of our customers.


I just want to give a HUGE thanks to Rand for doing this. Do yourself a favor and follow him on Twitter and check out the SEOmoz blog if you haven’t.


This post was written by...

Jon Cooper – who has written 129 posts on Point Blank SEO.

Jon Cooper is a link builder based out of Gainesville, FL. For more information on him and Point Blank SEO, visit the about page. Follow him on Twitter @PointBlankSEO.

Relax - I send out free emails full of
cutting edge link building tips.
  1. Kane Jamison says:

    Not sure if Rand will be around here answering questions, but:

    “We have access to a lot of link data from sources like Google and Bing Webmaster Tools, and will see as low as 40% (particularly for newer sites or those who link build from far-off, less well-traveled parts of the web)”

    1 – Is this a sample of GWT/BWT data for sites that SEOMoz or partners control and allow access to, or is this coming from a wider study of sites across multiple industries and market shares?

    2 – Also, regarding the “far-off, less well-traveled parts of the web,” I have a subjective hunch that OSE data tends to crawl fairly well for some industres like tech, and may miss things more often for more obscure industries. You could make the argument that that’s OK, since it would be the same bias for industry competitors, but I’m wondering if the OSE team knows of any particular areas that the data is lacking, and what types of things they’ve done to counteract that (such as adjusting seed sites that they crawl from, etc.)

    • Kane Jamison says:

      As a followup, a quick sampling of 5 small biz sites I manage shows a range of 27% to 51% for OSE external links to root domain VS. GWT linking domains. I consider that a pretty good ratio since GWT now shows quite a bit of junk links.

    • Rand Fishkin says:

      #1 – Yeah, this is a sample from our sites and those of partners. We don’t yet have access to a large number of Google Webmaster Tools accounts through oAuth or some such in the app, but that might be a good idea (opt-in only, of course) for future comparisons.

      #2 – I think this is a bias based on the link graph itself (certain industries, like tech, are naturally better-linked-to) rather than a bias inherent in how Linkscape crawls (which is wholly content-agnostic). We do see any URL on which link information is requested that we don’t have, and have been working on plans to discover and crawl these (though the real problem is finding the links that point to them).

      Thanks for the questions and the feedback on comparing OSE vs. GWMT!

  2. I had no idea the cost of how much it cost to run a linkscape update. Holy cow!

  3. Thanks for the insights into OSE, Rand. In terms of an expanded Linkscape Index, I’m SUPER excited. The spammy links are a part of every site’s ranking story, and they help to shine light into SERP trends within the industry – though it’ s easy to misinterpret or misjudge these trends. Even for the more seasoned link researchers.

    With that said, I’m curious what everyone else thinks about having a more “spammier” Linkscape index? Do you think it will make competitive backlink analysis easier or harder for you? As someone who lives in excel and Google docs, it’s hard for me to look past the “the more data to play with the better” side of things. But sometimes with more data, I can definitely spend too much time playing, lol.

  4. My advice on the vote system of inbound.org : at least make the vote button responsive. I don’t know how much time it takes in the US, but in France it takes more than 3 second to make the vote count. Speed up the ajax request or make the vote appear when we click on it instead of waiting for the server response.

    OSE has far less backlinks in its index than other tools, but I believe OSE returns the links that count. This can be a problem in some case, when the competitors are using BH tactics and you can’t see these spammy links unless you pay a premium account on majestic or ahrefs.

Interview with SEO Evangelist Rand Fishkin - Point Blank SEO