At the SMX East 2011 "The Great Correlation vs Causation" panel, industry experts speak about ways marketers can measure the SEO impact of various factors and why creating a human-friendly site experience counts.

Ever since data about the correlation of social sharing factors and top search rankings was shared at SMX Advanced, marketers have been debating the extent to which social data impacts their search rankings. At SMX East New York 2011, insiders are talking about the great “correlation versus causation debate,” and businesses might benefit from their testing tips and some overall best SEO practices.

To start, it’s important to clarify what causation and correlation mean, says SitesWithoutWalls.com’s Kristine Schachinger. In causal relationship, you can directly measure that A causes B, but in correlative relationships you can only measure the “strength of the effect of A on B,” according to Schachinger. The “confidence interval” is the best way to determine whether one SEO factor has a strong correlation with ranking.

Micah Fisher-Kirshner, senior SEO manager of Become.com, suggests that marketers looking to measure correlations should start using a scatter plot – there are going to be a lot of variables, and starting with scatter plots versus linear graphs is the way to better ensure all bases are covered in terms of what impacts SEO. Marketers should test their sites when they see differences in ranking, and boost their confidence factors by considering outliers and data logistics brought up by Schachinger, Fisher-Kirshner and Mitul Gandhi (founder of seoClarity), such as:

  • The sample size (how many pages studied?)
  • Multiple online campaigns that can be at play
  • Offline campaigns that might be influencing online engagement
  • Seasonal search trends
  • Weekend versus weekday search
  • World events impacting search activity
  • Analysis metrics (are you selecting the right variables?)
  • TEST TEST TEST – length of time factors studied is key

When marketers see a correlation, Fisher-Kirshner suggests they act on it – a correlation is a starting point, not an end, for marketers that really want to boost their SEO. He says once marketers have data, they can look at building more linear regressions – but only when there’s enough data to start making statistically significant theories and when the quality of the data enough to move forward with action based on correlation.

Eric Enge, president of Stone Temple Consulting, and Tony Wright, of WrightIMC, seem to take a different approach, advising that marketers be wary of “shiny things” in SEO. Enge is wary of acting on correlations, posing the following questions:

  • Does correlation cause causation?
  • Does cause lead to correlation?
  • Do correlations cause traffic? (Which may increase propensity to convert…)
  • Aren’t the signals just going to change anyway?

“Correlation versus causation don’t matter,” he says. Enge advises marketers to focus on creating good websites, and Tony Wright points out that you shouldn’t follow every trend that you read about. This harkens back to an SMX Advanced session on building sustainable SEO strategies as a Google Survival tip.

But session moderator (and Search Engine Land’s editor-in-chief) Danny Sullivan encourages marketers not to be blind to SEO correlation opportunities. He points to his experience of seeing Ford in the top 10 search results simply because he is friends with the company on Google+ as an example of an easy correlation businesses can jump on (when Google+ offers branded business pages, that is…).

Overall, the panel seems to agree that testing is necessary to determine ways to get an SEO edge, but they also agree that sustainable SEO best practices are a necessity – though nuances of these practices related to targeting and industry will be different for different businesses. Nonetheless, keeping the user in mind is a must.

Sullivan talks about how focusing on the user is key. He points out that Bing, Google and all major search engines are perfecting their ability to improve the human experience. This is something Brafton has touched on before, reporting that Matt Cutts has said Google, itself, gives SEO advice “to make the web a better place.”

Moral of the story: Test and act as “correlating factors” can still likely bring value beyond SEO to your site, but make sure you have the SEO basics down.

Katherine Griwert is Brafton's Marketing Director. She's practiced content marketing, SEO and social marketing for over five years, and her enthusiasm for new media has even deeper roots. Katherine holds a degree in American Studies from Boston College, and her writing is featured in a number of web publications.