Over the last year, the online marketing community has been abuzz with talk about using negative press to boost one's rankings in search engines like Google. It all started with a New York Times story about a man whose horrible customer service skills and illegal actions had supposedly boosted his rankings in the search engines. The general proposition was that links are good for search engine rankings, and making people extremely angry is one way to get them to link to your site.

The issue again heated up this April when Kris Roadruck, an SEO well known for his questionable methods and head-on confrontationalism, created a post picking a fight with the majority of the SEO community, claiming that SEO methods that Google approved of were worthless. The post prompted fiery responses from respected industry experts, including this one by Rand Fishkin. Roadruck later claimed that he had picked a fight on purpose, not really believing in his position, simply to attract links. For those who don't know, links from external pages are a major element in improving one's search engine rankings. Kris claimed that his feel-good piece had attracted few links and limited traffic, and thus that picking a fight might be a good SEO tactic.

Is being a bully a viable marketing tactic? I was not convinced.

How the Times Got it Wrong

Let's look at the New York Times piece first. Did this sun glasses shop really gain rankings by attracting negative reviews? If business was booming, and the site was ranking so high, one must wonder where the site owner found the time to bully, harass, and threaten customers who wanted to return items.

Furthermore, looking at the backlink portfolio, we see that the site actually had very few links before the NYT story broke. In fact, most of the links were obtained through third-party discussion of the Times story, and the site has a mere 27 linking root domains. For a supposedly established e-commerce site, that's actually quite pathetic. I run a little six-month-old site in my free time that has more links, higher authority, and probably more traffic.

What about rankings, though? Didn't the NYT story say that the offensive site was ranking first in Google? Yes, and it probably does rank highly for some terms, but probably only in New York. Whenever a user conducts a search in Google, Google will often give a bonus to users who live near the service provider. If she had turned off location tracking on Google, lived somewhere else, or manually changed her location, the site probably never would have appeared.

How Kris Roadruck Got it Wrong

Having interacted with Kris on a few occasions, I can say that I genuinely like him. I also believe that both the initial post and the test were misguided. Kris claimed that he had posted another "lovey" piece about friendship, and it had received very little attention. He used this as a sort of control group, comparing it to his controversial piece. Let's just assume, for the sake of this discussion, that he expended equal effort in promoting both pieces to relevant readers. (Spoiler: he did not).

It doesn't take an SEO Oracle to predict which piece will get more traffic. First, Kris is an expert on search engine optimization, and not on psychology or friendship - the posts he creates in his field of expertise have a running chance already, because other SEOs watch for his SEO-related posts. Second, the test headline was interesting, and the "control" piece was downright dull. One must wonder how SEOmoz's interesting, yet positive pieces fare comparatively in terms of attracting links.

Time to Prove It

I decided to write a piece that would be controversial, and measure the "efficacy of negative-emotion link bait" as opposed to positive link bait. I decided on and created a piece that could be informational to some, while striking a very strong positive or negative emotion for others.

Posting on a third-party blog run by a friend, we see that the piece worked amazingly well. It's hard to see with the huge spike distorting the data, but the blog owner reports that a return visitors have increased average daily traffic by about 20%. With thousands of visitors coming almost entirely from referring sites, we were able to classify 95% of referring domains in one of three categories:

• Site owners, social media sub-groups, and readers who identified strongly with the piece

• Those who disagreed strongly with the post

• General information, and non-targeted social media sites

Analytics figures were quite telling. Of the 95% of referring domains that were easily classified, about 60% of our referrals came from our "general" population and "general" social media referrals. This is not surprising, as there are always fewer people involved with sub-topics than with general topics. Their average time on the site, however, was at about 1:30. That's actually a pretty good average from a general channels.

Our "agreeing" segment represented about 25% of the traffic. These users spent an average of 3:50 on the site! They were also about 80% more likely than the average user to re-share the content in some way, based on the data we were able to compile.

Our "combative" segment sent the remaining 15% of traffic. Clearly they were not interested in reading opposing viewpoints: average time on site was about 30 seconds. They were, however, about 10% more likely than the general population to re-share the content in some way. It was not what we would call the good kind of sharing.

My take away from the data is as follows: controversial topics are good, especially for those who agree with your position. On this point, I'll have to disagree with Kris and the Times in prematurely concluding that disagreement is good. Controversy is good, but lasting benefit comes from giving the customer and the reader what they want.

The number of backlinks received from the post was stellar. Judging by the sheer number of referring domains directly to the post, we estimate that we received more links from this one post than the site featured in the New York Times has total. Conclusion: appealing to what people are passionate about works. There are, of course, plenty of reasons to seek positive press.

Return Customers and Reputation

Branding matters. If a legitimate, unique e-commerce site had been positively featured in the New York Times, they would have seen a spike of interest. They also would have made several sales from interested customers, and picked up a healthy amount of new customers. I also suspect that they would have picked up more links. Brand loyalty is a commodity: Starbucks places the value of their brand at over $300 million. We have known for a very long time that branding matters, and that angering customers will hurt return business and dampen word-of-mouth advertising. Whether online or offline, treating customers well just makes sense.

The Machines are Learning

As Google gets grows in complexity, many SEOs, myself included, feel that the search engine is getting better and better at determining the context of a link. If, for example, your link is on a site dedicated to exposing scammers, and the language on the page is purely negative, Google appears to give this link less weight in ranking your site. As Google gets better at determining the intent of a link, negative PR will decline even further in its effectiveness.

In the world of online marking, I suggest that the only worthwhile press coverage is positive press coverage.