Study is about brand loyalty to certain web sites, not the quality of a domain name.
A study by Microsoft over “domain bias” in searches is making the rounds today. Unfortunately I think domainers are misinterpreting what it says. I’m even going to disagree with Michael Berkens on the importance of this one.
The study is titled “Domain Bias in Web Search” and it shows that people are more likely to click on listings associated with certain domain names.
Score one for domainers, right?
Not really.
The authors use the term “domain” to mean web site. The study doesn’t show anything about how a domain like Baseball.com is better than MyCoolBaseball.com.
Instead it just shows that people trust certain web sites, such as Wikipedia or Amazon, and they’re more likely to click on links to those sites.
If I’m looking for information about a vacuum cleaner, I’m more likely to click the trusted link to Amazon.com than somerandomvacuumsite.com.
What is surprising is that users click on results from reputable domains even when more relevant search results are available. Our experiments are geared towards proving that domains can so drastically influence perceived relevance that users will favor some domains, regardless of content. Viewing content on the Internet as products, domains have emerged as brands. And users have developed such fierce brand loyalty that their clicks are tainted by domains.
This study is about brand loyalty to certain web sites (and hence their domains), not to the “quality” of the domain name.
Now, this doesn’t mean I don’t believe in the power of a good domain in search. If I’m searching for a vacuum, my first choice might be Amazon.com. But then I’m much more likely to click on Vacuum.com than SandysVacuumEmporium-Online.com as a second choice.
My point is that this study doesn’t say what we want it to say. Much like a “direct navigation” study from many years ago that showed high conversions on direct navigation (thanks to repeat traffic, not direct search) that domainers attributed to type-in search, this study won’t really help us out.