The research and statistics thing as it relates to the Internet and traffic and blah blah blah is ridiculously imperfect and highly malleable. As Internet marketers we all just need to keep that in mind as we get inundated with data.

Research is starting to seem as valid as the number of followers is for how influential one is on Twitter.
.-= Frank Reed´s last blog ..Is Aimless Meandering Just a Part of Social Media? =-.

My own “research” would be very much inline with iCrossing’s–which is part of my observation. Site’s that have been SEO’d tend to have a higher percentage of Google traffic.

I always appreciate the cynical eye toward any facts, figures, and research – but I have an even more cynical view of the ComScore data. I think there are biases in any research and the ComScore data may just be biased to skew away from Google data (I’m not saying it does, but there are surely biases there too). I think ComScore, Hitwize, Compete, and the like are “decent” at directional data, but often actual figures are questionable – such as predicting an exact search market share.

It’s also been my experience that Google sends significantly more traffic than its “market share”. I’ve managed plenty of small business sites that followed this trend – and much like this research, all the Fortune 500 sites I’ve had access to their analytics show the same larger share.

What about the sites you’ve managed? Have you seen things more inline with the market share figures that has led you to question this?

Yes, the data included all brand and generic keywords. The share was also calculated for natural search only. This also indicates that searchers who go to enterprise sites may use Google more often, although now I’m curious about brand vs. generic distribution in natural search. Maybe the topic of a future study : )

Rob

I hadn’t even considered the “Enterprise” skewing. Any data on the number of those queries that were “brand” keywords. I would suspect that Google sees a higher share of “IBM” or “Coke” searches, which could be another factor that skewed the share higher for Google.

I put this study together with a colleague at iCrossing, so hopefully I can help shed some light on the data we found. Your observation that the findings are biased is absolutely correct. The Search Engine Land post stated that the share was for *enterprise* level sites that generally fall into the category of a Fortune 1000 level business. It is by no means intended as a proxy for all search share across the web, but we did cover a wide range of verticals in the enterprise space.

When we ran the numbers, we thought there was a compelling story here that search share for large enterprise sites is likely very different from that of the “all sites across the web.” I found the individual ranges very surprising in many cases (especially for sites that derived a substantial amount of traffic from Bing and Yahoo), and many other sites were receiving shares well into the 80%+ range from Google. But when averaged out, what you get is a slightly higher share of search traffic from Google, and lower from Bing and Yahoo.

As far as the aspect of Google bias in optimization, I’ve been on the record many times in speaking, and in my MediaPost column in saying that we optimize for all engines. Maybe this Google skew towards enterprise is more of a reflection of Eric Schmidt’s comment that “the internet is a cesspool, and brands are the answer”, and the data provides some proof that their emphasis on big brands is greater than the other engines, however slight.

Again, hope this helps shed some light on the purpose and findings of the study.

Thanks,
Rob

Pretty much any site I’ve ever looked at the analytics for has a good 80% or more of their search engine traffic from Google. And that’s often without them ever having done any SEO.
.-= Jill Whalen´s last blog ..Advanced Forum Thread of the Week =-.