Phew – After 7 long months slogging away, we will finally officially release the long awaited white paper “Chasing the Long Tail of Natural Search [1]” next week Monday (Aug 7th) at SES San Jose and the Etail Philadelphia show.
One is always a little cautious about postulating grand theories into the wide world. But after studying over 1 million unique unbranded keywords across 25 major retailer search programs, we couldn’t resist – referring to the concept we outline as “Page Yield Theory.” This is an underpinning notion that the “long tail” of unbranded search keyword traffic is inextricably linked to the website’s number of uniquely indexable site pages. To those of us who subscribe to the “every-page-should-sing-its-own-song” philosophy, that seems like an obvious statement.
Yet the challenge behind it, and the impetus for the research, arose from the fact that many (unoptimized) well-branded multichannel retailers have 10’s/100’s of thousands of unique and indexed website pages. However most of their natural search traffic (usually over 90%) comes from searches related to their own company name. How could such strong brands and massive websites produce such little traffic for generic terms, terms other than the company name?
Big deal you say — it’s all relative, Brian. Big brands are supposed to have lots of brand searches. Yeah, I know. But here’s my riddle – why do the vast majority (usually over 95%) of those thousands of pages also yield no natural search traffic? What’s the connection?
What we find is that big brand retail sites happen to be laden with freeloader pages: sales people on the payroll, not earning their keep (worse – not willing to make a cold call. Even more worse – not given a phone or a list to make the cold call.)
So “Page Yield Theory” is our humble attempt to frame the correlation of these two phenomenon – suggesting that a small # of unbranded term traffic is a predictable indicator that a large website also has a small # of unique pages yielding search traffic. And vice versa – large amounts of unbranded traffic is caused by an increasing number of unique pages yielding search traffic. Some readers may have no idea how many pages are yielding traffic. No worries – based on the data, we provide a matrix that models these connections, based on the known mix of brand and non-brand natural traffic experienced in an average month.
So what of it? I think this “theory” raises important questions, and maybe a few creative ideas, for managers of large dynamic websites, around how to optimize for the long tail of unbranded keyword traffic. The scale involved practically demands that traditional SEO approaches evolve overnight into more of a “yield management” discipline featuring KPIs, robust reporting platforms, and testable strategies.
I can hear some screaming – but wait! Science? Reporting? Testing? Brian, I thought SEO was more art than science?
Ok, you’ll just have to read the story.
Brian