Analyze This, That, and the Other

Recently I was appointed chair and host of Search Engine Strategies (SES) London, which, of course, I’m looking forward to immensely. But that’s not the only chair I’m honored to have thrust under me. I’m also proud to have been appointed chair of the Search Analytics track at the upcoming eMetrics Marketing Optimization Summit in Washington D.C. (Any more chairs and I’ll have a full dining set!)

Together with all the other great stuff being ushered in by Google and universal search, search analytics will make not just search more interesting but analytics, too. At this time in search, analytics tends to be tied mainly around PPC (define) campaigns and ROI (define). I must say, as a quick aside, many companies I’ve dealt with seem to take only the most cursory glance at their PPC campaign stats. Then they simply continue to hurl money at search, when they should be paying just a little more attention to this guy.

Back in the day, marketers viewed SEO (define) as little more than smoke and mirrors. We got our analytics out of a log file with a text editor (can anyone else remember just how dull it was trying to read log file referrer data in Notepad?). People still talked about hits back then, and nobody was all that interested in a bunch of text or even a few pie charts they couldn’t make sense of anyway.

It’s been interesting to watch the industry evolve from a brigade of rank checkers and log analyzers into bona fide marketers and marketing analysts. Pre- and post-click, all of a sudden we make perfect partners. Of course, in SEO it’s mainly search engine referrer data people seem to be interested in. On the paid side, it’s mainly tracking tagged pages. But in the true sense of getting the big picture about our clients’ online performance, we must look at the entire marketing optimization process.

I’ve written before that I believe search doesn’t work well in a silo. It must be part of a totally integrated marketing effort. This way, we can get a measure of how it really performs by comparing it across the entire mix. Not only that, other elements such as e-mail marketing and affiliate marketing can frequently add a considerable boost to search marketing, and vice versa.

Web analytics tools’ level of sophistication has reached near mind-boggling proportions. Ask an analyst with a powerful set of tools what he can tell you about your Web site. He’ll most likely respond with: “Everything!”

And there you go. You’re back peering into a series of graphs and charts with a blank stare on your face, trying to figure out exactly what you’re really looking at.

Web Analytics: An Hour a Day” is an extremely well-written book that can help you figure out what you’re looking at. It’s by Avinash Kaushik, whose blog Occam’s Razor is a must-subscribe for analytics junkies. The book gets off to a flying start by declaring traditional Web analytics is dead (while admitting the announcement is two years too late), then explaining with a huge depth of clarity exactly what Web analytics should be.

Having started as data sourced from log files, Web analytics primarily contained technical information, not business information. Kaushik quickly dismisses a lot of the information early analytics focused on, such as page views, exit pages, and distracting metrics such as screen resolution. “We are in the midst of a metamorphosis for our industry; Web analytics is not what it used to be,” he says.

What should Web analytics be about? Key performance indicators (KPIs) have been the cornerstone of traditional Web analytics. Now, says Kaushik, it’s about key insights analysis (KIA). KIA examples include:

  • Click density analysis: Segmenting different types of users across your site
  • Visitor primary purpose: Measuring why people come to your Web site, not just how
  • Task completion rates: Migrating away from clickstream data toward successful task completion
  • Segmented visitor trends: Segmenting customers and behavior for a richer understanding of their interaction with your Web site
  • Multichannel impact analysis: Measuring the impact of other channels, such as TV, radio, and print, on traffic

According to Kaushik, if you wake up any Web analyst in the middle of the night and ask what he measures, his first two words will be “conversion rate.” But conversion rate is an unworthy obsession, he says. This was a very interesting section of the book indeed, considering I’m a conversion freak myself. If you solve for conversion rate, he goes on to say, are you solving for all your traffic? Are you improving the Web site experience for all your customers? The most likely answer to these questions is a big whopping “no.”

Fortunately, he has a logical argument for making these statements and certainly helps throw a new light on, what is for many in this industry, an intriguing subject. With a forward by my buddy Jim Sterne, this really is an enlightening how-to book.

The perfect companion book is written by my ClickZ colleagues, Jason Burby and Shane Atchison. See Rebecca Lieb’s post for an overview.

Mike is off this week. Today’s column ran earlier on ClickZ.

Join us for SES Search Engine Marketing Training Workshops on May 6, 2008, at Crowne Plaza Denver in Colorado.

Want more search information? ClickZ SEM Archives contain all our search columns, organized by topic.

Related reading

Brand Top Level Domains