Analytics news and expert advice every Thursday.
For 2011, Resolve To Stop Being Average & Get Real!
Isn’t this the time of year we always get a little introspective? What worked this year and what didn’t? I suppose New Year’s is really the time for that, but go with me here. Let’s get introspective about one of the things that drives us crazy all year: reporting.
Today, reporting is largely considered a way of conveying the facts and statistics of our web site, search efforts or other endeavors, but the truth is that none of it really means anything, and at its worst, doesn’t represent “facts” at all. Here are the types of things we talk about in reporting:
- How many people came to our web site
- How many visits those people had
- How many pageviews those people had, if they didn’t bounce
- What percentage of the time that traffic returned some revenue or other valuable event
And it’s not absolutes that matter, right? We rarely talk about the total number of pageviews, because getting 12 million pageviews isn’t all that informative. We want to know how many pageviews each visitor had on average or how many they had in the average visit, so we can see whether the site really drove a lot of interest or not. Same thing with conversions. We want to know the relative measure of performance for our traffic: does search drive “productive” traffic? Are these things trending upward or downward? You know what? It doesn’t freaking matter. Because averages are lies.
There is no such thing as an average. No average clickthrough rate. No average conversion rate, no average bounce rate, page views per visit, percent new visits, any of that. Why? Because there is no such thing as an average visit or visitor. Nobody is capable of looking at 5.82 pages in their session. Nobody in the world is 12% interested in digital cameras and 6.9% interested in washing machines in the same visit. Nobody converts 2.9% of the time. They either do these things or they don’t.
And nobody knows this better than you lovely, attractive, and highly intelligent search people, because you understand with certainty that there is no such thing as an average visit, otherwise there would only be one keyword driving traffic to your site (I’ll let you use your imagination about what that keyword might be!) But we know that thousands if not hundreds of thousands of keywords drive traffic, and these different keywords represent myriad needs, interests, or intents. And a lot of these keywords don’t have a prayer in the universe of creating a real conversion, but they will absolutely provide value to your company and your visitor (with the possible hopes of a conversion down the road) if your marketing and your site are doing their job.
What I’d urge us towards in 2011 is a new way of reporting (and acting on that reporting) that revolves around the key concepts of information architecture. There are two ways you can break down and understand every single visit and visitor on your site that allow you to determine the success of your efforts in search and on the site. Let’s use those.
A Quick Primer On Information Architecture
The most basic and fundamental components of IA are users, context and content. Here are some descriptions that will come in handy for the purposes of this article, but there is a lot more to learn about IA if you’re interested, and I’d imagine that some IAs may debate some of the finer points of these definitions. But since I have the keyboard at this particular moment, here we go.
Users are just people. People with histories, preferences, issues with their in-laws, and mean-ass pets that they swear are well-behaved. I don’t mean uniques or visitors; I literally mean real, actual people, like the ones you see in the grocery store.
Context, then, is what their current situation, need or pursuit is all about. “I need a new dishwasher,” is a context. “My pet won’t stop pooping in the kitchen,” is a context. “My digital camera’s pictures suck compared to Fred’s” is a context. Context is the reason users choose the keywords they choose, attempt to navigate where they navigate, and view what they view. But don’t confuse context for users, or people, because different people with the same context will behave differently. A self-proclaimed smarty-pants will shop for a digital camera very differently than a friendly novice; they will use different types of information to narrow their choices, learn more, compare prices and ultimately convert. So let’s just say context describes the current need. And a “visit” or “session” is literally the intersection of a user and a context.
Content is what you deliver. Your site is content. It’s the good that your site provides to those users with their contexts, and depending on the type of content you have (and the organization of that content), you will have visits that sit on the spectrum between satisfying or not satisfying the visit. SEOs ride companies pretty hard about creating better content around various topics, and they should. Because for the vast array of information-gathering searches/visits/intents, most sites are just trying to drive a conversion rather than really trying to help this particular context.
And The Point Is?
But wait, isn’t all of this just a fancy way of saying, “Use segments?” Well, yes. But no, too. Because we are not just making segments willy-nilly. We are picking specific types of segments related to the two factors in IA that help us understand how productive our visits are. We want to understand our users (your org may have defined personas) and the contexts (which often relate directly to funnel position, or how close to conversion they are), and report on those breakdowns. And what we are reporting on is whether our content is working for these segments.
As an example, let’s say that you do sell digital cameras on your site. How can you determine a sophisticated user with a high funnel position? How about people using Linux, or people running Chrome on a Mac… that sounds like a techie to me. And how about if they search for “digital camera reviews”? There’s your techie high-funnel visit. How did your site do? What about your nontechnical (AOL connection using IE 7) visits with the same search?
Obviously we are talking about some pretty small groups of people here, and no, I don’t advocate making a 4,000 page report every week. But start with some users and contexts that are “no duh” segments and work your way backwards to a realistic subset of personas and funnel positions, and ask yourself if you have the content to satisfy their visit. You’ll also see that measuring everything against conversion rate is totally stupid. Obviously, many of your segments have no prayer of converting. But what you will begin to see is the current: the force that moves people down through the funnel in successive visits. But you’re finally measuring and reporting on something that isn’t the “average” person; you’re talking about something that really tells you if what you are doing is working for real people with real needs that your business can satisfy. You goal shouldn’t be to outline every segment but to drive the point home that breaking things down into pieces means a whole lot more than keeping it “high level.”
There is no getting around the fact that this is going to take a ton of time, especially when you first get started. But in my experience, this reporting paints a clear picture of where the site is strong and where it is weak (and also where your paid and natural search efforts are weak/strong), which changes the focus of the conversation from radical guesses about how to improve pageviews per session for the “average” visitor toward conversations about how to more appropriately work with specific types of people and their specific needs.
If you want even more to think about for 2011, take a look at my predictions (and hopes) for web analytics in 2011, and weigh in on those, too. We’re in for some really exciting changes in 2011.
How close to this do you think you can get next year?
Some opinions expressed in this article may be those of a guest author and not necessarily Search Engine Land. Staff authors are listed here.