Mythbusting on Web Analytics
Posted by:
Misty McLaughlin on
April 21, 2008 at
2:41PM EST
Last week, I attended the ever-wonderful Information Architecture Summit, in pursuit of the latest tricks and trends for creating usable, intuitive web presences.
A few insightful sessions on Web traffic analytics have got me thinking about the myths we marketers, webmasters, execs, and development folks hold dear about our data. Whether it's Google Analytics, Visual Sciences, Urchin, Webtrends, Analog, Clicktracks - however you collect your data - these myths keep us from making the most of what our metrics can tell us.
Myth #1. Analytics tell us why our (inscrutable) users do what they do.
Traffic data gives us the "what" - and only part of the "what" - of how visitors use our site. It doesn't give us the "why." Yes, we can use the "what" of visitor behavior to speculate about the "why" of user motivation, but we don't know, for instance, why someone spends a particularly long time in a certain section of the site. Is she fascinated and reading every word - or so bored she walked away completely?
Myth #2. Accuracy is what we're shooting for.
If only. Analytics help us look for trends - not hard-and-fast, statistically significant numbers that can give The Irrefutable Facts. They do a great job of showing us growth or contraction over time, changing visitor behaviors, overall visitor loyalty. But the rules for how each analytics program collects data are constantly changing, which makes for good benchmarking, but, for instance, poor t-tests.
Myth #3. Traffic data is all numbers - hard quantitative stats, my friends.
Well, okay, you're right - in today's world, that's still mostly the case. When people talk about measuring and benchmarking traffic data, they're typically referring to big, impressive numbers. But some of the most interesting developments in the analytics world are actually around qualitative user data, such as that collected through tools called Web Use Recorders. These tools record individual browsing sessions to give you a more qualitative view, from your constituents' perspectives, of your site: where their attention is drawn, what content is overlooked, how they scan your homepage.
So now what?, you ask. If our traffic data doesn't tell us "why," isn't always accurate, and isn't "just the numbers," why do I analyze this data at all.
The answer: Triangulation with other user research methods and data sources. Traffic data is a critical - but singular - pillar of user research. Without alternative techniques like user interviews or surveys, usability tests or focus groups, even the best-laid plans of metricians and analysts can only answer part of the eternal question:
Who are our constituents, what do they want, and why do they do what they do online?
(1) Comments
|
Posted by: Pagealizer on April 24, 2008 1:24AM EST
For landing page analysis try Pagealizer Pagealizer shows how long people stay on your page (actual bounce rate), how far visitors scrolled down the page and where they clicked. Check us out :) |
|
Are you a Convio client?
Get answers to product questions, submit & rate product ideas, join “Birds of a Feather” discussions & more.
|
