Performance Marketing Tips

Let's Get Down to Business

Posts Tagged ‘engagement

Interview: Evan LaPointe of Atlanta Analytics

leave a comment »

atlanta analyticsIn the first of a recurring series, here’s a miniview (mini interview) with Evan LaPointe of Atlanta Analytics.  I posed 3 rapid fire questions to Evan to get his take on analytics today.  Be sure to check out Evan’s blog www.atlantaanalytics.com.

PMT:  Since you’ve been “doing stuff” online for 15 years (about as long as I have), what is the most profound change that you’ve seen in the use of data in the digital space?
EL:  I think the most profound change I’ve seen is people FINALLY starting to reject data without context.  Just this week, I’ve been working on a client obsessed with exit rates.  But visitors have to exit somewhere.  Without the context of whether the page provides an appropriate circumstance for exit rates or not, they tell you nothing – and sometimes, dare I say, certain pages with insanely exit rates are GOOD!.  I’m glad to see the days of metrics without context slowly disappearing.

PMT:  What is your favorite KPI and why?  (please elaborate or this will be a really short interview)
EL:  Favorite metric? I hate to do this, but it’s got to be bounce rate.  Why? Because it is the ultimate “content rejection” metric.  When I use bounce rates in the context of inbound traffic sources, and especially keywords driving traffic to this page, it’s usually immediately actionable.  If we’re not offering clear calls to action based on what the user has already said they’re looking for, we’re running the wrong way around the track from the very second the gun is fired.  Making small changes to align the page with what visitors expect can have enormous revenue impacts.

PMT:  What do you think is the hardest thing to measure and analyze that could provide the greatest amount of value, especially if it were easy to track? (I’m fishing around the ‘engagement’ watering hole on this one so please try to come up with something new and totally mind blowing)
EL:  What is the hardest thing to measure that would have the greatest value? Well, you’re clearly shifting material that should be on my blog to your own, but I’m okay with that 🙂

I’d say that answering the “why” questions would provide tremendous value.  If we could have less-intrusive survey [qualitative] data to layer against web analytics [quantitative] data (like 4Q survey, but where we don’t just get polar love/hate participation), not only would it mean great data, but it would more importantly mean fewer arguments in the conference rooms which HAVE to account for millions upon millions of dollars of lost productivity every year before we even consider the performance impact on the site.  This industry is rife with terrible opinions and otherwise very talented executives who are clueless about online behavior, paired with a huge absence of data to help the enlightened to make their case definite.

How we get that data, I haven’t a clue.  Hopefully, everyone will begin trusting analysts’ assessments of why things happen and these arguments will be solved through multivariate testing rather than hours of blabber.  A few sites do surveys well, but participation is always going to be an issue, if I’m being realistic.

If we’re looking at measuring “engagement,” I’d just quit.  The only way to get something simple is to degrade the insight of each of the points that make up engagement.  A great example is Peterson’s effort to create this crazy ass engagement metric.  The formula looked like we were trying to predict the trajectory of the lunar lander, not help a web site.  The truth is that this metric is ultimately useless because we lose sight of the details that make it up – and details are what we’ll act on. Creating a mushed-up metric takes us backwards, in my opinion.  It’s like having a number to assess the performance of your car’s parts.  If that number drops, we are no closer to diagnosing and fixing the problem. We just know that something, somewhere is wrong – which is less helpful than it sounds.

To learn more about analytics pick up the Kindle edition of Web Analytics: An Hour a Day from Amazon.com!

Written by Greg M

July 26, 2009 at 11:51 am