Simple logic for web analytics

Posted by Anh Tu Phuc Hoang on July 29th, 2014 at 8:47 am

When I started out years ago with an internship at General Electric I worked in the Project Management office performing statistical analysis (along with development duties). Fresh off a degree in Economics I was going to bring a lot of fire-power and wow everyone with my prowess of Excel's advanced statistical functions.  I created some pretty impressive charts and graphs and I think I may have touched upon what it was my boss wanted to find out only for his persistence in hammering it home. And that is how many analysts begin their career, overshooting the mark with brilliant charts leaving people to ask "what was the question?".

Having recently completed Google’s Analytics Fundamentals and the eCommerce: From Data to Decisions modules I realized what experienced project managers and analysts know. No matter what you believe to be the more interesting challenge sometimes the numbers point to an inefficiency that is given the lowest priority because of the ease of fixing it.

And how does that work, for example in fixing a seemingly ineffectual obstacle that will improve everything prior to it. A case is in looking at the high bounce rate on a Thank You page which comes after the payment processing at the end of an ecommerce transaction. You would think once they have your payment who cares right?

Websites that have tackled their Thank You page however have seen dramatic improvements in ecommerce conversions. And you can Google tons of anecdotes about it, its a phenomenon. The relationship between improved future conversions and lowering the bounce rate of the final Thank You page is a melting pot of good vibes that might be called leaving a good impression. If people still linger on your website after they have paid for their purchase you can imagine what an incredible impression it is making.

That is the sort of simple logic that you learn after taking Google’s Analytics Fundamentals and the eCommerce: From Data to Decisions modules with a bit of life experience. They say that is statistics function, which is not psychology which is really what we try to do when we create stories beyond the coverage of the hard numbers.

Of course many major internet retailers have tackled this issue today. And what you experience on Thank You pages after you have paid feels like someone ushering you from the cash register back into the store in a no-pressure sort of way (by way of written copy and links).  If your Thank You page is seeing a high bounce rate it feels more like a brick wall (“wham bam thank you ma’am, next in line please”). That is the cheap features touted about internet ecommerce and also the simplicity of practicing analytics well. And in improving a Thank You page it is very much like polishing a diamond when you have already invested so much on the other parts of your website.

Leave a comment