Sunday, January 9, 2011

The Great Analytics Experiment, Part One

In an attempt to better my shop (and in the spirit of science), I'm undertaking a Google Analytics Experiment.

Hypothesis: Listing items based on top keyword results will result in more hits, lower bounce rates, more time on the site, and more pages per visit (or any combination thereof).

Step One: What are my top keyword results?

As it stands, here are the top five keyword searches (ranging from November 30th until now) that people were looking for when they found my shop:

  • pi earrings
  • golden ratio necklace
  • math jewelry
  • pi bracelet
  • golden ratio jewelry


People are clearly looking for two things here: pi jewelry and golden ratio jewelry. So, I'm going to shift the focus in my math jewelry just on the golden ratio and pi for the next round of listings. (I guess people have very little love for the square root of two, or the Feigenbaum constants. What a pity.)

Anything related to science or chemistry is curiously absent from that keyword list, but I think it's far more likely because I simply have more math items in my shop than I do science items.

Step Two: What are my stats like now?

There's lots of web traffic stats to look at. I'm going to focus on four areas, though I may whittle that down as I go along:

  • Total Visits
  • Time on Site
  • Bounce Rate
  • Pages per Visit


Total Visits is pretty obvious. So is Time on Site. Bounce Rate is how often people leave a page for somewhere else without clicking any further. It's calculated by dividing visits that lasted for only one page by the total number of visits. So, for example, if 100 people found my Etsy storefront and 60 of them continued to browse Etsy without clicking on any of my items or my profile, my store would have a bounce rate of 60%. Pages per Visit is also pretty obvious.

Why did I pick these stats?

They're all important, of course! You can't have any sales without any visits, period. So increasing Total Visits—whether repeat or new visitors—will increase your chance for a sale. And the longer those visitors are lingering in your shop, the more likely they are to buy something (and the more likely it is that they like what you do, and will consider buying in the future even if they don't on this visit). So you want the Time on Site to be pretty high, too. Likewise, if they're leaving after only one page, they probably don't like what they see all too much—so you want a low Bounce Rate as well as more Pages per Visit. My (shamefully low) stats from the same timespan, courtesy of Google Analytics, are as follows:

  • 419 Visits
  • 2.09 Pages/View
  • 65.16% Bounce Rate
  • 1:31 Average Time On Site


Step Three: Procedure

For the next month, I will focus on listing golden ratio and pi jewelry in my math section. Any new pieces I make will only be pi- or golden ratio-based. At the end of the month, I will see how the new data (from the date of the first listing until forty days after) compares.

No comments:

Post a Comment