Some (Practical) eCommerce Google Analytics Tips
By Tim Wilson on in Analysis, Web Analytics with 3 Comments
A short, partially self-promotional post — two links and one, “Hey…look out for…” note about sampling.
Post No. 1: Feras Alhlou’s 3 Key GA Reports
Feras Alhlou of E-Nor recently wrote an article for Practical eCommerce that describes three Google Analytics reports with which he recommends eCommerce site owners become familiar. The third one in his list — Funnel Segments — is particularly intriguing (breaking down your funnels by Medium).
Post No. 2: (Log Rolling) 5 Custom Events for eCommerce Sites
I also recently published a Practical eCommerce article with some handy (I claim) tips for eCommerce site owners running Google Analytics that describes five of my favorite custom events for eCommerce sites.
“Hey…look out for…sampling (with conversion rates)”
Sampling in Google Analytics is one of those weird things that people either totally freak out about (especially people who currently or previously worked for the green-themed-vendor-that-has-been-red-for-a-few-years-now) or totally poo-poo as not a big deal at all. Once Google Analytics Premium came out, Google actually started talking about sampling more…because its impact diminishes with Premium.
I actually fell in the “poo-poo” camp for years. The fact was, every time I dug into a metric in a sampled report — when I jumped through hoops to get unsampled data — the result was similar enough for the difference to be immaterial. I patted myself on the back for being a sharp enough analyst to know that an appropriately chosen sample of data can provide a pretty accurate estimate of the total population.
And that’s true.
But, if you start segmenting your traffic and have segments that represent a relatively small percentage of your site’s overall traffic, and if you combine that with a metric like Ecommerce conversion rate (which is a fraction that relies on two metrics: Visits and Transactions), things can start to get pretty wonky. Ryan at Blast Analytics wrote a post that I found really helpful when I was digging into this on behalf of a client a couple of months back.
Obviously, if you’re running the free Google Analytics and you never see the yellow “your data is sampled” box, then this isn’t an issue. Even if you do see the box, you may be able to slide the sampling slider all the way to the right and get unsampled data. If that doesn’t work, you may want to pull your data using shorter timeframes to remove sampling (which throws Unique Visitors out the window as a metric you can use, of course).
Be aware of sampling! It can take a nice hunk of meat out of your tush if you blithely disregard it.