Who cares if there's no such thing as visitors?
September 2019
Yesterday I went on at length about the limitations of tracking visitors to your website.
I strongly urged you to use air quotes around the word “visitors” and even suggested that you call them “device-browser-cookie-lifespans” instead.
So, technology’s imperfect. What’s the big deal?
In order to decide whether we care or not, let’s look at the best and worst case scenarios for visitor tracking.
Best case scenario
Everyone who converts on your site does so in a single visit. (And the poor souls who don’t convert only visit once, too.)
In this case, visitors == sessions, and you don’t have to worry about cookie deletion, Incognito mode, or multiple device usage.
Your metrics are spotless, your experiment results are reliable, and all is right with the universe.
This might be you if: Your reported “returning visitor” count is zero.
Worst case scenario
Everyone who visits your site does so on multiple devices. In fact, they all decide whether to convert while on mobile, but they actually do so on desktop.
In this case, your desktop experience has minimal impact on your conversion rate, even though that’s where the conversions are happening.
Imagine you’re running an A/B/C test on the homepage, for desktop and mobile visitors. And imagine that experience C is far superior to A and B.
Mobile visitors who see C will decide to buy in overwhelmingly higher numbers. Yay! 💰
Except … they’ll all return on desktop to complete their purchases.
And when they do so, they’ll be randomly assigned to A, B, or C again. Because they’re considered new visitors on this new device.
So your results will be flat - no winning experience detected. You’ll never know how amazingly great that C variation was.
This might be you if: Your mobile conversion rate is 0%.
Okay, so I’m sure your returning visitor count isn’t zero, and your mobile conversion rate isn’t 0%.
These ridiculous scenarios don’t really apply to anyone.
But if we think of them as two ends of a spectrum, you can probably place yourself closer to one end or another.
And if you’re closer to the Worst Case Scenario, there are some steps you can take to mitigate the confusion that these device-browser-cookie-lifespans introduce to your analytics and experimentation.
We’ll look at that next. Have a lovely weekend!