Every so often I find an article about web analytics that is a great reminder of something. This article by Sayf Sharif does a great job of describing the problems of web attribution and how direct traffic (which has no attribution) can mask the customer journey.
It gets a bit technical, but it’s definitely worth a read.
Here are some thoughts on this article:
It goes to great length to expose the problems of direct attribution in web analytics and how it can creep in from many directions. The list of ways that traffic can be tagged as “direct” is handy and a good reminder for anyone who spends a lot of time with web metrics.
Calling direct attribution “dark” traffic is a great way to describe it. It’s easy to forget that direct/none is not something specific. It is everything that is unknown. For us, it is cloaked by ambiguity.
I’m glad to see the admonition to tag everything. This is something I’ve been preaching for a long time (download my UTM spreadsheet for help).
Diving direct traffic into “surface, shallow, deep” based on entry page as a way of gaining insight is an interesting idea and something worth playing with. Another way might be based on engagement to get a sense of where the quality is happening.
The ideas around cohorts for understanding long term conversion activity is interesting, but a little off topic.
Overall a really good article. The direct traffic is present in every web analytics setup. Attribution is one of the core challenges in analytics, and this article does a good job of digging into the mystery of when no attribution shows up. ¯\_(ツ)_/¯