Why use underscores in URLs with UTMs?

Question

Back in the day (2005), Matt Cutts said "use dashes, not underscores" in URLs. But that was before search engines got smart, and now it seems it doesn't really matter what you use in canonical URLs.

But what about URLs with UTMs? How can dashes vs. underscores help us understand traffic in Google Analytics?

B2C published a blog post in September 2013 analyzing the different ways that different companies use UTMs in email marketing materials. At the end of the Gilt City email, the author says:

An underscore should be used in “san-francisco” part of utm_campaign. This tells us it is one word and separates it from the other information.

When the author gets to the Zillow email, he mentions:

Separating different pieces of information by a dash (“-”) helps to parse them easily in analytics tools.

I'm modestly savvy with Google Analytics, and I'm having trouble seeing the advantage of using an underscore to link compound phrases such as promoted link, sponsored story, sidebar ad or other multi-word phrases that I might choose to regularly use in my company's UTM schema.

Is there an advantage in Google Analytics to linking these words one way or the other?

Answer 1

You only get 5 standard UTM parameters (without creating any custom ones) and all of them have predefined roles. But sometimes you need to send additional tracking information that does not fit into any of them. e.g you might want to track the date when an email campaign was sent. For this, something like utm_date would be very helpful. Since it's not readily available, you can add this information to other parameters (e.g. utm_content).

If you follow the convention that compound phrases are connected by underscore, and multiple phrases by dashes, you can later parse these parameters using regular expressions in Google Analytics to create more complex reports.

For example, sponsored_story-13_june_2013 can be split into "sponsored_story" and "13_june_2013" later. This will help you create a report for "sponsored_story" irrespective of when the emails were sent.

UTM parameters are special compared to any other URL parameters. Google explicitly ignores them while indexing. They are purely for analytics. No one except the analytics tools (including Kissmetrics, Mixpanel, etc...) uses them.

Having said that, a lot of the above recommendations are just conventions. You can come up with your own conventions if you like, but make sure to use them all the time. Good, well-defined conventions, consistently implemented can give you much cleaner reports and you will avoid a lot of headache later on.

P.S. I am the author of that B2C blog post.