The trouble with Attribution

As a marketer, one of the goals is to create ways to establish accurate Attribution.

In marketing, attribution is the identification of a set of user actions (“events” or “touchpoints”) that contribute in some manner to a desired outcome, and then the assignment of a value to each of these events. Source

During my conversations with marketers (especially those spending a lot of money!), the reason to establish accurate “attribution” is not just to measure to effectiveness of a channel(i.e. which channel gets leads) but also effectiveness of their team. I am not sure how far this is true, but if it is the case, then this is worrying. This is because in my experience, attribution as a metric should be taken with a pinch of salt.

Let’s look at 3 channels where your goal is to get someone to sign up for a trial version of your product. The way someone can do this is fill up a form on your website. The promotions you’re running to get users to this form are:

Channel 1: Google Adwords

Channel 2: Promoted tweets on Twitter

Channel 3: A sponsored post on a 3rd party blog

You’ll be spending $$$ on these channels and the way you’d measure how effective a channel is when someone clicked on a link in that channel (for example). So, if Person Y clicked the link on the tweet and then you’d account for that channel to be responsible to “convert” Y.

However, in the real world, people don’t just see and ad and click on a link. What might happen is let’s say someone attended an event offline (where your tool was discussed) and then saw the sponsored post. They read it and thought it might be something that they were interested in. Then, they saw a promoted tweet (but didn’t do anything here) and clicked the link there. So while the last channel would get the attribution, the first two channels also hold responsibility to convert the user.

One of the challenges I’ve had as a marketer for an open source tool, is that we actively try and NOT implement any kind of tracking that might broach privacy of users. This makes it incredibly hard to measure attribution. But in a strange way, it has also liberating because the team is open to experiment with new and more creative ways of marketing.