Your Attribution Model Sucks. And It’s No Big Deal.

Your Attribution Model Sucks. And It’s No Big Deal.

While on my high school newspaper, I learned about the 5 Ws – Who, What, When, Where, and Why.

Organizations want these answers too. They imagine we consultants have a magic attribution wand we wave. Suddenly we know everything about everyone who interacts with them online. But perfect, even really good, attribution is almost impossible to achieve.

Little girl with a magic wand, waving it in front of a computer.
You can’t wave a magic wand to get attribution!

Here’s an example from my own life.

  • I listen to Leah Pica’s podcast on a regular basis. I’m always months behind. In June I listened to a Podcast where she mentioned the DA Hub (a digital analytics conference).
  • I remembered that someone was discussing the DA Hub in Measure Slack a few days earlier. I look to Measure Slack all the time for industry information.
  • In Trello, I made a card called “Look into DA Hub” on my “List o’ Stuff to do when you have a free five minutes.”
  • About a week later, I Googled DA Hub, went to the website, and joined the mailing list to get more information.

So let’s recap. Leah Pica mentioned the DA Hub on a podcast. Two months later, I listened to it. A week later, I Googled the DA Hub, and signed up for their mailing list. So  there was a good four months in between Leah discussing the DA Hub to the point where I signed up for the mailing list.

I know my story is anecdotal, but it demonstrates that folks go about their world in all sorts of different ways.

I expect that the folks who run the DA Hub do a bang-up job with their tracking. They are digital analysts, after all.

Yet chances are they have no idea what my “conversion journey” is. At best, they may know I typed “DA Hub” into Google. But that was the least interesting part of the journey. Leah got me there, with an assist from Measure Slack.

If DA Hub only used  website data, the forces that drive conversion would never be credited. The data alone would conclude she doesn’t have influence. And the data would be dead wrong.

Which leads to an unpopular option: work on attribution, use data, but don’t be fanatical. Don’t spend more time and money tracking than doing what you actually do.  Go ahead, add UTM codes. Develop a level of governance around them. Do your best to track those channels you spend a lot of time and money.

But understand, your data is just a small piece of a big puzzle. Don’t beat yourself up because you have a few unattributed conversions. The bulk of your ad-tech may be downright wrong. Don’t bet the bank on it. Don’t ignore industry leaders, best practices, and word-of-mouth, just because you can’t put UTM codes at the end of them.

Make the best decisions you can with the information you have.