Ever since John Wannamaker’s immortal quote; ‘“I know that half my advertising isn’t working, I just don’t know which half,” advertisers have been on a mission to find out what works. Over the last century and a half, the single biggest obstacle to accurately answering this question has been attribution. Many variables contribute to a success, but how do I know which ones contributed the most, and which contributed little to nothing to the result? As marketers, we like simple solutions. Sound bytes we can easily recall and recite back weeks later when asked to justify tough decisions. “Its not rocket science” is the standard complaint when anyone makes things too complex to fit in a 144 character tweet. Well, sorry folks, but THIS IS ROCKET SCIENCE. It’s a bitch of a problem that’s kept the descendants of Mr. Wannamaker awake for the last 150 years, so take a deep breath before reading further.
If you have started researching marketing attribution solutions, you’ve probably come across several articles explaining why a last click approach is wrong and why the author’s solution is better. These kinds of articles are a helpful primer on why marketing attribution modeling is important, but if you’re considering an investment of $200,000+, you may want to read something more in-depth than a sales pitch dressed up with a short discussion about math. So what are the options and relevant questions? To understand the answer, it is helpful to understand that the core functions of an attribution solution are data management and mathematical modeling. This can be accomplished in a platform (Adometry, VisualIQ, Convertro, etc.) or using a more customized combination of data management and data science. Below we provide considerations intended to inform your investigation with both types of solutions in mind.
Read the full document