The Perfect Multitouch Attribution Model Doesn’t Exist
I recently did some online car shopping. My lifestyle had changed, and the fun little car I had been driving would no longer meet all my needs. I was immediately frustrated by the site I was using, which didn’t even offer the right filters for me to properly narrow my search. I was just looking for an all-wheel–drive electric convertible that could seat four adults comfortably and tow at least 3,500 pounds. To my shock and dismay, I didn’t find a single one!
I get it. People have different needs and tastes. But for most of us, outside our home, education, and Star Wars figurine collection, our vehicle is likely one of the biggest investments we’re going to make. And they still expect us to buy more than one to fill different needs?
I was thinking about this experience again as I worked with a client to plan a new multitouch attribution model to better analyze the impact of their marketing tactics. This was a big undertaking, and something they had been working toward for quite some time. They were carefully thinking through how different key points in the buying cycle should be weighted, whether the order in which the tactics were consumed should matter, and whether some higher-touch interactions like in-person meetings should receive more credit than more impersonal ones. Their goal was to build the perfect model for their business — one that they could use to analyze all their demand generation tactics moving forward.
They were surprised to hear that I don’t believe there is any such thing, nor should there be. If we can’t even build a car that can be all things to all people, why would we expect one model to be able to answer every tactic performance question? We work with clients to understand the specific intention of their marketing tactics, and then judge their performance within that custom context.
It’s also important to remember that no attribution model is capable of generating a precise value for the return from one individual tactic in a complex system like a B2B sales cycle. All of these models deliver estimates that we can leverage to better understand relative performance, not precise return.
For the same reason I can’t buy a peppermint–salmon–lasagna smoothy and I don’t mix chocolate milk, orange juice, and vodka (anymore), there isn’t one perfect attribution model that can mash every objective and audience together and spit out the one “right” answer. Best practice demands the flexibility to customize marketing attribution, leveraging multiple purpose-built models to answer different business questions.
Along with Conrad Mills, I’ll be talking more about how customizing multitouch attribution models can support better business decisions in our presentation “Beyond Attribution: Tactic Mix Planning Using Tactic Efficiency” at the upcoming Forrester B2B Summit North America. Learn more about this year’s B2B Summit and register.