Well, I’ve been spending a lot of time lately thinking about attribution. Some people think about their summer plans or the Oscars or Spring Training. I, on the other hand, have been thinking about attribution. The underlying questions are really quite simple, aren’t they? What is driving our pipeline and revenue? Every marketer wants to know this so we can stop doing things that aren’t working and double-down on what is working.
The challenge, of course, is that at any giving point there are so many variables in-play that attributing a desired outcome (lead, deal, revenue, etc.) to a specific marketing activity or series isn’t that easy. Sure, you can fall back on a basic attribution model of first or last-touch, but that isn’t really effective… we all know that “email” can’t be originator of all revenue this year!
So, I recently read an interesting post titled “How can the Reverend Bayes help you find out if your campaign worked?” The authors discuss how regression, interrupted time series analysis (ITS) and ultimately Bayesian structural time series (BSTS) can be applied to marketing use cases to truly identify cause-and-effect. After all, the authors wisely state that “correlation does not always mean causation”. Just because we changed our pay-per-click strategy and revenue went up (correlation) doesn’t necessarily mean that PPC was the reason (causation). I’m always intrigued when simple questions require very complex statistical and analytical models. Clearly, this is beyond the capability of most marketing teams.
Then, it hit me. Maybe, I’ve got this all wrong. Let me explain.
Like most marketing executives, we are hyper-focused on KPIs, metrics and even increasingly complex attribution models in order to prove to ourselves and everyone else that marketing is having a material, positive impact on the company. However, maybe the pendulum has swung too far in this direction. Maybe, just maybe, we’re becoming so hyper-focused on conversion rates, metrics and other proof-points that we’re sacrificing other critical elements of our role and department.
Maybe we’re spending too much time in spreadsheets and data models and not enough time in customer meetings?
Perhaps we’re looking for answers in complex regression equations and not looking hard enough at our competition?
Could it be that we’re so focused on multi-channel attribution that we’re missing insight on what content our buyers really find useful?
Are we spending more time talking to our own data scientists and not enough talking to prospective customers and partners?
Are we too worried about “doing things right” and not enough concern about “doing the right things” in the first place?
Relax! I’m not saying numbers and metrics and attribution aren’t important. Of course they are, today more than ever. However, I am concerned that marketing executives (and our teams) could become too removed from the actual markets in which we compete and the customers we serve.
As we look at our workload, review status reports and look at days of meetings, let’s be sure to ask ourselves those questions listed above. Let’s make sure we continue to make advances in important topics like marketing attribution and return on marketing investment while continuing to stay close to what made us great marketers in the first place; having the deep feel for all of the dynamic elements of the market, our competitors and customers.