> Imagine you’re a marketing-degree graduate in middle management. If you’re lucky, you may have done some statistics courses. In your day-to-day, you likely don’t apply these to a great amount. You hear about this great thing called “Machine Learning” which is sweeping marketing departments. You might have read a white paper or two about it, heard about it at a conference, or been schmoozed by someone from IBM trying to sell you Watson (ugh). You decide to talk to a consultancy about it, or are referred to one by upper management, who’s been living from cocktail-to-cocktail on the consultant’s corporate AMEX for their entire career.

Lets be real - when you’re in this position, the consultants say a lot of words to you that you don’t understand and then you say yes to a contract because the PowerPoint looks good. Not only do you get to mention your huge new budget on your resume, thanks to the huge invoice, but you also get to mention you implemented MACHINE LEARNING at your company. But what’s next?

I have been working in advertising industry (background in software engineering) for 3+ years as an engineer (for one of the four biggest agencies). I think the author overestimates how much data we can/do collect, and how much we leverage on it. But the excerpt above is absolutely spot on. I'm a mid-level manager and I sometimes baffled at how much bullshit some companies try to sell to our clients (e.g., some big CPG companies), and how much the management folks from our clients and our advertising agency/company alike buy into that bullshit. "ML/AI" and "Media attribution" are really big now. But I believe almost 99% of the people who are leading these efforts (at least in my company) have no clue as to how to do it--thus, the snake oil salesmen are taking advantage of this.