Marketing mix modelling: Is the most unethical work a data scientist might be asked to do?
What has been your experience in MMM?
I know DS are asked to use Bayesian to get any random explanation that you make the sales spending look valid?
Many cases see them manipulate data to generate this as well.
Almost every DS I met hated it. What's your experience?
Clients already have a graph in mind, and till u manipulate ypur data enough to show them exactly what they want, you have to keep iterating
😂 Fck. But what is the intuition behind keeping this funded? Is it that leadership just doesn't know how this shit works yet?
Leadership blindly trusts in the ability of the data science heros! They dont go over the nitty grittys. And as long as it makes "sense", they keep their praise on middle managers. Middle managers are under pressure to bring things which make business sense to the table. Hence this gross manipulation is the norm! And most companies dont do MMM on their own. They use 3rd party vendors, so the blame jn case something goes off can be squared off without any internal butchering
Yeah I completely agree. It makes me feel nauseous when I have to "lie" or "manipulate" data to make them feel like they are doing theiur job right.
