Bifalgorithm post

If you work with marketing data as a data scientist, data analyst, or even just a marketing professional, you want to know about these advancements in marketing mix modeling. Marketing mix models are basically the holy grail that marketers are looking for to get estimates of ROI for all of their channels at once. But they suffer from problems of multi-coloniality and endogeneity, making the results typically unless you can put in priors. Now this all gets really technical, but traditionally these models would have parameters that are put in as priors that are not very interpretable and especially difficult for non-technical stakeholders to understand. But there was a recent research paper that was put out by Google, which talked about how you can take ROAS or return on ad spend that you can easily get out of experimentation or through discussions with some of your marketing stakeholders and use that to inform your models. Now, when you use these ROAS priors, your stakeholders are going to feel so much more comfortable that they know where the inputs to this model are coming from. Now, if you guys are interested, I have a much longer video on YouTube about this that you can dive in as well as a host of other data science tutorials and causal inference tutorial. But this stuff is so hot right now. I've been posting about marketing mixed models on LinkedIn and literally getting job offers just from hiring managers reaching out to me for my posts on LinkedIn. This stuff is hot and they really do not teach it in school.

If you work with marketing data as a data scientist, data analyst, or even just a marketing professional, you want to know about these advancements in marketin…