search

pymc-labs / pymc-marketing

Bayesian marketing toolbox in PyMC. Media Mix (MMM), customer lifetime value (CLV), buy-till-you-die (BTYD) models and more.
https://www.pymc-marketing.io/
Apache License 2.0
704 stars 195 forks source link

Expose optimization of media transformations for alternative priors specification #1038

Open wd60622 opened 1 month ago

wd60622 commented 1 month ago

The media transformation are defined in pytensor could be leveraged in a function optimization in order to get parameters from more intuitive statements. For instance,

"At (scaled) spend of 1, I will have a contribution of 0.25"

saturation = LogisticSaturation()

# Some lam and beta values
most_likely_parameters = saturation.desired_output_at_value(spend=1, desired_output=0.25)

I think something like this could be possible

What do marketer intuition tend to look like?

sonriks6 commented 1 month ago

This is exactly what I need!

wd60622 commented 1 month ago

This is exactly what I need!

Hi @sonriks6 Thanks for the feedback. Are there any other media spend spends that resonate well that might be missing?

sonriks6 commented 1 month ago

Yes, we look at the ROI level expected regarding the net revenue and the media driven sales, which by default are estimated to contribute 25% (that's the marketers intuition)