Is the response curve constructed using ad-stocked media or with original spend media?

[AruunnAnanda]

Hi there,

I would like to know if the X-axis of response curve depicts ad-stocked media or original spend media?

Thanks!

[michevan]

Thanks a lot for flagging this! I just wanted to share a link to our new "support" section in the readme to give you some more information on how we can and cannot support such questions at the moment.

[steven-struglia]

I believe it's using the original spend data looking at lightweight_mmm.py self.media = media and plot.py media = media_mix_model.media in the plot_response_curves function.

[slavakx]

I didn't debug the code, so I don't know if it includes the original or adstock spend but I think it should include the adstocked spend because we should see the effect of adstock on our spend

[steven-struglia]

I'm unsure if the model ever changes the media_mix_model.media parameter through the transformation functions besides the scaler used prior to feeding the data into the model. Though I do agree that we should be able to see the effect of adstock on spend inside the saturation curve, we should be able to see both ideally. Some sort of a parameter inside of the saturation curve function would be great to have. I have an edited saturation curve function that has a confidence interval on the spend vs. KPI curves that I'll share soon, and I can add an argument to the function like transform_adstock = True or False so we can do both.