I have two independent variable such as printed and digital channel which is actually derived from one column (channel_type) and dependent variable is sales. Have two goals in mind.

1. Derive the contribution of printed and digital channel

2. Find the frequency and reach curve (response curve) to decide the optimal number of both type of channel.

Want to apply market mix modelling , the problem is we can't directly apply linear regression as there is issue with saturation. Since dataset is at Customer * Week level.

Could you please suggest best model to use or analysis to use for this problem