Oscillations in Lag and Carryover Curves After Adding Control Variables to MMM Model

[nolepegoalborrego]

I'm building a Marketing Mix Modeling (MMM) model. With previous datasets, the lag and carryover curves (i.e. the delay before the impact of media investment shows up, and the residual effect over time) usually have a smooth, intuitive shape—similar to a beta function—with a buildup period followed by gradual decay.

However, with my most recent dataset, after introducing control variables, the generated curves started to show unexpected oscillations. You can see this behavior in the attached screenshot. When I remove the control variables, the curves revert to the expected smooth pattern. Any ideas why this might be happening?

Regards.