Good question, i’m also looking into ways to better forecast our marketing spend & sales. When Googling’ I came across a marketing forecast built with an online tool called Causal app.

[https://mackgrenfell.com/marketing-forecaster](https://mackgrenfell.com/marketing-forecaster)

I’m not a data scientist so not 100% sure how the Causal app works, but apparently it uses a method called Monte Carlo simulation which runs a lot of simulations based on input ranges (so in your example on CVR, instead of using 2% CVR as input, you can give a range between 0,5% and 2,5% and this gives you an output with a probability distribution).

But also following this thread for any other tips.