Textbooks
1. Bayesian Data Analysis, 3rd Edition https://www.math.muni.cz/~kolacek/docs/bayesian_data_analysis.pdf
- This textbook is seen as the best textbook for learning Bayesian statistics. The first 5 chapters provide the fundamentals and some parts could be quite technical.

Papers
1. I've tried using the R package CausalImpact, which is an implementation of the methodology explained this paper (https://research.google.com/pubs/pub41854.html https://google.github.io/CausalImpact/CausalImpact.html). If you have time series data at some group level (e.g. city level) and are interested in estimating the effect of a treatment on one city, the algorithm finds a combination of other cities (i.e. control cities) to forecast what the response variable of the treated city would be without the campaign; this is the synthetic control. The difference between the forecasted and observed response in the treated city is the causal effect. If the time series data is messy or if the effect size is too small, then you wont be able to detect a stat sig effect.

2. This dude writes about lots of marketing issues related to attribution or causal inference. https://research.google.com/pubs/JimKoehler.html