Discussions

I noticed some fluctuations while plotting the average total spend and raw visit counts of some Starbucks stores we have pooled from our study.

Spend

For spend, there is a weird fluctuation, a sudden drop at the beginning of 2022 and a peak in the middle 2022 when I plot for the two groups we study:

[Image](file:///Users/xguo/Dropbox (Harvard University)/Xiaowei-Isamar/Starbucks/Dewey/images/avg_spend_starbucks_bygroup.png)

The same patterns are seen with the country-level transaction data (from Entire US Panel Summary):

[Image](file:///Users/xguo/Dropbox (Harvard University)/Xiaowei-Isamar/Starbucks/Dewey/images/avg_spend_US.png)

Is it possible that this is driven by the source and sampling of spend data? We aim to compare between two groups of Starbucks stores (treated VS contral). I wonder if you have any further advice on this?

---

Foot-traffic

There seem to be a significant drop after 2023, which is also consistent when looking at the normalized variables (see below).

[Image](file:///Users/xguo/Dropbox (Harvard University)/Xiaowei-Isamar/Starbucks/Dewey/images/raw_visit_counts_starbucks.png)

[Image](file:///Users/xguo/Dropbox (Harvard University)/Xiaowei-Isamar/Starbucks/Dewey/images/Normalized_visits_by_state_starbucks.png)

I would appreciate any guidance you could provide on how to interpret and address this.