In the 1980s I got enough money to replace my Color Computer 3 with a 286 based PC doing a data processing project for my uncle who worked for this place

https://casinoballroom.com/

I entered what people filled up on maybe 2000 survey cards. I made a very primitive looking report and my uncle was interested in only one data item. My mom thought the report didn't look like it was worth very much and thought I should have done "better" but my uncle was very happy.

I found out years later that he was having a disagreement with his boss about whether or not it was worth advertising in

https://thephoenix.com/

my uncle believed that people who visited that venue didn't read the Phoenix and the survey proved it was right. They were spending upwards of $30K a year advertising in that paper so the ROI was fantastic.

Do you mean ROI of the project itself, or ROI of specific tools/methods/approaches etc?

Prescriptive analytics for Pharma.