Location,TX 77379,USA

Machine Learning for Profitable Ad Sets

Matt Clayton | Business Intelligence & Analytics

💪 This tool used supervised machine learning through logistic regression to identify unprofitable and promote profitable ad sets quickly.

⚙️️ To build it required a working knowledge of statistics, plenty of SQL, and crafty spreadsheet skills.

😌 The result was a model that accurately predicted unprofitable ad sets allowing them to be shut down in days rather than a week or more. This allowed ad spend to be redirected to profitable ad sets netting higher margin.

My favorite part about this tool is that it was implemented entirely in Google Sheets, although it could have easily been built with R or Python. The advantage being that the data and insights are more tangible to less technical audiences when shared within spreadsheets. Trust through transparency!

Leave a Reply

Your email address will not be published. Required fields are marked *