The aim of this short project was to optimize demographic targeting for a financial product, the main question being: What features define those who convert the least?
The data was extracted from Everflow, which doesn’t provide customer-level data. Analysis was therefore done on a per-click basis. The clicks were classified as:
Converted: 993 clicks
Not converted: 8,920 clicks
Unique: 9,320 (10.16% CVR)
Not unique: 593 (7.76% CVR)
On top of click classifications, other data provided were city, phone model, source websites, type of credit card, housing payments, and monthly income.
While there were differences in traffic and conversion rates for the last four features, the most pivotal one proved to be monthly income.
Madrileña from Manila 