Campaign Effectiveness Regressions

Overview

Probability of Conversion

If the marketer can indentify the number of people that can be exposed to a given campaign, she can then compare the number of conversions (new customers) to the the total number of people exposed to the campaign. This gives a rough estimate of the probability of conversion.

If in addition the knowing which customers converted, the marketer can identify other factors that may affect the conversion probability (for example, target income or demographic etc....), then she can apple the tools of Logistic Regression to try to estimate the effect of each factor on the probability of conversion.

Revenue

When the result of a campaign is not a binary event, such as a new customer or not, but rather the results represent a continuum, the more approprate type of regression is the OLS Regression