I have run an ordinal logistic regression. To keep it simple and to the point, I have found that customers with whom we perform a particular sales event are 1.5x more likely to upgrade. Of the customers who have upgraded, through linear regression I have found that they are willing to pay $10 more for said upgrade than the ones who have not had the sales event.

If I wanted to say something like: for each customer on which we perform the sales event we will obtain $XX more in revenue (due to them upgrading AND paying more than the standard customers), how would I calculate it?

It is more or less combining the odds ratio with the additional revenue, but I do not know how to calculate what the combined effect is here.

Any pointers would be appreciated.


1 Answer 1


(I'm assuming if they don't upgrade, you make $\$0$.) This is really quite straightforward. If every customer you give your sales pitch to has a $p$ probability of upgrading, and each upgrade will net you, on average, an extra $\$10$, then you expect to make, on average:
\begin{align} d &= 10p\qquad \text{(per person)} \\ d &= 10Np\quad \,\text{(total)} \end{align} where $d$ is dollars, and $N$ is the number of customers you give your sales pitch.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.