sci.op-research

Hi experts,

i'm trying to develop a model based on the neg binomial ditribution,
to report profit and loss.

The scenario is that an advert is shown on a website, with an initial
click-thru rate. In my mind the advert will model the distribuiton
because a user will either click on the advert or wont and the mean
for the distribution will be the average click thru rate.

What i'm trying to find to model is the number of adverts i should
show on that site before making a loss, based on the average cost of
an advert and the average return if somebody clicks thru and buys
something.

So essentially i want the model to say after 512 adverts have been
shown on the website using the specified distribution the probability
of making a profit is less than that of making a loss, hence do not
buy more than 512 adverts at the start of a campaign

Can anyone give me any clues on how i would approach this? Ideally i
wan't to build this out into a programme say in python to do this
automatically.

I'm struggling with finding the profit loss threshold

Many thanks

Mike