News for hackers. Stuff that matters.

A comment on Quantity discounts on a virtual good: results of a massive pricing experiment.

Re:Quantity discounts on a virtual good: results of a massiv... by kooshball on Saturday July 09, 2016 @02:33AM (#12060408)
Super interesting.

> Subjects were randomized into one of the four possible price schedules, with a 20% probability of assignment to the standard discount or the radical discount and a 30% probability of being assigned to the enhanced discount or the deep discount.

Does anyone know why they would split the experiment group into 20%/30%/30%/20% instead of 25% each?

[ Parent | on HN ]
Re:Quantity discounts on a virtual good: results of a massiv... by Dwolb on Saturday July 09, 2016 @03:21AM (#12060532)
It might have to do with response rates.

If you feel response rates to your experiment are going to be low for a certain group, you make the group bigger so you can have a more robust sample in which to characterize respondents of that group.

[ Parent | on HN ]