Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Advice on conditional pricing

I have a study where I have price linked to 2 attributes, brand and pack size. The client wants to make it so that the smaller the pack size the more expensive the product is per kilo. So for example the following prices could be shown for the same price level. eg.

Brand 1
Large pack (1kg)

Brand 1
Medium pack (500g)

Is the best way to approach this to have a 3 way conditional price and calculate interactions between brand and price/ brand and pack size in HB? Most of the literature I have seen on conditional pricing only involves additional prices for features e.g. extra RAM +$10 etc so any advice would be welcome.

asked Jun 21, 2012 by anonymous
retagged Apr 21, 2014 by Walter Williams

1 Answer

0 votes
Conditional pricing seems like it could work for this fine.  You could set it up as 3 separate attributes:

Package Size

And, then make a conditional price table, with conditional price based on package size.  

Depending on your sample size, and how many brands and package sizes you have, you might also consider collapsing brand and package size into a single factor (attribute), if you want the ability to include the 3-way interaction effect among brand + package + price.  But, that's icing on the cake.

Conditional pricing is, in my opinion, probably MOST used when doing FMCG studies and brand + package + price, so I'm surprised you don't think it is discussed much in the literature.
answered Jun 25, 2012 by Bryan Orme Platinum Sawtooth Software, Inc. (174,440 points)
Thanks for your answer, unfortunately there are many package sizes and the client wants to keep base size costs down as usual!
My point about the conditional pricing was that most of the examples I had seen were additive, rather than what is really a ratio in this case so I wasn't sure if that would work.
Sounds like the 3-way interaction won't work.  The idea of displaying the dollars per pound as well in the concept shouldn't be of concern.  Just make sure to follow the rules about developing your conditional pricing table (as expressed in our software documentation on conditional pricing).  If you do something funny in there, it could lead to a required 2-way interaction effect between package size and price, and that could lead to a lot of expense for you in terms of numbers of parameters to estimate in the model.  You want to give yourself the best opportunity to fit the most amount of variance using main effects.