BookmarkSubscribeRSS Feed
Lapis Lazuli | Level 10

I would like to perform some basic statistical test to establish whether certain customer segments are more price sensitive than other. For each customer segment (CustomerSegmentId) I have samples of how many units were bought of one specific product (NumberOfUnits) at each price (Price). The data structure is as follows:


CustomerSegmentId Price ProductId NumberOfUnits


Certain customer segments have much lower samples than others, making it an unbalanced problem. This means that I should use PROC GLM rather than PROC ANOVA using code along those lines:  



proc glm data = SomeData;

  class CustomerSegmentId ProductId;

  model NumberOfUnits

    = Price CustomerSegmentId ProductId;




I know that this community does not exist to answer statistical questions but the only site I am aware of Cross Validated:


is not very responsive (please suggest other sites).


Is the above a good starting point? Also how do I perform post hoc tests to answer questions as to whether CustomerSegmentId=1 is more price sensitive than CustomerSegmentId=2?


I also had a look at choice set approaches, which use for example logistic regression. Unfortunately, I only have observational data in this format:


TargetProductId ComparableProductId TargetPriceProductPrice ComparableProductPrice CustomerSegmentId TargetProductBought

1          2          23         25         1          0

1          3          23         25.50    1          0

1          4          23         21         2          1


Here we look at a target product at the time and we can establish if another comparable product of a customer was viewed. We know the price of the target product and the comparable product. We also know if the target product was bought by the customer belonging to a certain segment (TargetProductBought = binary).


Perhaps one could fit a logistic regression model using these product pair data (there would also be independent variables for each customer segment etc.)? I am aware of great publications by Warren F. Kuhfeld, e.g.:


but I am not sure whether my data described above could be used.


Any feedback would be very much appreciated. Thanks!

Jade | Level 19

You might want to move this topic to the SAS econometrics and Forecasting forum in the "analytics" group.  I suspect your most knowledgeable respondents will be over there.


One comment I would make though.  Your models as specified, regardless of post-hoc test choices, presumes a linear effect of price on demand.  If you have lots of price points, you probably should consider non-linear effects.

The hash OUTPUT method will overwrite a SAS data set, but not append. That can be costly. Consider voting for Add a HASH object method which would append a hash object to an existing SAS data set

Would enabling PROC SORT to simultaneously output multiple datasets be useful? Then vote for
Allow PROC SORT to output multiple datasets




Registration is open! SAS is returning to Vegas for an AI and analytics experience like no other! Whether you're an executive, manager, end user or SAS partner, SAS Innovate is designed for everyone on your team. Register for just $495 by 12/31/2023.

If you are interested in speaking, there is still time to submit a session idea. More details are posted on the website. 

Register now!

What is Bayesian Analysis?

Learn the difference between classical and Bayesian statistical approaches and see a few PROC examples to perform Bayesian analysis in this video.

Find more tutorials on the SAS Users YouTube channel.

Get the $99 certification deal.jpg



Back in the Classroom!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 1 reply
  • 2 in conversation