Have an idea?

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

ACBC exercise including segment with different base price

We are doing an ACBC study of automotive client. This study has 3 segments each having different base price.

We have used base price pragmatically and its working fine in summed price.

However, we noticed that base price is not adding in utilities. Does it happen that base price never gets added in utilities or is it happening in this case.

Should we report output of each category separately?
asked Aug 11, 2017 by Saurabh Aggarwal Gold (27,930 points)
retagged Aug 11, 2017 by Walter Williams
Could you please explain more what you mean by "base price is not adding in utilities"?
Example: Base price of Segment 1 is 50,000, Segment 2 is 60,000, Segment 3 is 70000.

So, my prices in linear price values should be atleast 50,000. But two values in price attribute which are coming are 20,000 and 90,000. We have varied the summed price just by 10%.
We generally recommend you vary the price with random shock about +/-30%.  Varying price by only +/-10% could lead to poor estimation of price slope.  So much depends on how large the base price is compared to the prices attached to the attribute levels.  If the base price is a large proportion of the total summed price, then you can make price shocks a bit smaller and still obtain reasonable precision of the price slope.  

We describe these issues regarding summed pricing and how the amount of random shock affects the precision of the price slope in http://www.sawtoothsoftware.com/download/techpap/price3ways.pdf

1 Answer

0 votes
When you say, "But two values in price attribute which are coming are 20,000 and 90,000." I assume you are talking about the two extreme price values that the software automatically fills in the 30x1 table under HB estimation "Attribute Coding + Price".

Regarding the two values in that table, the software does its best to guess what the min and max prices used were, by examining the questionnaire settings.  But, if you do something complicated such as scripting your base price to be different for different segments of respondents, then it can guess incorrectly.  This does not bother proper data collection and simply may be reset properly by you before computing utilities under HB.

Just change the min and max values to cover the full range of prices that were actually shown across all respondents.  These extreme end points are only used for converting the price slope to discrete utility endpoints that the software moves forward into the market simulator to be used as endpoints along the price slope.  When the simulator computes utility for any price point along this continuum, it uses a straightline interpolation that is faithful to the original linear utility estimation.  So, everything is OK.
answered Aug 11, 2017 by Bryan Orme Platinum Sawtooth Software, Inc. (133,765 points)
THanks, this helps