# Conditional Pricing with variance price level

Hi Sawtooth support team,

We're conducting a CBC study for a serial of  wide range product from low-end to high-end level, and going to implement Conditional Pricing to get rid of prohibition. Can I set variance price level for difference product? For example:

We have 2 set of product, low-end and high end product, for low-end product, we want to offer only 3-level price, and for high-end product, we are going to offer 7.

My question here are:

1) Is good practice for using variance price level for different product? What's disadvantage would be?

2) How can I restrict 4 other-level for low-end product? Prohibition?

3) Can I assign the same pricing for multi-pricing level? If I keep 7 price level for low-end product, and set the same price on level 1&2, 4&5, 6&7, is this works? But how can I get the performance share in market simulator with specify pricing? (now we have 2 price level with the exactly same price indeed).

asked Nov 11, 2013

It would be a strange thing to do to use 3 price points for low-end product and 7 price points for the high end product.  There is something in conjoint analysis called the "Number of Levels Effect" (NOL) that biases attributes with more levels to get more "importance".  That means that if you used 3 price points for the low-end product and 7 for the high-end product, you risk biasing the elasticity for the high-end product upward relative to the low-end product.

With conditional pricing, the number of price levels for the low-end product and high-end product must be the same.

With alternative-specific designs, you could create two price attributes (one to apply only to the low-end product and one to apply only to the high-end product).  That would give you the flexibility to specify 3 levels in one attribute and 7 in the other, but again, this is not advisable.
answered Nov 12, 2013 by Platinum (173,090 points)
+1 vote
Leen,
I'm not form SS support, but:
3) Yes, you can show same labels for different levels. But you have to estimate interactions between price and brand and you should estimate price as part-worth attribute.
2) Yes, the prohibitions are another way to do it, but I don't believe you get right estimates of main effects for your price attribute in this design.
1) You put yourself into many troubles using this strategy. And as far as i know it's not the best practice. Sometimes gains from this strategy can be worth the trouble, but in most situation's there are better ways to provide your respondents with reasonable choices.

You can also use attribute specific (from advanced designs module) design and nest different price attributes for different brands, but doing so you restrict the data used for estimation of the price parameters (compared to one common price attribute). But if you have only two price "channels" then it will be better then estimation of the  full brand*price interaction.

Anyway, do you really need this approach for your study? What do you want to accomplish with such a design?
answered Nov 12, 2013 by Bronze (2,830 points)