Using a conditional pricing table allows you to show prices in the CBC tasks that tend to be more realistic to the features included in the product concepts. But, it leads to some additional things to have to think about if you try to interpret the part-worth utility scores.

One thing to keep in mind is that if you have 3 price points (e.g. low, medium, and high) and 6 brands...and you are using a conditional pricing table (18 cells) for the 6 brands x 3 price points, then even though it looks to you that there are potentially 18 price levels running around, the software still only keeps 3 levels of price in its experimental design. The fact that there are potentially 18 prices shown to respondents is just a display trick using the conditional pricing lookup table.

So, when you estimate utilities, you obtain utilities for 6 brands and 3 prices as before if not using conditional pricing tables. HOWEVER, you can no longer interpret the utility of the 6 brands in the traditional way you do with traditional conjoint...in an "all else equal" main effects mindset. In other words, you have to interpret the utility of a brand part-worth considering the average price that this brand was shown at. So, the brand part-worths are no longer disentagled from the price points.

Luckily, this only poses challenges for interpreting what the part-worth utilities mean. If you just move the part-worth utilities into the market simulator and just pay attention to the resulting shares of preference, everything works out properly with the model. As long as you are just looking at share of preference simulation results rather than trying to directly interpret the meaning of the part-worth utilities, it's essentially the same as conducting regular conjoint analysis.

Regarding your question whether you can use -10% to +10% price variations, if you mean that you are going to construct your conditional pricing tables by making the low price for each brand be -10% from the middle price and the high price +10% from the middle price (assuming just a 3-level price attribute), then this is one way you could do it. The general recommendation is to build your conditional pricing tables in a very proportional way, such that the variation from the mean price for each brand in the conditional pricing table is essentially equal (though some brands may have a premium starting middle point and others may have a discount starting middle point).

If you have more questions that I haven't answered well here, there are two key areas I know of in Sawtooth Software documentation where we explain how to interpret your utility results when using conditional pricing:

In the SMRT Market Simulator Documentation:

https://sawtoothsoftware.com/help/lighthouse-studio/manual/index.html?hid_web_cbc_cp.html
and, in pages 2 to 4 of the following technical paper:

http://www.sawtoothsoftware.com/download/techpap/price3ways.pdf