WTP assumes that with your price attribute, lower prices are better. It assumes that to equalize an enhanced share it should raise prices to accomplish it. With Salary as the "Price" attribute, more is always better...so this poses a problem.
One possible solution would be to reverse the scale of the Salary attribute to "trick" the software into estimating WTP using a Salary attribute. In the market simulator, you would click the "Attribute Info" icon. Then, for the Salary attribute, you'd mark it as the "price" attribute and that it is continuous. Next, we'd modify the "Value" fields to flip the scale.
For example, let's say the salary levels in your study were:
Two ways to flip the scale:
1) Put negative symbols in front of the salary values, leading to values for Salary of:
2. You could flip the scale by maintaining the same increment spacing between each of the levels (e.g. a 10 unit difference between levels 1 and 2), but reversing the direction, such as:
When you specify your scenarios (the product "specifications") you'll of course want to specify the salary amounts on the flipped scale (according to either method above).
Now, the WTP algorithm will try to decrease the salary to drive the share of preference back down for an enhanced product (job offer) on other features.