In general, we discourage researchers from trying to compute WTP, as it is fraught with problems and likely will overstate willingness to pay. We have written an article that discusses the challenges involved and warns against calculating WTP. That article is at: https://www.sawtoothsoftware.com/download/techpap/monetary.pdf
That said, when conditional pricing is in place, and IF you must calculate WTP, we would follow the market simulation approach where you find the indifference price between two alternatives by finding the price that makes respondents on average indifferent between the two offerings (50/50 share split). This is discussed in the white paper I just referenced.
This means allowing our market simulator (choice simulator) tool to automatically handle the issue of conditional pricing. The software knows how to interpret and deal with conditional pricing.
When using conditional pricing, it is important to be on the lookout for potential significant interaction effects between the price variable and the conditional attribute. Interaction effects can be specified in addition to the main effects in the utility estimation dialog within logit, latent class, or HB utility estimation.