An attractive business model for service brokers and aggregators is the creation of composite services built from services offered by third parties. These providers generate profits by charging their customers a premium above the charges they themselves pay to the third party providers. When multiple providers offer services with the same or similar functional and non-functional properties, selecting services with lowest associated costs is clearly beneficial. However, this is not a trivial task when, as is often the case, third party providers have complex service charging models incorporating discounts based on the context in which their service is used. This context can refer to, for example, which other services a service is composed with, or the time interval in which the service is invoked. We present a service selection algorithm that takes into account time-sensitive intra and inter provider discounts which serves to minimize the expected cost to the service aggregator of offering a composite service within a specified time interval.