Markets for credence goods are classified by experts alone being able to identify consumers' problems and determine appropriate services for solution. Examining a market where experts have to invest in costly diagnosis to correctly identify problems and consumers being able to visit multiple experts for diagnosis, we introduce heterogeneously qualified experts. We assume that high skilled experts can identify problems with some probability even with low effort, e.g. due to education or experience. In a laboratory experiment we show that, against our theoretical predictions, this does not lead to a drop in experts' willingness for high diagnostic effort. However, consumers generally distrust experts' diagnosis effort as they buy less often after their first recommendation than it would be optimal and show frequently other non-optimal behavior patterns, e.g. receiving recommendations but do not buy service. Our results indicate that, at some level, introducing higher qualified experts increases the quality of diagnosis, as well as consumers' trust resulting in more and quicker service purchases.