Service recovery has attracted increasing attention in recent years as a re
sult of the premise that service failures are inevitable, but dissatisfied
customers are nor. However. many methodological obstacles. e.g the question
of how failure and recovery incidents are collected, have not been overcom
e vet In this article, the author suggests a process approach by which not
only dissatisfied or complaining customers are surveyed but due attention i
s Paid to a representative sample of both satisfied and dissatisfied custom
ers. This approach is supported by the corresponding results, which is not
surprising, since failure incidents and recoveries ave indeed specific to i
ndvidual processes. An analysis of the effect of good recoveries resulted i
n the recovery paradox being found in all but one process type.