The evolution of cooperative behavior is one of the most tackling challenges within evolutionary biology. By providing a public good, cooperators have an reproductive disadvantage compared to cheaters which safe the cost of giving the benefit. In simple evolutionary dynamics cooperators are hence outcompeted by cheaters and become extinct. A broad variety of mechanisms have been proposed to solve the dilemma. Here, we study the role of fluctuations on the evolution of cooperation. In recent years the role of fluctuations for the outcome of the evolutionary dynamics moved more into focus. In this work we expand standard approaches of evolutionary game theory and consider the interesting case, where the strength of fluctuations within the system depends on the current state of the system, i.e. the fraction of cooperators. By this, fluctuations are asymmetrically enhanced and couple back into the evolutionary dynamics. As a net effect of our results, the selection pressure towards a defective behavior is suppressed and we are able to find transient maintained cooperation. We confirm our results with extensive stochastic simulations and give rules of thumb in which parameter regimes our model can explain the evolution of cooperation without more advanced patterns of behavior, like punishment or reciprocity.