On-farm trials present several challenges for analysis depending on how the trials are conducted. In Wisconsin, on-farm foliar fungicide trials have been conducted since 2007 with a primary goal to obtain new information about the use of foliar fungicides under grower production conditions. To accomplish this goal, a protocol was developed that outlined a minimum set of requirements for conducting a trial including the use of replication and randomization of treatments. Over the past four years, a total of 42 trials were conducted using either small plot (similar to those used on research farms) or large strip (field length) methods. The number of replicates per trial ranged from two to four. Analyses have been conducted in two different ways. The first was a site-by-site analysis at the end of each growing season in order to provide growers results as quickly as possible. The second analysis was a three-stage analysis to specifically address the question, “what is the probability that a foliar fungicide application paid for itself?” This approach integrated Bayesian statistical methods to estimate the probability across different costs/prices for fungicide and seed. Results indicated that in only 1 of 42 trials was there evidence of an effect of a foliar fungicide when trials were analyzed individually (P<0.05). When combining trials for analysis, results indicated that two active ingredients, azoxystrobin + propiconazole (Quilt Xcel, P=0.0247) and pyraclostrobin (Headline, P=0.0701) had relative yield increases of 4.7% and 1.8%, respectively. Lastly, the probability that a foliar fungicide paid for itself was highest when fungicide costs were lowest and potential yield highest, but varied greatly by active ingredient. Results from these trials have been very beneficial to address a specific grower question using methods that most closely match their production practices.