This paper reports on an industrial success story of using mutation testing for evaluating and improving test suites for smart contracts. ERCx is an industry scale, hand-written test suite and framework for testing smart contracts,developed by Runtime Verification. Hand-written tests often overlook edge cases. To improve test suites, mutation testing employs a systematic approach of introducing small, syntactic changes to a program. These changes, called “mutants”, act as proxies for real faults. Mutants that are not detected by the test suite indicate potential weaknesses and can be used as test goals. We show that mutation testing finds test cases that were missing from ERCx, even after substantial time and effort investment from developers. Based on “live mutants”, we found five missing tests. One of them was deemed critical by the ERCx developers indicating that all contracts must pass this test and all five are now part of the ERCx test suite. The generated mutants revealed interesting correlations between the tests. Using this correlation model, we developed a mechanism for test minimization without sacrificing coverage. The minimized test suite is equally effective in finding bugs in real-world programs while being faster to run.