Context: Managing uncertainty while expressing model transformations is problematic. Indeed, we are constrained to express various transformation specifications implementing the different possibilities. These possibilities are driven by the need to realize the relevance of each scenario to choose the best one (uncertainty on a transformation scenario) or result from the need to propose other alternatives to a given scenario if it is not feasible (uncertainty on the feasibility of a transformation scenario). In both cases, we face maintainability issues related to handling separated and frequently changed transformation specifications. Objective: This paper gives a global overview of our approach to deal with uncertainty in model transformations while focusing on maintainability aspect. Methods: We have proposed a new approach for dealing with uncertainty in model transformations. Basically, our approach makes use of partiality to allow expressing a transformation specification that covers different possibilities. The current paper focuses on the impact of our proposal to enhance changeability. This has been demonstrated by carrying out comparative experiments involving three other transformation techniques while considering the effort required to implement a change. Results: Our experiments show that our approach has proven useful and effective for implementing changes, mainly for complex ones. Conclusion: This paper provides an overview of our approach for managing uncertainty within model transformations. Mainly, it focuses on the impact of our proposal to enhance changeability. The experiment results reveal that our proposal allows expressing highly changeable specifications.