Abstract
Modelling has become a routine part of ecological and evolutionary research, yet its practice often lacks a clear conceptual framework. I propose that modelling can be fruitfully understood as experimentation. Like empirical studies, modelling projects involve treatments, levels and responses: parameter regimes or data manipulations serve as treatments, replicated runs yield summaries and comparisons across conditions reveal main effects and interactions. This framing sharpens design, reduces mission creep and clarifies communication. I outline common layers of abstraction, highlight good modelling habits and argue that thinking like an experimentalist fosters rigour, reliability and credibility without constraining creativity.