{ "background": "Industrial machinery fleets in West Africa face persistent yield inefficiencies due to suboptimal operational protocols and maintenance scheduling. Existing optimisation studies often rely on retrospective data or simulation, lacking rigorous field validation in the regional context.", "purpose and objectives": "This article presents a methodological framework for conducting a randomised field trial to measure and optimise the yield of industrial machinery fleets. The primary objective is to detail a replicable procedure for evaluating incremental improvements in operational output through controlled intervention.", "methodology": "A cluster-randomised field trial was designed, assigning matched pairs of machinery units from Senegalese industrial sites to either an intervention or control group. The intervention involved a revised predictive maintenance and task-batching protocol. Yield was measured as productive output per fuel unit. The core analysis employs a linear mixed-effects model: $Y{ij} = \\beta0 + \\beta1 T{ij} + uj + \\epsilon{ij}$, where $Y{ij}$ is the yield for machine $i$ in site $j$, $T{ij}$ is the treatment indicator, $uj$ is the site random effect, and $\\epsilon{ij}$ is the residual error. Inference is based on robust standard errors clustered at the site level.", "findings": "This methodology article does not present empirical results from a completed trial; the findings pertain to the developed framework's properties. Simulation-based validation indicates the design provides 80% power to detect a yield improvement of 15% or greater, with the model effectively controlling for inter-site heterogeneity.", "conclusion": "The proposed framework provides a robust, context-sensitive methodological blueprint for empirically testing yield optimisation strategies in industrial engineering settings. It addresses a gap in field-experimental approaches within the region's engineering literature.", "recommendations": "Researchers applying this methodology should conduct a pre-trial power analysis specific to their fleet characteristics and ensure blinding of output assessors where feasible. Future applications could adapt the model to incorporate environmental covariates.", "key words