This dissertation aims to develop a model by using experiment planning techniques applied to cold-rolling process to manufacture special steels utilized in the carbonated beverage market segment. The process consists of downgauging hot-rolled coil, which is further pickled for oxide removal, into a coil with the final thickness required by customer. It involves controlled and uncontrolled variables, which interact with one another, thus affecting the response variable, i..e, product´s final thickness. Some of the fundamental questions to be asked include: How do controlled variables affect response variable? What are the most influential ones? Are there interactions among such variables? Is it possible to work out an appropriate model to address this problem? The fragmented factor planning technique provides answers to the questions raised, by breaking new grounds in terms of using this powerful tool in manufacturing processes , where situations similar to those suggested by this study are fairly common. It includes a description of the problem, a bibliography review with the required steps to conduct the experiment planning, choice of controlled variables which make up the experiment as well as its execution, result analysis, an approach to Generalized Linear Models, and validation of model by adjusting process analysis as shown in the result analysis. Applying this methodology and implementing the proposed changes bolstered the process in such a way that even when disturbances of uncontrolled variables occur the thickness remains within the specified tolerances. Additionally, a further reduction in variation ranging from ± 1.5% to ± 1.0% was achieved, where technology investment would be required, thereby making steel products more expensive and less competitive against other competing products, thus ensuring competitiveness.