Design of experiments has been applied in manufacturing industry for several decades. Nevertheless, there is still room for exploring its cost‐effectiveness features. For instance, many industrial experiments involve factors with levels more difficult to change than others. This naturally leads to the harder‐to‐change factors being less reset than the easier‐to‐change factors. These randomization restrictions commonly generate the so‐called split‐plot type designs. Methodologies for designing and analyzing such experiments have advanced lately, especially for 2‐level designs. However, practical needs may require the inclusion of factors with more than 2 levels. In this paper, extensions and variants of mixed‐level split‐plot designs are presented. An industry application to evaluate the impact of temperature, type of lubricant, type of pavement, and type of rubber on performance and quality of tires is discussed. Due to lack of previous adequate planning, the actual experiment was executed as an unreplicated split‐strip‐strip‐plot design. As a consequence of the inadvertent split‐plotting, evaluation of the significance of temperature and lubricant is impaired. Other design alternatives, along with their properties, advantages, and disadvantages are shown. This paper aims to contribute to a better understanding of how statistically designed experiments may be used, in a practical way, to manufacturing planning and optimization.