We propose a sizing methodology for electrical storage to operate loads in PV-powered industrial processes. Our approach is based on the solar variability of the installation site and its effect on the load. By pinpointing the detrimental effects on the load, the storage can be sized to tackle these effects, resulting in improved performance and contained storage unit. The method performs a day classification followed by a study of effects on the load. Once the problems are identified, the type of day is selected as input for the sizing procedure. This way, the problems will be mitigated with associated benefits for all day types, while keeping the storage unit as small as possible.