Abstract: A computer-implemented method of making a semi-custom product for a user, includes the steps of: (a) providing a set of data files, each data file representing a distinct variant of the product; (b) providing personal data from the user, the personal data including at least first and second distinct user attributes; (c) providing product preference data from the user, the product preference data including at least first and second distinct product attributes; and then (d) ranking the set of data files with (i) the personal data and (ii) the product preference data to identify a best fit data file, the best fit data file representing a variant of the product that most closely meets the product preference data based on the personal data.