Fundamental Factor Models In Practice

Hello all, I’m having difficulty in understanding how one would calculate a fundamental factor model in practice. Specifically coming up with the standardized betas first and then estimating the factor returns second through a regression.

Suppose you were modelling a mutual fund’s return through a fundamental factor model. In order to get the factor sensitivities would you not have to know how the entire universe of mutual funds’ have behaved relative to, for example, the HML factor? It feels like I’m missing something.

you would basically use the benchmark average and compare it to the company/security in question, crafting it like a z-score by also using the standard deviation inclusive of the benchmark (eg. debt/equity ratio sensitivity would be calculated by [(d/e mutual fund1) - (weighted average d/e of universe)]/(st deviation of universe)