When trimming variables are feasible?

Hello everybody, this is my first publication so if i'm writing in the wrong area, i'm sorry.

i'm doing my first academic work and i'm a bit confuse with my data, my work uses the financial information from brazilian open capital firms so i have a huge range in the data.
I consider some values from my variables that i create with the original data are outliers (for example: the range from the payout firms is from -20000% to 1200%), so i'm looking to trim these variables to have lower ranges.
Continueing my example, if i trim the payout in the 5% above and below, my range goes from -0.12% to 1.20%.
Can i have yours opinion about this situation? i don't have with who discuss. thanks

Array