A data set containing 2,124 corn silage (CS) gathered throughout the years 2014 to 2018 (YH) was analyzed by near infrared spectroscopy at the Valacta Forage Laboratory (Sainte-Anne-de-Bellevue, QC). The nutritional value and fermentation profile were used to perform a principal component (PC) analysis using SAS 9.4, (SAS Institute Inc., Cary, NC). Six PC were extracted by adopting criterion eigenvalue higher than 1. The PC would correspond to: PC1) Carbohydrates and maturity of CS; PC2) homolactic fermentation, protein solubilization and starch digestibility (DIG); PC3) fermentation length (FL), heterolactic fermentation or other secondary fermentations; PC4) NDF DIG; PC5) Protein degradation and heating; and PC6) L. buchneri fermentation. The PC subject scores were then annually averaged for CS sampled from the same farm from October 20th to August 31st of each YH (2014 to 2017). Then, each farm was characterized for their annual average milk yield (MY) and composition. Only farms (n = 43) with a minimum of one annual average for each respective YH including PC scores, percentage of CS in rations (% CS), MY and composition were retained and analyzed. Mixed model was used to relate MY and composition to average annual PC scores and % CS where FARM, HY and FARM × HY as random effects, PC1 to 6, % CS and PC1 to 6 × % CS as fixed effects in the model (JMP 13, SAS Institute Inc., Cary, NC.). As main results, Annual average milk yield per cow (kg/cow/year) was affected positively (P < 0.05) by %CS, PC2 and by the following significant interactions PC2 × %CS and PC4 × %CS. Annual average milk urea (mg/100 mL) was affected positively (P < 0.05) by %CS, PC1, negatively by PC2 and the following interactions PC2, PC3 and PC6 × %CS. Average annual log-transformed somatic cell count was only affected positively (P < 0.05) by PC5. This exploratory research shows the potential of using historical data from dairy farms to explore the effects of silage analysis on dairy herd performance and developing better data driven tools.