This study in dairy cow feeding explored the effects of growth stage and chop length for grass‐clover silage, and of concentrate protein level, including the interactions of these three factors. The experiment was conducted with 48 Norwegian Red cows according to a cyclic change‐over design with four 4‐week experimental periods and eight dietary treatments in a 2 × 2 × 2 factorial arrangement. Round bale silage from a mixed sward of timothy, meadow fescue, and red clover was harvested at two growth stages (H1, H2) and offered to dairy cows either unchopped with median particle length of 152 mm or chopped to 19 mm median particle length before feeding. A low protein (LP) concentrate containing 7% lignosulphonate treated soybean meal (SBM) and a high protein (HP) concentrate containing 12% solvent extracted SBM were compared. Silage was offered ad libitum and supplemented with 8 kg concentrate per day. Silage dry matter intake was 2.0 kg higher for H1 than H2 silage, which supported a higher yield of milk, energy‐corrected milk and the milk constituents, fat, protein, and lactose. All these measures of milk production, as well as feed energy efficiency and N use efficiency, increased when silage was fed chopped rather than unchopped. Lignosulphonate treated SBM spared N in the diet without milk yield reductions. For efficient milk production with high energy and N utilisation, we recommend chopping to 20 mm for timothy, meadow fescue, and clover silages harvested in late stem elongation or in reproductive growth stages.