Dietary trials may link macronutrient intakes to health outcomes, but adherence to dietary targets requires advice based on an understanding of food composition and consumption patterns. Using data from a weight loss trial, we hypothesized that structured advice would be required for significant fat modification to occur. We compared participants' food choice patterns in response to advice based on a structured "whole-of-diet" model vs a general approach to healthy eating. Overweight participants (n = 122) were randomized to 2 advice arms (saturated fat [SFA] < 10% energy [E]): (1) general low fat (LF) control-(a) isoenergy, (b) -2000 kJ; and (2) structured LF high polyunsaturated fat (PUFA) (∼10% energy PUFA; PUFA to SFA ratio ≥1) (LF-PUFA)-(a) isoenergy, (b) -2000 kJ. Intakes of E and fat and fat from food groups (percentage of total fat intake) were compared at baseline, 3 months, P < .05. Baseline diets were similar, with most fat from high-SFA foods (59%): meat and milk-based staple meals and high-fat snacks. By 3 months, all groups reduced E and met the SFA target. Polyunsaturated fat targets were met by the LF-PUFA groups only (P < .001), enabling targeted between-group differences. In response to general advice, LF groups simply switched to LF alternatives of the same foods (P < .05). In comparison, LF-PUFA groups shifted fat intake to high-PUFA choices (54%), consuming more fat than controls from nuts (P < .001), whole grains (P < .001), and oils and spreads (P < .05). Significant reductions in E were achieved regardless of advice, but significant shifts in dietary fat profile relied on structured whole-of-diet advice on a range of meal and snack food sources of fat subtypes.