Logistics and Transport Costs in Milk-Run Systems

A milk-run delivery system uses a single vehicle to collect (or deliver) small lots from multiple suppliers on a fixed route, rather than direct one-to-one shipments. Originating in dairy collection, milk runs are now common in manufacturing and retail supply chains to support just-in-time (JIT) supply. By consolidating loads, milk runs can greatly reduce travel distance and fuel use. For example, Toyota Motor Manufacturing Indonesia found that a 10-day milk-run loop (covering 3,265 km) replaced multiple direct deliveries (3,812 km), cutting total distance 14% and reducing transport costs by 43%.

Milk-run networks are thus designed for efficiency: one vehicle makes stops in sequence and returns to the depot, maximizing load factor and minimizing idle trips.

Transport cost is typically broken down into several main components, most of which also apply in milk-run operations:

Vehicle Operating Costs

This covers fixed costs plus variable ones. Firms assign a blended rate per kilometer. One study used Rp 11,000/km plus a fixed Rp 1,500,000/day. High utilization, above 85% in milk-runs, spreads these costs across more cargo.

Inventory Carrying Costs

Frequent small deliveries reduce on-site stock, freeing working capital and cutting storage overhead. For just-in-time (JIT) operations, this indirect saving is often as valuable as the transport savings.

Collectively, direct transport costs (fuel, labor, maintenance, tolls) dominate.

For example, a logistics breakdown lists “fuel, vehicle maintenance, driver wages, tolls, and carrier fees” as the core transport costs. Milk-run planning explicitly targets these: by shortening routes and filling trucks, it minimizes fuel use and driver hours per unit shipped.

Consider a real case from Toyota Motor Manufacturing Indonesia:

Table 1

By consolidating pickups into one loop, Toyota cut mileage by 14%, labor time by 13%, and total transport costs by 43%. On top of that, vehicle utilization jumped to over 90%.

Designing a cost-efficient milkrun involves solving a complex routing problem. Common mathematical models include:

Heuristics and Metaheuristics

Because VRPs are NP-hard (nondeterministic polynomial time), practical solutions for real networks rely on heuristics. Tabu Search, Ant Colony Optimization, Genetic Algorithms, and saving heuristics are common. For example, one study used Tabu Search to optimize an automotive milkrun, improving load rates and cutting travel distance significantly. These methods trade optimality for tractable run-times and can incorporate business rules.

Time-Indexed Models and VRP Variants

When pick-up/delivery times matter, the model becomes a VRP with time windows (VRPTW). This adds scheduling variables (arrival times) and increases complexity, often requiring specialized solvers or heuristics. Some research also considers “progressive lanes” or collaboration across companies (4th-party milk runs) to share routes and costs.

In practice, many companies use specialized transportation management software that encodes these models. For instance, one TMS with built-in milkrun planning calculates optimal loops to minimize cost. Overall, cost-minimization formulations for milkruns are variants of the VRP, aiming to reduce total distance or time (and thus fuel and labor cost) subject to capacity and timing constraints.

Automotive Manufacturing

Leading automakers rely on milk runs for parts supply. Toyota has long used milkrun loops to supply plants, which reduced its parts logistics cost and lead time. In one case Toyota Motor Manufacturing Indonesia saw a 43% transport cost reduction (see Table 1) by switching to milkrun pickup. Audi and other OEMs integrate milkruns (often combined with rail) in Europe to cut carbon and cost on their supply routes. Webasto and Valeo have reported similar benefits in their parts networks.

Retail and FMCG

Some retail chains use milk runs to restock urban stores from city depots. By routing one truck through multiple stores, they lower last-mile expenses. (Detailed public cases are scarce, but the principle is analogous to Amazon’s internal milkruns). Third-party logistics providers also bundle shipments for clusters of customers on recurring routes, sharing costs among orders.

Retail and FMCG

Some retail chains use milk runs to restock urban stores from city depots. By routing one truck through multiple stores, they lower last-mile expenses. (Detailed public cases are scarce, but the principle is analogous to Amazon’s internal milkruns). Third-party logistics providers also bundle shipments for clusters of customers on recurring routes, sharing costs among orders.

Intralogistics (Factory/Plant)

Within factories or distribution centers, milk runs are implemented by conveyors or AGV trains. Replacing point-to-point forklift moves, milkrun systems (sometimes called “tugger trains” or “Mizumashi” systems) dramatically cut internal transport costs. For example, an intralogistics study reported that switching to milkrun trains reduced operating costs ~20–40% and cut CO₂ emissions ~33%. Regular, fixed loops of material carts help factories avoid bottlenecks and lower forklift idle time.

These examples consistently show that, after upfront planning effort, milk-run logistics yield lower total cost and leaner inventories compared to disjointed shipments. Most reported studies confirm reductions in fuel use, distance, and time when shipments are consolidated. For instance, an internal parts study noted that fully-loaded milkruns can cut transport costs by roughly one third.

In calculating milkrun cost, firms combine the above components into cost formulas. A simple cost-per-route expression might be:

For example, the Toyota Indonesia case used:

They also estimated handling time (1 min per m³) to compute route time, which influences driver pay. Many linear models assign a cost per distance and cost per hour, summing these to get a total.

More advanced mathematical optimization models (CVRP) build cost into the objective function. For instance, an MILP might minimize Σ (cost_per_km * dist_ij * x_ij + driver_cost_per_hr * time_ij * x_ij), subject to capacity and routing constraints. In these models, the “cost” per leg often includes fuel, wear, and labor prorated by distance/time. Some models even add penalty terms for late deliveries or carbon emissions, reflecting environmental cost.

Table 1 above illustrates concrete cost results; other studies model cost reduction in percentage. In general:

🔹Consolidated milkruns shrink variable costs by reducing empty miles.

🔹High load factors reduce the share of fixed truck costs per part.

🔹Frequent stops increase handling labor, but this is usually offset by overall shorter routes.

Milk-run logistics turn multiple small shipments into one consolidated loop, which fundamentally changes the cost structure. By pooling deliveries, milkruns cut fuel, labor hours, and empty mileage while increasing vehicle utilization. Key cost drivers remain fuel consumption, driver wages, vehicle operating cost, and loading/unloading time, but each is optimized via routing. Modern approaches use VRP-based optimization (MILP and heuristics) to design cost-minimal routes. Real-world cases (automotive, retail, internal logistics) consistently show large savings in transport cost (20–50%+) from adopting milk-run versus point-to-point methods. In summary, milkrun delivery requires more planning but typically yields substantially lower overall logistics cost in JIT and multi-stop supply scenarios.