In-Plant Logistics Milk Run Implementation —— A Systematic Approach from Route Design to On-Time Delivery

1. A Production Line Crisis Caused by "Chaotic Delivery"

A certain automotive parts company (hereinafter referred to as "the company") experienced three production line stoppages in a single week, and the cause was surprisingly not equipment failure or incoming material defects, but rather the inability to keep up with material delivery. On average, line-side operators spent forty minutes per day leaving their workstations to collect materials from the warehouse; delivery personnel pushed heavily loaded carts back and forth through the workshop, with routes crisscrossing and blocking each other; whenever the production plan was adjusted, materials were piled in the aisles, preventing forklifts from passing, and safety officers posted corrective action notices daily. More troubling was that during urgent material shortages, delivery personnel could only prioritize the loudest requests, often leaving the truly needed materials untouched.

Production line stoppages cost thousands of yuan per minute, and management finally realized that the company's logistics system was still at a primitive stage of "manual labor and on-demand service." After the lean improvement team entered the scene, their diagnosis was straightforward — this was not due to a lack of effort from the delivery personnel, but rather a flawed delivery model. What the company needed was not faster material handlers but a delivery system that operates on a fixed schedule, similar to a bus route, which is the focus of this article: in-plant Milk Run.

Milk Run, literally translated as "milk delivery route," originated from the practice of milkmen delivering milk and collecting empty bottles at fixed times and routes in the early days. When applied to a factory, it involves delivery vehicles or personnel following a fixed route, time window, and frequency to cycle through and deliver materials to each workstation while also collecting empty containers. This method, though seemingly simple, is a critical leap from "push-based handling" to "pull-based delivery" in lean logistics.

2. Why "On-Demand" Delivery Inevitably Leads to Chaos

To understand the value of Milk Run, we need to identify the three major flaws in traditional delivery methods.

The first flaw is opaque demand. Only the operator knows when and how much material is needed. When they shout, the delivery person runs a trip, but no one tracks how much material is loaded or if the trip is fully utilized. Delivery personnel are always "putting out fires" and never know where the next demand will come from, leading to low efficiency.

The second flaw is irregular routes. Delivery personnel decide the order of deliveries based on experience, resulting in different routes each day. This flexibility actually wastes a lot of time: crisscrossing paths, repeated trips, and traffic jams during peak hours. The longer the handling distance, the higher the logistics cost, and the longer the material is in transit, making it harder to control line-side inventory.

The third flaw is the inability to trace responsibility. If materials are delivered late, is it a planning issue or a delivery issue? If the wrong material is delivered, is it a picking issue or a verification issue? Without a fixed schedule and handover standards, problems often lead to "blame games," making it difficult to implement improvements.

Milk Run addresses these issues simultaneously: it forces demand transparency with fixed delivery frequencies, eliminates unnecessary handling with fixed circular routes, and clarifies responsibilities with schedules and handover documents. It transforms "people looking for materials" into "materials finding people" and "random service" into "fixed schedule and route," which is the most basic and effective step in lean logistics.

3. Route Design for Milk Run: Draw the Map, Set the Stations, Schedule the Runs

The first step in implementing Milk Run is to draw a complete in-plant logistics map. Many companies skip this step and directly schedule runs, resulting in winding routes and even worse on-time delivery rates. The map should include four types of information: the types and quantities of materials required at each workstation, the storage space and material placement positions at the line side, the width of the aisles and the type of handling equipment, and the distance and travel time between workstations. This map serves as the basis for all subsequent decisions.

The second step is to determine the delivery stations. More stations are not always better, and not every workstation needs a station. Typically, stations are set up at workstations with high demand and ample space; workstations with low demand and close proximity are combined into "virtual stations," where the delivery person can stop once and distribute to multiple workstations. The number of stations directly affects the time for one complete cycle. Too many stations can result in a cycle time exceeding thirty minutes, making it difficult to increase delivery frequency.

The third step is to schedule the delivery runs. When determining the delivery frequency, two data points are crucial: the duration of material that can be stored at the line side and the transit time from the warehouse to the line side. The basic principle is "small batches, high frequency" — the higher the frequency, the lower the line-side inventory, but the higher the requirement for delivery stability. A common formula used in practice is: the delivery interval is approximately half the time the line-side storage can support, ensuring safety stock while allowing for delivery fluctuations.

The fourth step is to draw the circular route map. The route should follow the principle of "single loop, clockwise or counterclockwise, no retracing," connecting all stations into a loop. The starting and ending points of the delivery are both in the warehouse dispatch area. The delivery person loads the materials at the starting point, delivers them station by station along the loop, and collects empty containers, then returns to the endpoint to reload and start the next cycle. The route map should be posted on the delivery vehicle and in the delivery area to ensure everyone knows "when this vehicle will arrive at my station."

4. On-Time Delivery Rate: The "Heart Indicator" of Milk Run

Once the routes are drawn and the schedules are set, the success of Milk Run depends on one key metric — the on-time delivery rate. This metric measures the accuracy of delivery vehicles arriving at each station according to the planned schedule, and it is the undisputed "heart indicator" of the Milk Run system.

The criteria for measuring the on-time delivery rate should be clearly defined. A common approach is to allow a ±2-minute deviation from the planned schedule, with deliveries within this range considered "on-time." Deliveries that are early or late by more than two minutes are recorded as "not on-time." Early deliveries are penalized because they can also disrupt line-side order — materials arriving early occupy aisles and increase inventory, which is essentially the same problem as late deliveries. The recommended statistical cycle is by shift, with each shift recording the actual arrival times at each station, summarizing daily, and analyzing weekly.

With on-time delivery rate data, the next step is to analyze the reasons for deviations. Late deliveries typically stem from three issues: 1) delays in the loading process, such as picking, verification, and scanning; 2) bottlenecks in the route, such as intersections, narrow aisles, and elevator waits; 3) temporary adjustments in production plans causing demand fluctuations, leading to additional tasks for delivery personnel. Early deliveries are often related to delivery personnel "rushing" — loading materials early and starting ahead of schedule, which seems proactive but actually disrupts the rhythm.

Improving the on-time delivery rate involves a set of standard actions. The first step is to establish a delivery schedule board, visualizing the planned arrival times for each station to ensure both delivery personnel and workstations are aware. The second step is to conduct a "one incident, one analysis" for late deliveries, recording the cause, clarifying responsibility, and setting a deadline for corrective actions. The third step is to implement a buffer mechanism, reserving flexible time at the end of the route to absorb minor fluctuations. When the on-time delivery rate stabilizes above 95%, line-side inventory can be further reduced, forming a positive cycle of "on-time delivery rate improvement — inventory reduction — site improvement."

5. Loading Rate and Balancing: Preventing Delivery Vehicles from "Running Empty"

The on-time delivery rate addresses the "timeliness" issue, while the loading rate addresses the "economic" issue. If a delivery vehicle only carries one-third of its capacity, even if it arrives on time, the logistics cost remains high. The loading rate is the ratio of actual load to rated capacity, generally required to be stable at 70% or above, with a rate below 50% prompting a review of the route or frequency settings.

There are two directions to improve the loading rate. One is to consolidate stations and reduce empty runs, grouping workstations with similar material requirements into the same cycle to avoid "making a trip for one piece of material." The other is to promote container standardization, using uniform turnover boxes and material racks instead of a variety of packaging, making it clear how much material is loaded and how much is left.

Complementing the loading rate is the balancing of deliveries. Ideally, the delivery volume for each shift should be relatively stable, avoiding extreme fluctuations like "idle at the beginning of the month, busy at the end." Achieving balance hinges on production leveling — breaking down the production plan into hourly increments, which naturally levels material demand, stabilizing delivery frequency and load volume. Many companies find that the root cause of delivery chaos lies not in the logistics department but in the planning department: large fluctuations in planning make it difficult for logistics to optimize. When Milk Run is deeply implemented, it inevitably forces production plans to become more leveled, a typical example of lean improvement's "one move affects the whole."

6. Empty Container Recovery: The Overlooked "Other Half" of Milk Run

The term "circular" in Milk Run refers not only to the circular route of the delivery vehicle but also to the circular turnover of material containers. Delivery personnel deliver full containers to the line side and simultaneously collect empty containers and racks, returning them to the warehouse. This is the "return load" of Milk Run. Empty container recovery, though seemingly minor, is crucial for the sustainability of the cycle — if empty containers pile up at the line side, full containers cannot be delivered, and the entire cycle will be blocked.

The key points in managing empty containers are threefold. First, establish clear recovery standards, specifying the maximum number and duration of empty containers allowed at each workstation, with excess triggering a recovery alert. Second, maintain a container ledger, tracking the quantity, location, and turnover status of turnover boxes and racks to prevent them from "drifting" to other areas. Third, collaborate with suppliers to integrate empty container recovery into the external logistics cycle, achieving a closed loop of "full containers in, empty containers out."

A straightforward way to assess the effectiveness of empty container recovery is to check if the aisles are clear at the end of the shift. If the material is returned to its place and empty containers are cleared daily, it indicates a healthy container turnover rhythm. If the line side is always cluttered with overnight boxes, it suggests a problem in the recovery process that needs priority improvement.

7. Information Systems and Digitalization: From "Paper Schedules" to "Real-Time Dashboards"

In the early stages of Milk Run implementation, a paper schedule and a whiteboard can suffice. However, as the scale of deliveries increases, paper management becomes a bottleneck: schedule changes cannot be updated in time, information transmission lags during anomalies, and manual data aggregation is slow and prone to errors. At this point, information systems need to be introduced.

The first step in digitalization is to establish a delivery demand signal. Common practices include using electronic kanban or pull systems: when the material at a workstation reaches the set replenishment point, the operator scans the material card or presses a call button, and the demand information is transmitted to the warehouse in real-time. The warehouse then picks and loads materials according to the sequence and shift schedule. Compared to manual shouting, this method allows demand information to be recorded, analyzed, and optimized, providing a data foundation for frequency and loading rate improvements.

The second step is to visualize delivery execution. Install positioning terminals on delivery vehicles or use barcode scanners to record arrival times, with delivery routes and on-time rates displayed in real-time on dashboards. Managers can see at any time "which vehicle is late, which station is congested." This step is not just about monitoring but also about data accumulation — several weeks of arrival time data can support quantitative decisions for route optimization and frequency adjustment.

The third step is to integrate the data chain between planning and logistics. When the production planning system is linked with the delivery system, any change in the plan automatically updates the delivery schedule and material demand, eliminating the disconnect where "the plan changes, but logistics is unaware." It is important to note that digitalization is an amplifier, not a replacement for Milk Run — if the process itself is not lean, the system will only accelerate the chaos. Running the cycle manually first and then considering digitalization is a more stable approach.

8. Implementation Path: From a Pilot Line to Full Plant Rollout

The rollout of Milk Run should avoid a "full-scale" approach. The correct method is to pilot first and then gradually expand.

The first step is to select a typical production line for piloting. The pilot line should meet three conditions: relatively stable material types, minimal production volume fluctuations, and adjustable line-side space. The pilot cycle generally lasts four to six weeks, with the first two weeks dedicated to route and schedule design, the middle two weeks for trial operation and data collection, and the last two weeks for standardization and effect evaluation.

The second step is to validate the effects with data. Compare key metrics before and after the pilot: on-time delivery rate, line-side inventory value, handling distance and time, and production line stoppages. Data is the only language that can convince management and relevant departments. If the pilot results are solid, the rollout will be more persuasive.

The third step is standardization and horizontal expansion. Compile the route map, schedule, station definitions, and container standards from the pilot line into standard documents, incorporating them into the site management baseline, and then gradually replicate them to other production lines. During expansion, note that "each line has its own solution" — material characteristics vary by line, so solutions cannot be directly copied, but the design methods and improvement strategies are consistent.

The fourth step is to ensure supporting mechanisms are in place. Milk Run will affect the work methods of multiple departments, including the warehouse, planning, and production. Establish cross-departmental coordination mechanisms, such as daily logistics meetings and anomaly escalation processes. Additionally, incorporate the on-time delivery rate into the performance indicators of relevant positions to ensure that improvement results are institutionally supported.

9. Conclusion

Looking back at the company's transformation journey: from frequent production line stoppages to an on-time delivery rate stabilized at over 97%; from line sides cluttered with materials and aisles blocked, to empty containers cleared and aisles clear by the end of the shift; from delivery personnel relying on experience to put out fires, to operating on a fixed schedule and route. The surface change is smoother logistics, but the fundamental change is the transition from "reactive response" to "proactive rhythm" — delivery is no longer a subordinate to production but a synchronized pulse.

The value of Milk Run lies not in the few routes drawn but in how it uses a fixed rhythm to link demand, handling, inventory, and recovery into a closed loop. When materials arrive at the line side like a bus at a stop, stoppages decrease, inventory drops, and the site becomes cleaner. More importantly, managers can free themselves from "putting out fires" and focus on value-creating improvements. This is the simplest yet most profound power of in-plant logistics Milk Run.

Delivery like a bus, on-time is the rhythm — making materials run like a shuttle bus.

Knowledge code: 7.4.2

Version: v20260804

Author: Quality Think Tank

Quality Think Tank is dedicated to providing systematic professional knowledge, methodologies, and practical tools for quality management practitioners, helping companies continuously improve their quality capabilities.