Facility location and layout

Where an operation is located and how its space is arranged shape its costs, its service and its ability to grow for years. A distribution centre placed a few hours closer to its main markets can save millions in transport costs and offer next-day delivery; a clinic placed on a bus route serves patients who do not drive; a factory layout that forces materials to criss-cross the floor wastes labour, damages parts and hides problems. Location and layout decisions are expensive to reverse, so they deserve careful analysis even though many of the factors involved are hard to quantify.

This lesson covers the factors that drive location decisions, three quantitative methods (factor rating, locational cost–volume analysis and the centre-of-gravity method), location decisions for services, the main types of layout, and assembly-line balancing. The crossover analysis from lesson 2 and the bottleneck concept from lesson 3 reappear here.

Location decisions and their drivers

Location decisions are usually made in stages: first the country or region, then the community, then the specific site. At the country level, firms weigh labour costs and productivity, market access, trade agreements and tariffs, exchange rates, political stability, regulation, taxes, infrastructure and culture. At the community level, they weigh the availability of skilled workers, wage rates, utilities, transport links, quality of life for employees, local attitudes and incentives. At the site level, they weigh land cost, size and zoning, access to highways, rail, ports or airports, environmental conditions, and room for future expansion.

Different operations weigh these factors differently. Manufacturers of heavy, low-value products such as cement or beverages locate close to markets because transport costs are high relative to value; processors of perishable or bulky raw materials, such as fish plants, pulp mills and mines, locate close to their inputs. Firms that depend on scarce skills locate in talent clusters: software in Toronto–Waterloo, Montréal, Vancouver, Seattle or Austin; aerospace in Montréal and Wichita; automotive along the Ontario–Michigan–Ohio corridor. Clusters offer specialized suppliers, shared knowledge and deep labour pools, which is why they persist even where costs are high.

Global location strategies have shifted in recent years. After decades of offshoring production to low-cost countries, many firms are reconsidering. Rising wages abroad, long and fragile supply chains exposed by the pandemic, shipping disruptions, tariffs and trade tensions, and concerns about forced labour and emissions have encouraged reshoring (bringing production home), nearshoring (moving it to nearby countries such as Mexico) and friend-shoring (moving it to politically allied countries). Lesson 12 discusses the supply chain risks behind these moves. The right answer depends on total landed cost and risk, not labour cost alone.

Factor rating

The factor-rating method introduced in lesson 2 is the most widely used location tool because it can combine quantitative and qualitative factors. The team lists relevant factors, assigns weights that sum to 1, scores each candidate location on each factor (for example from 1 to 100), multiplies scores by weights and sums them. Its value lies as much in the discussion it forces as in the final number. Sensitivity analysis is essential: if a small change in a weight reverses the ranking, the decision is a close call and deserves more investigation.

Locational cost–volume analysis

When the main differences between sites are in fixed and variable costs, the crossover method from lesson 2 applies directly. Fixed costs include rent or construction, property taxes and overheads; variable costs include labour, materials and transport per unit. Plotting total cost against volume for each site shows which is cheapest at each volume. For example, if site A has fixed costs of $300,000 a year and variable costs of $75 a unit, site B $600,000 and $45, and site C $1,100,000 and $25, then A is cheapest below 10,000 units, B between 10,000 and 25,000 units, and C above 25,000 units. At an expected volume of 20,000 units, B is the low-cost choice.

The centre-of-gravity method

When transport cost is the main concern, the centre-of-gravity method finds a location that minimizes the volume-weighted distance to the points it serves, such as the stores a distribution centre supplies. Each destination is placed on a coordinate grid (a map with x and y coordinates) and weighted by the volume shipped to it. The centre of gravity is the weighted average of the coordinates:

where \(d_{ix}\) and \(d_{iy}\) are the coordinates of destination \(i\) and \(V_i\) is the volume shipped there. The result is a starting point rather than a final answer: it uses straight-line distance, ignores the road network, and may land on a lake, a residential area or an expensive site. Analysts then examine nearby real sites with road and highway access, often using geographic information systems (GIS) and network optimization software that model actual travel times and costs.

Locating services

For services, location is primarily about revenue rather than cost, because customers must travel to the service or the service must travel to them. Retailers, restaurants, banks and clinics study the demographics, incomes and purchasing power of the surrounding area, traffic counts, visibility, parking, transit access and the presence of competitors and complementary businesses. Gravity models of retail, such as the Huff model, estimate the probability that a customer will shop at a store based on its size or attractiveness and the customer's travel time to it and to competing stores.

Public services face different goals. A fire station or ambulance base is located to keep response times below a target for as much of the population as possible; a library or community centre to maximize access, often with attention to equity for underserved neighbourhoods. Online services have changed the picture: many retailers now locate stores partly as showrooms and pickup points and locate fulfilment centres to cover the largest population within a one- or two-day delivery radius.

Types of layout

Layout is the physical arrangement of departments, workstations, equipment and storage within a facility. A good layout minimizes material and people movement, uses space efficiently, makes the flow of work visible, supports safety and communication, and is flexible enough to adapt as products and volumes change. The main types follow the process types from lesson 3.

In a process layout, the main design problem is to place departments so that those with heavy traffic between them are close together. A load–distance score multiplies the number of loads moved between each pair of departments by the distance between them and sums the products; alternative layouts are compared by their scores, and analysts swap department locations to reduce the total. Hospitals, for example, place emergency departments next to imaging and close to operating rooms. Cellular layouts, a feature of lean operations (lesson 11), regroup a process layout into cells, often U-shaped, that make families of similar parts; they cut travel, waiting and work in process and let a small team own a product from start to finish.

Office and service layouts increasingly balance collaboration and focus. Open-plan offices encourage interaction but can reduce concentration; many organizations now use a mix of open areas, quiet rooms and bookable spaces, especially with hybrid work. Warehouse layouts use the principle of placing the fastest-moving items closest to the dock, a version of ABC analysis (lesson 9), and many now include automated storage and robotic picking, discussed in ROB-440.

Line balancing

In a product layout, the central problem is assembly-line balancing: assigning tasks to workstations so that each station's work fits within the required cycle time and idle time is minimized, while respecting precedence (some tasks must come before others). The required cycle time is set by demand and is called takt time, from the German word for the beat of music:

The theoretical minimum number of stations is the total work content divided by the takt time, rounded up. Because tasks cannot be split and precedence must be respected, the actual number is often higher. Heuristic rules assign tasks one station at a time: among tasks whose predecessors are complete and that fit in the time remaining at the station, choose the one with the longest task time (or the most followers). Efficiency is the total work content divided by the number of stations multiplied by the cycle time; balance delay is one minus efficiency.

# Line balancing with the longest-task-time rule
tasks = {  # task: (seconds, predecessors)
    "A": (40, []), "B": (30, ["A"]), "C": (50, ["A"]), "D": (40, ["B"]),
    "E": (20, ["C"]), "F": (30, ["D", "E"]), "G": (25, ["F"]), "H": (35, ["G"]),
}
available = 7.5 * 3600          # seconds of production time per shift
demand = 480                     # units per shift
takt = available / demand
total = sum(t for t, _ in tasks.values())
print(f"Takt time: {takt:.2f} s; total work content: {total} s")
print(f"Theoretical minimum stations: {total/takt:.2f} -> {-(-total//takt):.0f}")
done, stations = set(), []
while len(done) < len(tasks):
    left, station = takt, []
    while True:
        ready = [k for k, (t, p) in tasks.items()
                 if k not in done and all(x in done for x in p) and t <= left]
        if not ready:
            break
        k = max(ready, key=lambda x: tasks[x][0])   # longest task first
        station.append(k); done.add(k); left -= tasks[k][0]
    stations.append((station, takt - left))
for i, (s, load) in enumerate(stations, 1):
    print(f"Station {i}: {', '.join(s):<8} load {load:3.0f} s  idle {takt-load:4.1f} s")
n = len(stations)
print(f"Stations used: {n}; efficiency: {total/(n*takt):.1%}; balance delay: {1-total/(n*takt):.1%}")

Takt time: 56.25 s; total work content: 270 s
Theoretical minimum stations: 4.80 -> 5
Station 1: A        load  40 s  idle 16.2 s
Station 2: C        load  50 s  idle  6.2 s
Station 3: B, E     load  50 s  idle  6.2 s
Station 4: D        load  40 s  idle 16.2 s
Station 5: F, G     load  55 s  idle  1.2 s
Station 6: H        load  35 s  idle 21.2 s
Stations used: 6; efficiency: 80.0%; balance delay: 20.0%

The line must produce 480 units in 7.5 hours of production time, so takt time is 56.25 seconds. The work content of 270 seconds implies a theoretical minimum of 5 stations, but the heuristic needs 6, giving 80 % efficiency. In this case 6 is in fact the best possible: task A (40 seconds) must come first and no other task can join it without exceeding takt time, leaving 16 seconds idle at station 1, which already exceeds the total slack of 11.25 seconds that a 5-station line could afford. To reach 5 stations, the firm would need to redesign tasks (for example, splitting C or combining A with part of B), speed up the slowest tasks, or run a longer shift.