The Adjacency Intuition in a Distribution Center Layout
A regional distribution center moves 4,000 orders a day through four areas strung along one main aisle: Receiving (R), Reserve storage (S), Pick-and-pack (P), and Dispatch (D). The industrial engineer's from-to chart, in pallet-loads per day, reads: R→S 120, S→P 100, P→D 90, R→P 30, S→D 20, P→S 15 (returned totes). One-way travel along the aisle is 40 m from R to S, 30 m from S to P, and 25 m from P to D. Aisle travel is costed at $0.80 per load-meter. The DC manager wants Pick-and-pack moved next to Receiving on the grounds that 'receiving feeds picking,' which would give the sequence R, P, S, D with 30 m R–P, 30 m P–S, and 25 m S–D. No flow volumes are expected to change.
1. Compute the daily flow cost of the current layout and of the manager's proposed layout.
Read the worked answer
Current layout (R–S–P–D) distances
Adjacent: R–S 40, S–P 30, P–D 25. Non-adjacent: R–P 70, S–D 55, R–D 95.
| Flow | Loads | Distance (m) | Load-m |
|---|---|---|---|
| R→S | 120 | 40 | 4,800 |
| S→P | 100 | 30 | 3,000 |
| P→D | 90 | 25 | 2,250 |
| R→P | 30 | 70 | 2,100 |
| S→D | 20 | 55 | 1,100 |
| P→S | 15 | 30 | 450 |
| Total | 13,700 |
Cost = 13,700 x $0.80 = $10,960/day.
Proposed layout (R–P–S–D) distances
Adjacent: R–P 30, P–S 30, S–D 25. Non-adjacent: R–S 60, P–D 55, R–D 85.
| Flow | Loads | Distance (m) | Load-m |
|---|---|---|---|
| R→S | 120 | 60 | 7,200 |
| S→P | 100 | 30 | 3,000 |
| P→D | 90 | 55 | 4,950 |
| R→P | 30 | 30 | 900 |
| S→D | 20 | 25 | 500 |
| P→S | 15 | 30 | 450 |
| Total | 17,000 |
Cost = 17,000 x $0.80 = $13,600/day.
The proposal is $2,640/day worse (+24%), or about $660,000 per year on a 250-day basis - for a change that was justified by intuition alone.
2. Explain why the intuition failed and state what you would recommend to the DC manager.
Read the worked answer
Why it failed: the manager reasoned from a single flow (R→P, 30 loads) and ignored the dominant path. The material in this building moves as a chain: R→S 120, S→P 100, P→D 90. That is 310 of the 375 loads/day. The proposal buys a 40 m saving on 30 loads (1,200 load-m) and pays 20 m on 120 loads (2,400 load-m, R→S now crosses P) plus 30 m on 90 loads (2,700 load-m, P→D now crosses S). The tail wags the dog.
Recommendation:
- Reject the swap as proposed. Keep the process sequence R–S–P–D; the current layout already places the three highest-volume flows on adjacent pairs, which is the correct structural answer.
- Reframe the question. The right question is not 'which areas are related?' but 'which flows dominate the cost?' In this DC the answer is the main chain, so the remaining improvement opportunity is not sequence but distance - shorten the aisle, or remove the R→S move entirely by cross-docking.
- Attack the 120-load R→S flow. If even a third of receiving volume were cross-docked or placed in a forward pick face adjacent to P, the largest single cost line falls. That saves more than any permutation of the four areas.
- Check non-flow factors before finalizing. Reserve storage usually needs the tallest clear height and the fewest columns; dispatch needs dock doors and trailer apron. If S physically cannot go anywhere but the middle bay, the flow calculation is confirming a constraint rather than choosing a layout - say so explicitly so the study is not re-litigated next year.
- Treat the 15-load P→S returns stream separately. It is 4% of volume; routing returned totes back along a dedicated path rather than the main aisle is cheaper than distorting the layout around it.
3. How large would the R→P flow have to become before the manager's swap actually pays off?
Read the worked answer
Hold every other flow constant and let x = R→P loads/day.
- Current layout: R→P travels 70 m, contributing 70x load-m.
- Proposed layout: R→P travels 30 m, contributing 30x load-m.
- Saving from the swap = 40x load-m/day.
- Fixed penalty from the other flows = 17,000 - 900 (the R→P term in the proposed layout) minus (13,700 - 2,100) = 16,100 - 11,600 = 3,300 load-m/day.
Break-even: 40x = 3,300 -> x = 82.5 loads/day.
So the swap only wins once R→P exceeds roughly 83 loads/day - nearly triple today's 30. That is a useful thing to tell the manager: the idea is not wrong in principle, it is wrong at these volumes, and the threshold is far enough away that this is not a close call.
The threshold is also worth recording in the layout file. If a future operating change - say a new customer segment that requires receiving inspection before putaway, or a kitting operation fed directly from receiving - pushes R→P past 83 loads/day, the manager's intuition becomes correct and the layout should be revisited. Also recheck the cost rate: at $0.80/load-meter the absolute numbers are sensitive to labor rate and to whether the 40 m is a real travel distance or a centroid-to-centroid estimate. The ranking of layouts is robust to that error; the dollar figure is not.