Consolidating two DCs into one: the freight line the business case forgot
Northbridge Building Products distributes 12,000 SKUs to 640 independent contractors from two leased DCs (Atlanta, 180,000 sq ft; Birmingham, 110,000 sq ft). Next-day delivery is promised on 96% of order lines; actual on-time-in-full is 88% and fill rate from stock is 91%. A consultant proposes consolidating into one 260,000 sq ft DC near the interstate junction, which modelling shows would cut inventory 18% and warehousing cost 9% through cube and labour efficiency, while raising average distance-to-customer from 140 to 290 miles. Freight is currently 6.4% of revenue and would rise to an estimated 8.1%. Revenue is $310M, gross margin 24%, inventory $48M at a 21% carrying cost. Warehousing and handling runs at 2.9% of revenue. The largest 20 customers generate 61% of revenue and buy in emergency, job-lot quantities. The CFO wants the savings; the sales director says service will collapse.
1. Quantify the annual cost impact of the consolidation on a like-for-like basis. On these numbers, is the CFO right?
Read the worked answer
Put all three lines on the same annual footing.
- Inventory: $48M x 18% = $8.64M of working capital released. At a 21% carrying rate that is +$1.81M/year of avoided cost (plus a one-off $8.64M cash release).
- Warehousing: 2.9% of $310M = $8.99M base; a 9% cut is +$0.81M/year.
- Freight: 6.4% of $310M = $19.84M today; 8.1% = $25.11M. That is -$5.27M/year, a 27% increase in the freight bill.
Net = 1.81 + 0.81 - 5.27 = -$2.65M per year. Consolidation destroys roughly $2.6M of value annually, so on the stated numbers the CFO is looking at two savings lines and ignoring the third.
Sensitivity, which is what a good answer adds: the decision only flips if the freight penalty is under about 0.84 percentage points of revenue, or if the inventory reduction is closer to 35% than 18%. Neither is credible with 290 average miles and a next-day promise to emergency buyers. Ask for the freight model by lane and by customer before rejecting the idea outright, because the inventory gain is real and may be capturable in a different network design.
2. What does the arithmetic miss, and what network design would you actually recommend?
Read the worked answer
The arithmetic treats service as a constant. It is not.
- The promise is already broken. OTIF is 88% against a 96% promise and stock fill is 91%. A consolidation that adds 150 average miles to a next-day commitment will make an existing failure worse, and the 20 customers worth 61% of revenue are exactly the ones buying emergency job lots - they are paying for responsiveness, not for the lowest cost-to-serve.
- Margins are thin. At 24% gross margin, losing a handful of those accounts costs more than the warehousing saving.
Recommended design instead of the binary choice: a two-echelon network. Keep a slimmed Atlanta DC as the stocking point for the top 20% of SKUs (which normally cover 80-90% of order lines) and for the top-20 accounts; convert Birmingham into a cross-dock or forward-stocking point replenished from Atlanta nightly; hold slow movers centrally. Model this - it typically captures 50-70% of the inventory benefit at a fraction of the freight penalty.
Separately, fix the 91% fill rate as a forecasting and replenishment-policy problem rather than a node-count problem, and check whether the 96% next-day promise is even the right promise for the job-lot segment. Also cost the transition honestly: dual running for 9-12 months, lease exit, and the inventory build before the cutover.