Holding cost: percent of unit cost vs flat $/unit/year
Use i (percent of unit cost) when capital cost dominates H. Use flat $/unit/year when the storage footprint dominates and unit value varies widely. Mixed SKU portfolios usually need both: i for high-value low-volume items, flat for low-value bulk goods[1].
Why the choice matters
Both forms feed Wilson Q* identically once you compute H. The difference is in how you defend the number to a CFO and how you scale across the SKU file. Percent form scales cleanly with unit cost; flat form scales cleanly with pallet count. Choose the form that matches the dominant cost driver.
Percent form: when it fits
- High-value goods where capital cost is the dominant component (luxury, electronics, pharma).
- Anything where insurance and shrinkage scale with value not volume.
- Multi-SKU portfolios with similar physical footprint per SKU.
Flat $/unit/year: when it fits
- 3PL pallet-storage pricing where you pay per pallet per month regardless of value.
- Cold storage where the BTU cost dominates the carrying-cost calc.
- Bulk staples (grain, salt, cement) where unit cost is low and storage cost dominates.
Mixed-portfolio blend
For mixed SKU files, run a hybrid. Compute H per SKU as max(i * C, F) where F is the floor implied by your storage cost per pallet divided by units per pallet. This avoids the failure mode where a cheap bulky SKU gets a tiny H because i * C is small but its physical footprint is enormous.