Start with an observation. According to supply chain reports published by multiple research firms in recent years, average inventory days for consumer goods companies have risen, not fallen, over the past five years. Demand volatility is intensifying, SKU counts are expanding, yet most companies still manage inventory with a single safety-stock formula. The formula hasn't changed; the world has. No wonder inventory keeps getting messier.
The root cause: many companies manage fundamentally different kinds of inventory as one pool. Safety stock held against demand volatility, strategic stock held for supply continuity, and cycle stock held for day-to-day turnover — three different decision logics, three different holding-cost profiles, three different optimization directions. Managed together, the result is predictable: what's overstocked stays overstocked, what's understocked stays understocked. Inventory stratification splits "a pile of stock" into three layers, each managed by its own logic.
Layer One: Cycle Stock — the Only Layer That Should Be "Calculated"
Cycle stock exists to meet the normal rhythm of sales, and its logic is the cleanest: the trade-off between economies of scale from batch purchasing or production and the cost of holding inventory. The economic order quantity (EOQ) model handles exactly this layer — inputs are demand rate, ordering cost, and holding cost rate; output is the optimal batch size.
The management imperative here is honest parameters. I've seen companies whose EOQ parameters hadn't been updated in three years: ordering cost still based on five-year-old freight quotes, holding cost rate set at 25% by gut feel, demand rate averaged from last year. Garbage parameters in, wrong batch sizes out. I recommend reviewing all three inputs every six months — especially demand rate. Once a SKU enters its decline phase, cycle stock must come down in step, or it becomes tomorrow's dead stock.
By the numbers, cycle stock typically represents 40%–60% of total inventory value — the largest layer, and the one with the most optimization headroom. Get this layer right and total inventory drops substantially.
Layer Two: Safety Stock — the Insurance Premium Paid for "Uncertainty"
Safety stock buffers against demand and supply variability. It is, in essence, insurance: every dollar of holding cost buys the outcome of "no stockout." And insurance should be sized like insurance — by the service-level target.
The standard formula: safety stock = Z × demand standard deviation × square root of replenishment lead time. Z maps to the service level (1.65 for 95%, 2.33 for 99%). The formula itself is fine. The problems are two common misuses.
Misuse one: one service level for all SKUs. Many companies set 95% across the board, with the result that long-tail SKUs consume a disproportionate share of safety stock. The correct approach is ABC-tiered service levels: 98%–99% for A items, 95% for B, 90% or lower for C. A stockout on a C item costs little to begin with — no reason to pay a fat premium insuring it.
Misuse two: using past volatility to predict future volatility. Standard deviation measures history, but promotions, launches, and seasonality make future volatility a different animal. My recommendation: for volatile SKUs, set safety stock as "base volatility + event uplift," and dial it back when the event ends. In cases I've tracked, safety stock levels that were "set and forgotten" averaged more than 30% above reasonable levels after two years.
Safety stock usually accounts for 20%–35% of total inventory. Its optimization direction isn't blind reduction — it's making every dollar count: insure high-value, high-volatility SKUs generously; insure low-value, stable SKUs lightly.
Layer Three: Strategic Stock — the Layer Least Suited to Formulas
Strategic stock exists for supply-chain disruptions, geopolitical risk, and critical-material shortages. In the years since 2020, this layer's prominence has exploded: chips, key raw materials, sole-sourced components — companies built strategic reserves across the board.
But strategic stock is the most poorly managed layer in many companies. Its decision logic shouldn't be a formula; it should be scenario analysis: if this material were cut off for three months, what would it cost? What does holding three months of usage cost? Compare the two, then decide whether to build the reserve and how large.
One manufacturing case I studied: a critical chip where one day of line-down losses ran about $500K, while holding 90 days of usage cost about $1.2M (including capital carrying and warehousing). The math said that as long as the disruption probability cleared a very low threshold, the strategic stock paid for itself. But the company's prior approach was "keep two months by gut feel" — no scenario analysis, no periodic review. When the 2021 chip shortage hit, two months of cover lasted six weeks.
Strategic stock typically runs 10%–25% of total inventory, but carries the highest holding-cost rate (long cycle, slow turn) — and must be reviewed regularly. Supply-chain risk is dynamic; when the risk recedes, strategic stock must come down, or it quietly morphs into dead stock.
Coordinating the Three Layers: A Decision Framework
Once the layers are managed separately, they need a coordination framework so they don't drift apart. I suggest this matrix for the annual review:
| Layer | Decision logic | Review cadence | Optimization direction |
|---|---|---|---|
| Cycle stock | EOQ: scale economies vs. holding cost | Semi-annual | Honest parameters, right batch sizes |
| Safety stock | Service level: insurance logic | Quarterly | Tiered coverage, event uplifts |
| Strategic stock | Scenario analysis: disruption loss vs. holding cost | Annual (immediate on risk change) | Reduce when risk recedes; don't feed dead stock |
The boundaries between layers deserve periodic scrutiny too. One trend worth watching: as supply-chain digitization improves forecast accuracy, the theoretical need for safety stock is declining — while geopolitical and climate risks are rising, pushing strategic-stock needs up. As one falls and the other rises, corporate inventory structures are shifting. Over the next five years, the best performers may not be the companies with the lowest total inventory, but the ones with the healthiest three-layer structure.
One open question to close: in your company, are these three layers accounted for separately? If the financials show a single "inventory" line, no layer can be optimized. The first step of stratified management is stratified accounting — make each layer visible before trying to manage it well.
By Dr. Chen, Supply Chain Researcher
