Safety stock (also safety inventory) is the buffer held above cycle stock to absorb demand or replenishment lead-time variability — including freight — and protect service level against stockouts.
If your answer to every delay is “raise the minimum,” you are funding transport uncertainty with sleeping capital. This guide locks definition, a lead-time-aware formula, a peso example, and the link to on-time delivery (OTD) for Mexico–U.S. shippers · not an ecommerce fulfillment tutorial.
- above cycle stock
- Buffer
- freight lead time counts
- σLT
- protects the promise · not the spreadsheet
- OTD
- MX–US shipper · not ecommerce-only
- CEDIS
Cluster: inventory · stock · OTD · distribution costs · logistics chain.
What safety stock / safety inventory is
In a distribution center (CEDIS), safety stock is the extra quantity planning leaves “just in case” — but with a written policy: which SKU it applies to, with which data, which service level it protects, and who reviews it. Without that, it is not safety stock: it is unaudited “just in case” inventory.
In Mexico both names show up in search and on the floor: inventario de seguridad (more planning/accounting) and stock de seguridad (more dock/WMS). Same cushion. Existences context: what stock is · asset frame: what inventory is.
Safety vs cycle stock
Mixing them into one “minimum” is the expensive mistake: you cannot tell whether you are covering the normal cycle or paying for freight uncertainty.
Cycle stock
What it covers: Expected demand between replenishments
Warning signal: Depletes “on schedule” and returns with the order
Typical owner: Planning / purchasing
Safety stock
What it covers: The unexpected (demand or lead time)
Warning signal: Only touched on spikes or delays
Typical owner: Planning + traffic (LT data)
Overstock / obsolete
What it covers: Nothing useful: capital or expiry
Warning signal: Does not turn · no review date
Typical owner: Finance + operations
Buffer · two layers
Cycle vs safety
If you merge them into one minimum, you do not know which uncertainty you are paying for.
Cycle stock
Expected demand
Between replenishments · depletes “on time” and returns with the order.
Safety stock
The unexpected
Demand or freight lead time · only touched in peaks or delays.
Alert
Without a review date, “just in case” is idle capital dressed as prudence.
How to calculate (with lead time)
Many ecommerce guides assume stable lead time. On Mexico–U.S. lanes, freight lead time (appointments, crossing, delays) is often as noisy as demand. Use a heuristic that includes both sources · and calibrate with your operation.
- Z
- Service-level factor (e.g. ~1.65 ≈ 95% cycle service).
- LT
- Average door-to-door lead time (days), including freight.
- σd
- Std. deviation of daily demand (units).
- D
- Average daily demand (units).
- σLT
- Std. deviation of lead time (days).
Compare with the thin form common in ecommerce tutorials:
Thin / ecommerce formula
Formula: SS ≈ Z × σd × √LT
Assumption: Nearly fixed lead time (σLT ≈ 0)
Risk on MX–US: Understates buffer when appointments/border move days
Formula with σLT (MX–US corridor)
Formula: SS ≈ Z × √(LT × σd² + D² × σLT²)
Assumption: Demand and lead time both vary
Risk on MX–US: σLT × D² often dominates · measure real freight
A typical reorder point (ROP) adds expected demand during lead time plus the buffer: ROP ≈ D × LT + SS. Without SKU/ABC policy, one ROP for the whole catalog lies.
Worked Mexico–U.S. example
Finished-goods SKU A at a Nuevo León CEDIS · average demand 40 cases/day · demand std. dev. 10 · door-to-door lead time 4 days (supplier + freight) · lead-time std. dev. 1.2 days (appointments + crossing swing) · Z = 1.65 · unit cost $220 MXN. Illustrative numbers · calibrate with your history.
Demand during LT
Calculation: 40 × 4
Result: 160 cases
Thin / ecommerce SS
Calculation: 1.65 × 10 × √4 = 1.65 × 20
Result: ~33 cases · ~$7,260 MXN
Combined variance
Calculation: 4 × 10² + 40² × 1.2² = 400 + 2,304
Result: 2,704
SS with σLT
Calculation: 1.65 × √2,704 = 1.65 × 52
Result: ~86 cases · ~$18,920 MXN
Gap
Calculation: 86 − 33
Result: ~53 cases · ~$11,660 MXN of capital the thin formula misses
If lead-time swing falls from 1.2 to 0.6 days (better appointments, fewer no-shows, timely POD), the D² × σLT² term halves and required buffer drops without blindly squeezing service.
What inflates the buffer on the corridor
In nearshoring, lead time is not only “factory days”: it includes dock, CEDIS appointment, crossing, transload, and delays. Raising pallets without attacking these causes is the most expensive way to buy calm.
Demand variability
How it inflates SS: Raises σd · more buffer on A SKUs
Fix before buying more: Forecast + ABC · not one % for everything
Appointments / dock
How it inflates SS: Raises σLT (day swings)
Fix before buying more: Written window · less WhatsApp as the clock
Border / crossing
How it inflates SS: Unpredictable door-to-door lead time
Fix before buying more: Measure GPS/POD · not catalog lead time
No-show / spot freight
How it inflates SS: Delays and emergencies
Fix before buying more: Assignment and evidence in the trip file
Late POD
How it inflates SS: Planning thinks it “has not arrived”
Fix before buying more: POD bound to the trip ID
OTD vs sleeping capital
Well-calibrated safety stock protects the OTD promise and completeness (fill rate / OTIF). Excess does not “improve service”: it occupies locations, freezes cash, and feeds distribution cost.
Low SS + noisy freight LT
What happens to service: Stockouts · broken OTD/OTIF · rush jobs
What happens to the P&L: Expensive spot freight · lost sales
High SS without attacking σLT
What happens to service: Fewer “inventory” stockouts
What happens to the P&L: Sleeping capital · occupancy · obsolescence
Calibrated SS + measurable LT
What happens to service: Defendable service by SKU
What happens to the P&L: Fewer rushes · fewer eternal minimums
How to calibrate the buffer
Practical order: measure real lead time first, then design the cushion. The other way around is buying uncertainty.
Policy cadence
Lead time, buffer, and review
Measure lead time
Real door to door
Classify SKUs
ABC + criticality
Set service
Fill rate / OTIF
Compute buffer
Formula or days
Review cadence
Owner and date
Operable policy checklist
Use it in S&OP or planning before raising another “because of the border” minimum.
Elige un paso para ver el detalle
Detalle del paso · 01
Separate cycle from safety in the WMS or policy
OCL and freight lead time
OCL Cargo is an autonomous transportation management system (TMS) with computer-use agents: it does not compete with your WMS for rack balances. It does not compute your EOQ or safety stock. It makes the transport leg measurable — tracking, proof of delivery (POD), appointments, and exceptions — so planning stops buying uncertainty blind.
Who owns exceptions? Your team. When applicable, OCL can stamp invoices and Carta Porte. Typical pilot 6–8 weeks; recovery pattern 5–7% when auditing 100% of the pilot flow · price signal from ~$50 MXN per shipment depending on scope.
Key takeaways5 points
- Safety stock / safety inventory = buffer above cycle stock for demand and lead-time uncertainty (including freight).
- Thin / ecommerce formulas (Z × σd × √LT) assume stable lead time · on MX–US lanes σLT is often missing.
- Usable heuristic: SS ≈ Z × √(LT × σd² + D² × σLT²) · calibrate with your operation.
- The buffer protects OTD / fill rate · excess burns cash and distribution cost.
- OCL does not compute EOQ or replace your WMS: it makes trip lead time measurable. Who owns exceptions? Your team.
Is your buffer funding freight uncertainty?
Related reading
Frequently asked questions
It is the buffer inventory held above cycle stock to absorb demand or replenishment lead-time variability — including freight — and protect service level (fill rate, OTIF, OTD). In Spanish it is also called inventario de seguridad or stock de seguridad.
Cycle stock covers expected demand between replenishments. Safety stock covers the unexpected: demand spikes, late suppliers, slipped appointments, or unpredictable border crossings. See also what stock is and inventory.
A usable form when both demand and lead time are uncertain is SS ≈ Z × √(LT × σd² + D² × σLT²). Z is the service-level factor; LT average lead time; σd daily demand std. dev.; D average daily demand; σLT lead-time std. dev. Calibrate with your operation — it is a heuristic, not a dissertation.
It often understates the buffer when appointments, border dwell, or carriers add day swings. The thin form assumes stable lead time (σLT ≈ 0). On the corridor, measure door-to-door freight lead time before “raising the minimum.”
Only if stockouts came from real demand or replenishment variability. If the problem is picking, appointments, or the carrier, the buffer hides the symptom, freezes cash, and inflates distribution cost.
Demand (mean and variability), real door-to-door lead time (purchase + freight + dock), target service level, and ABC by SKU. Without measured freight lead time, the formula lies.
It is not a warehouse management system (WMS): it does not compute your economic order quantity (EOQ) or buffer. It helps make freight lead time measurable (tracking, POD, exceptions) so planning does not buy uncertainty blind. Who owns exceptions? Your team. When applicable, OCL can stamp invoices and Carta Porte.