Stop stocking zero on critical parts

See criticality-based safety stock computed on your own materials.

Talk to an MRO expert →

Short answer: Industry estimates suggest the average asset-intensive manufacturer carries 20 to 30 percent excess MRO inventory and simultaneously faces stockout risk on 10 to 15 percent of critical parts, consistent with Verusen's experience across hundreds of implementations. Standard safety stock formulas require demand history, but most maintenance parts fail too infrequently to generate usable data; a bearing that fails twice in five years returns zero history, triggering zero stock and eventual line stops. Instead, calculate MRO safety stock using criticality (impact of stockout) and lead time, not demand forecasts, because spare parts don't sell on a schedule, they fail.

MRO safety stock: The minimum inventory of a maintenance spare part needed to buffer against the random timing of equipment failure and supplier lead time, sized by criticality and failure consequence rather than demand probability.

Why standard safety-stock formulas fail for MRO

Standard spare parts safety stock calculation formulas assume demand history exists; for a bearing that fails twice in five years, the formula returns zero stock, forcing manual override based on intuition instead of data. The root cause is a category error: ERP systems apply demand-forecasting logic to failure events, but a bearing that fails twice in five years is not 'demanded' the way a consumer good sells on a predictable curve.

Demand forecasting requires a distribution; failure is a point event. You cannot build a forecast from two data points separated by years. SAP, Oracle, and most ERPs ship with identical safety stock definition and reorder point methodology for both finished goods and spare parts, so when a spare-parts part number has sparse failure history, the system recommends zero stock. A leading gold mining company with 17 sites and three ERP systems identified $96.8M in MRO inventory excess only after applying criticality-driven logic instead of demand forecasting (based on Verusen customer results).

Replace demand history with a criticality-weighted calculation

Instead of waiting for failure history to accumulate, calculate a criticality score that weights the cost of downtime and sourcing lead time against annual holding cost. This score tells you which parts deserve safety stock regardless of how often they've failed.

1. List the part number, annual holding cost (storage, obsolescence, capital), hours required to source and install it, and the production line or equipment it supports.
2. Estimate the cost per hour of downtime if that equipment fails (revenue loss, labor delays, cascading stops across dependent lines).
3. Calculate: Criticality Score = (Downtime Cost per Hour × Hours to Source and Install) / Annual Part Holding Cost.
4. Normalize the result to a 1 to 10 scale, where 10 represents a part that will stop production immediately and cannot be sourced locally within 24 hours.
5. Enter the criticality score into your ERP or EAM reorder point field and set safety stock proportional to the score (high criticality: 2 to 4 units; medium: 1 unit; low: 0).

Real-world outcome. Georgia Pacific, a major pulp and paper operator with 110 US sites and roughly $1 billion in maintenance, repair and operations (MRO) inventory across four ERP systems, identified 2,900 materials at stockout risk using demand forecasting alone. Once it applied criticality-weighted logic, it centralized stocking decisions from hundreds of plant operators to a team of seven people and recovered 6,600 hours (based on Verusen customer results).

Criticality-weighted safety stock

Criticality-weighted safety stock assigns buffer sizing by operational consequence first, demand history second, solving the core failure of standard formulas: a bearing that fails twice in five years has no usable history, so the formula returns zero stock, the bearing fails, and the line stops for three weeks. A Fortune 100 pharma and bioscience company with 33 sites identified $26.5M in excess inventory by switching from demand-history formulas to criticality scoring, based on Verusen customer results.

How to implement criticality-driven MRO safety stock

1. List all parts in your high-consequence asset categories (compressors, bearing assemblies, control systems, hydraulic valves, etc.).
2. For each part, estimate the hourly production loss in dollars if that part fails, multiplied by the hours required to source and install a replacement from your preferred supplier.
3. Divide that consequence value by the annual holding cost for that part (use your finance team's carrying-cost rate — storage, capital, and obsolescence).
4. Normalize all criticality scores to a 1-to-10 scale so reviewers can rank parts by operational consequence, not demand frequency.
5. Sort parts into three tiers: Tier 1 (scores 8-10), Tier 2 (scores 5-7), Tier 3 (scores 1-4).
6. For each Tier 1 part, set reorder point in your ERP/EAM to: (average weekly demand × lead time in weeks) + safety stock buffer. For Tier 1 parts with short lead times, buffer with up to one full cycle of lead-time demand; Tier 2 parts take a partial buffer; Tier 3 parts use economic order quantity (EOQ) alone.
7. Enter the reorder point and safety stock values into your ERP/EAM reorder point field and activate automatic flagging when stock falls below threshold.

This is a one-time conversation with your maintenance team, not a data science project. Review classifications quarterly with operations; update if failure consequence changes (for example, a new upstream process now depends on a previously low-impact part).

Where AI improves the estimate

Manual methods often classify parts as overstock when they have failed only once or twice in years of operation, causing planners to cut safety stock below safe levels. A leading global gold mining company with 17 sites carried parts classified as overstock by traditional demand history; when AI recalculated by consequence instead, it uncovered $96.8M in true excess inventory across the portfolio, based on Verusen customer results. The difference: machines fail on schedules humans do not predict, but the consequence of failure is calculable and does not change. Criticality-driven spare parts management shifts effort from frequency-based formulas to consequence-based classification, the fastest way to surface the parts that actually protect uptime.

A Fortune 500 CPG manufacturer with 41 sites applied this framework and reduced material review time from over 20 minutes to 4 minutes per part, based on Verusen customer results. The speed came from pre-ranked criticality: reviewers spent effort only on Tier 1 and Tier 2 items; Tier 3 parts were managed by rule. When a part rarely fails, demand history says order nothing; criticality says order based on what happens if it does fail.

A worked calculation

For intermittent-demand MRO parts, the criticality-driven safety stock formula replaces demand history with downtime consequence and lead time, because standard formulas return zero when a part fails only twice in five years. A maintenance manager at a multi-site food producer managing tens of thousands of MRO parts across three ERPs faces this exact constraint: the historical demand model has no signal. Criticality scoring solves it by asking 'What stops if it's not available?' and 'How long until a replacement arrives?' instead of 'How often did it fail last year?'

MRO safety stock formula: demand-history vs. criticality-driven

Lead time is days from order placement to on-hand receipt, verified in your ERP or recent purchase orders. Criticality is 1.0 if failure stops the production line, 0.5 if it degrades performance without a full stoppage, and 0.1 for inconvenient but non-critical failures. The +1 unit buffer accounts for simultaneous in-service degradation and replacement delay, a real scenario on offshore rigs and multi-shift plants.

Real outcome. At Seadrill, a global offshore operator managing 17 rigs with Maximo, this criticality matrix replaced manual stocking decisions across a hub-and-spoke shorebase-to-rig supply model, enabling real-time stockout-risk flagging and reducing emergency offshore freight costs (based on Verusen customer results). Industry estimates suggest the small share of parts that are truly critical carries the majority of downtime cost, consistent with Verusen's experience across hundreds of implementations; concentrating criticality scoring on those high-impact items first yields faster uptime gains than optimizing the full spare-parts catalog.

Check the formula on your worst part

Bring one line-stopper; we'll walk the calculation live.

Talk to an MRO expert →

Setting review cycles

Tie review cycles to failure consequence and lead time, not a calendar: a major US energy company with 45,000 MRO materials achieved 100% audit capability for FERC compliance while identifying $40M in excess inventory by shifting from calendar-based reviews to criticality-driven triggers (based on Verusen customer results). Most plants inherit annual or semi-annual review schedules from paper-era practices. For intermittent-demand parts, yearly reviews trap inventory: a bearing that fails twice in five years sits at zero stock for 18 months, then the week you skip your review cycle, it fails and production stops for three weeks.

Criticality score and tier assignment

Calculate a criticality score for each part using this formula: Criticality Score = (Downtime Cost Per Hour × Hours to Source and Install) / Annual Part Holding Cost. Normalize the result to a 1-10 scale. Parts scoring 8-10 are Tier 1 (review every 30 days); 4-7 are Tier 2 (review when consumption threshold is hit); 1-3 are Tier 3 (deactivate if zero demand and zero stock). A Fortune 100 pharma and bioscience company with 33 sites identified $26.5M in excess inventory by switching from demand-history formulas (which return zero for intermittent parts) to criticality scoring (based on Verusen customer results).

Implementation workflow

1. Calculate downtime cost per hour (production revenue lost or customer penalty per hour of unplanned downtime).
2. Estimate hours to source and install for each part (supplier lead time plus on-site installation time).
3. Divide by annual holding cost (part cost × storage, capital, and obsolescence rate).
4. Map the result to a 1-10 scale and assign tier.
5. Enter the tier and review frequency into your ERP or EAM reorder-point field or maintenance dashboard as an automated trigger rule.

Most ERP systems fail at MRO safety stock review because they apply the same schedule to all parts regardless of failure consequence. Once you encode criticality into your review logic, line-stopping parts stay reviewed and low-consequence parts stop consuming review capacity, freeing your team to focus on inventory that actually affects production.

Where AI improves the estimate

Manual methods flag parts as overstock when they fail once every three years; AI recalculates by consequence and uncovers the true excess. A leading gold mining company operating 17 sites across three ERP systems identified $96.8M in MRO inventory savings by scoring each part on failure consequence, time-to-replacement, and substitutability instead of relying on demand variance alone (based on Verusen customer results).

Why demand-only formulas fail on critical, infrequent failures

Standard safety stock formulas measure variance around historical demand. A bearing that fails once every three years has almost no variance. The formula returns zero safety stock because there is no pattern to measure. When the bearing fails and inventory is empty, the production line stops for the time it takes to source and install a replacement. A Fortune 100 pharma and bioscience company with 33 sites identified $26.5M in excess inventory by discovering that traditional demand-history methods had marked critical, low-frequency parts as overstock (based on Verusen customer results).

How to implement criticality-driven MRO safety stock

Replace demand variance with a consequence-based decision rule. Calculate criticality as the cost of being wrong: downtime cost per hour multiplied by hours to source and install the part, divided by your annual part holding cost. Normalize the result to a 1 to 10 scale.

1. Pull failure history, downtime duration in hours, and lead times from your ERP or EAM without requiring a data cleanse first.
2. For each high-consequence part, multiply downtime cost per hour by the hours required to source and install; divide by annual holding cost to calculate criticality.
3. Score the result on a 1 to 10 scale: a part with high downtime impact and long lead time scores 8 to 10 and requires safety stock; a part with low downtime impact scores 1 to 3 and triggers zero stock, freeing capital for parts that actually protect production.
4. Enter the criticality score and target safety stock quantity into your ERP or EAM reorder point field; tie replenishment logic to consequence and risk, not the flawed assumption that infrequency equals low value.

Implement this workflow using an MRO inventory optimization strategy that automates criticality scoring across multiple sites and ERPs. The result is stocking policies tied to what actually matters: consequence and risk.

Go deeper: this article supports our pillar guide, MRO Inventory Optimization: The Complete Guide. Related: spare parts inventory management by criticality.

Further reading: safety stock formula methods, Institute for Supply Management, and spare parts inventory management guide.

Roll it out across sites

Scope network-wide safety stock policies with an expert.

Talk to an MRO expert →

Frequently asked questions