Safety Stock: A Distributor's Playbook

Safety stock is insurance, and like any insurance it can be under-bought or over-bought. A bearing distributor that holds too little stock misses orders and loses accounts; one that holds too much freezes capital in slow-moving shelves. The trick is not to carry more stock but to carry the right stock for the right segment at the right service level.

This playbook sets out how to size safety stock for a bearing range using segmentation, service-level arithmetic and the two kinds of variability that actually drive the number, then reduces it to a step-by-step procedure you can apply to a real catalogue.

Segment first: A, B and C

No single stocking rule suits a whole bearing range. The first move is to sort the range into three segments by turnover and by how standard the envelope is, then apply a different policy to each. A items are high-turnover, standard-bore parts that move predictably. B items are steady but narrower, with fewer customers and less regularity. C items are slow movers that are usually better ordered on demand unless a specific risk says otherwise.

The point of segmentation is that a single global service level wastes capital on C items and starves A items. Tying the stock policy to the segment lets the same budget produce far better availability on the parts that actually move.

Why the split matters

Segmentation lets a single budget serve three different risks at once. An A item is sized against turnover and replenishment frequency, a B item against steadier but thinner demand, and a C item is deliberately left un-stocked unless a specific risk justifies it. Without the split, a slow-moving special and a high-turnover mainstream bore compete for the same capital and the same shelf, and the mainstream part loses.

The service-level arithmetic in plain terms

Safety stock is the buffer that covers the gap between what you expect to sell during a replenishment cycle and what you actually sell. The size of that buffer depends on how confident you want to be that you will not run out, which is the service level, and on how variable demand and lead time are.

The Z factor

The service level enters through a single multiplier called the Z factor, which comes from the normal distribution. Its message is blunt: the higher the confidence, the more stock is needed, and the cost of the last few points is disproportionate.

Read the table carefully. Going from 90% to 99% confidence almost doubles the buffer, and that extra stock sits on a shelf for a long time. For most standard bearings a 95% service level is the sensible target, and the last points are reserved for parts whose absence genuinely stops a machine.

Using variability, not intuition

The most common sizing error is to plan around the average. A part that sells twenty pieces a month on average is not the same as a part that sells two hundred one month and none the next, even though both have the same average.

Demand variability and lead-time variability

Two spreads feed the safety stock calculation. Demand variability is the spread of how much you sell per period; lead-time variability is the spread of how long a replenishment actually takes. When lead time is reliable, only demand variability matters. When lead time swings - a production slot that moves, a batch that waits for the next heat - the lead-time spread enters the calculation as well, and it often dominates, because a late replenishment arrives after the stock has already run out.

This is also why a bearing that is fully interchangeable is easier to plan. The interchangeable number structure means one stocked designation covers several brands, so a single SKU absorbs demand that would otherwise be split across separate records and become artificially volatile. That structure rests on ISO 15 boundary dimensions, which is what makes two brands genuinely interchangeable on the envelope in the first place.

The lead-time components that actually move

Replenishment lead time is not one number. It is a chain of steps, and knowing which links move tells you where to hold buffer and where to negotiate instead. Ocean transit for a full container is typically 30–40 days on its own, which is why it is the component most worth planning around.

Two of these links - the production slot and the heat-treatment batch - are the ones that make lead time swing rather than simply run long. When both are tight, the lead-time spread is what forces the safety stock up, not the average lead time itself.

Cycle stock, safety stock and critical spares

Three different inventories are often lumped together under "stock", and each answers a different question. Cycle stock exists because ordering in economic lots is cheaper than ordering continuously. Safety stock exists to absorb variability. A strategic critical-spares list exists for a different reason entirely: a handful of bearings that would stop a plant if they were missing, held regardless of turnover.

Keeping the three separate stops the same budget from being spent twice. Cycle stock is driven by order economics, safety stock by service level and variability, and critical spares by the cost of a stoppage. A single-point-of-failure bearing on a machine with no spare can justify a critical-spares position even at zero turnover, because the penalty for a stockout is measured in downtime rather than in sales.

Critical spares are their own budget

A strategic critical-spares list is a third pot of money, sized by the cost of a stoppage rather than by a service level. It is small, it is reviewed separately, and it is the one place where a slow-moving part earns a permanent shelf position. The ratings used to qualify an acceptable substitute in that position follow ISO 281 for dynamic life and ISO 76 for static capacity, so the comparison is made on numbers rather than on brand alone.

When to stock a bearing you do not sell often

A slow mover is not automatically a "do not stock". A C item becomes worth holding when one of three conditions is true:

• Single point of failure. The bearing has no substitute in the machine and its absence stops the line.
• Long lead time. Replenishment is measured in weeks or months, so a stockout cannot be covered reactively.
• High downtime cost. The cost of a stoppage dwarfs the carrying cost of the part many times over.

When any of these applies, the part is a critical spare rather than a slow-moving item, and it is sized by downtime cost rather than by service level. The counterfeit and misrepresentation guide is worth reading before you qualify such a part, because a critical spare is exactly the position where a re-marked article does the most damage.

Review cadence

Stocking is a recurring decision, not a one-time one. Demand shifts, ranges grow and slow movers turn into obsolete part numbers on a shelf. A simple cadence keeps the plan honest: reclassify the ABC segments quarterly, run an obsolescence check on C items at the same time, and age the whole inventory at least once a year so that nothing sits undisturbed for a decade.

The catalogue data review shows how quickly a range grows, and our own deep groove range and cylindrical roller range illustrate how many designations a single category can carry, which is why a quarterly reclassification beats a fixed rule.

Step-by-step: sizing safety stock for a bearing range

1. Segment the range. Sort every designation into A, B or C by turnover and how standard the envelope is.
2. Set a service level per segment. Use 95% for A items, 90–95% for B items, and reserve higher levels for critical spares only.
3. Look up the Z factor. Read the multiplier for the chosen service level from the table above.
4. Measure demand variability. Use the spread of demand, not the average, for each segment.
5. Measure lead-time variability. Add the spread of the replenishment chain, remembering that ocean transit alone is typically 30–40 days.
6. Set the buffer. Combine the Z factor with the combined variability to size the safety stock, then round up to a sensible pack quantity.
7. Decide the order quantity. Choose a replenishment lot that balances order frequency against carrying cost, and keep cycle stock separate from the safety buffer.
8. Review quarterly. Reclassify segments, check C items for obsolescence, and re-age the inventory so that nothing lingers unnoticed.

Common mistakes

• Stocking on best-seller lists alone. A list of what sells is not a stocking plan; it ignores lead time, variability and the cost of a stockout.
• Applying one service level to the whole range. A single global setting starves A items and wastes capital on C items.
• Ignoring the cost of a stockout in a plant with no spare machine. Where a stoppage is measured in downtime, a critical spare is justified even at near-zero turnover.
• Never ageing the stock. Unchecked shelves hide obsolete part numbers and lubricant shelf-life limits, so a stock check should include an age sweep.
• Planning on the average. Average demand hides the spread that actually drives the buffer, and average lead time hides the swings that cause stockouts.

Safety stock is not a number to maximise or minimise; it is a choice about service level and risk. Segment the range, size each segment against its own variability and lead time, keep critical spares in their own category, and review on a fixed cadence. Done that way, the same working capital produces noticeably better availability. To see the full range this applies to, start from the bearing selector, compare house and named brands on the brand directory, and use the ANDGE range as a worked example of a broad house-brand catalogue. The distribution mechanics behind those listings are covered in the B2B e-commerce review, and our enquiry desk can help size a critical-spares list for a specific plant.