How Much Safety Stock Should You Hold? (2026 Formula + Tariff Guide)

What safety stock actually is

Safety stock is the extra inventory you hold to protect against surprises. It sits below your normal working stock and you only dip into it when something goes wrong: a demand spike, a late shipment, or a supplier who suddenly needs four weeks instead of two.

There are only two reasons you ever run out: demand was higher than expected, or your resupply took longer than expected. Safety stock is the buffer that covers both. Get the amount right and you avoid stockouts without drowning in cash you cannot get back.

Why 2026 changed the answer

For most of the last decade the smart move was lean: hold as little as possible, reorder often, keep cash free. In 2026 that advice flipped for a lot of small businesses.

Tariff volatility is the reason. 96% of small and mid-size businesses say tariffs have hurt their sourcing or supply chain in the past year, and the response has been a shift from "just-in-time" to what buyers now call "just-in-case" inventory: holding more stock as a hedge against price jumps and delays. Nearly three-quarters of SMBs have stretched their planning horizon from a 30 to 60 day window out to 90 to 120 days.

There is a second, quieter problem. The historical sales data you would normally use to forecast has become less reliable, because buying behaviour lurched during the uncertainty (customers pulled purchases forward, then went quiet). When your averages are shaky, your buffer has to do more work.

The takeaway: if you set your safety stock levels two years ago and never revisited them, they are almost certainly too low for how your suppliers behave today.

How much safety stock should you hold? Four methods

There is no single number. The right amount depends on how much your demand swings, how much your supplier's lead time swings, and how badly a stockout hurts. Here are four methods, from crudest to most precise. Pick the one that matches the data you actually have.

1. The simple days-of-buffer method

Multiply your average daily usage by the number of days of buffer you want to keep.

Safety stock = average daily usage x days of buffer

If you sell 20 units a day and want a 5-day cushion, that is 100 units. Quick, rough, and fine for slow, steady items. It ignores how much your supplier varies, so it is the weakest method for 2026 conditions.

2. The average-max method (best starting point for small businesses)

This is the most popular method for small businesses because it uses numbers you can pull from your own records without any statistics.

Safety stock = (max daily usage x max lead time) − (average daily usage x average lead time)

Say your average usage is 20 units a day over a 10-day lead time, but on your worst week you sold 28 a day and the supplier took 14 days:

• Worst case: 28 x 14 = 392 units
• Normal case: 20 x 10 = 200 units
• Safety stock: 392 − 200 = 192 units

It is deliberately conservative, because it sizes your buffer to your worst realistic case. In a volatile year, conservative is not a bad place to be.

3. The statistical (service-level) method

If you want to hit a specific reliability target, use the variability of your demand. This is the standard formula:

Safety stock = Z x standard deviation of daily demand x square root of lead time

Z is the service-level factor. It sets how often you are willing to stock out:

Example: your daily demand has a standard deviation of 5 units, your lead time is a steady 10 days, and you want a 95% service level.

Safety stock = 1.65 x 5 x √10 = about 26 units.

95% is a sensible default for your best-selling, highest-margin items. Push toward 99% only for products where a stockout is genuinely expensive.

4. The variable-lead-time method (the one that matters most in 2026)

Method 3 assumes your lead time is steady. In 2026, for most people, it is not. When both your demand and your supplier's lead time move around, you need this version:

Safety stock = Z x √(average lead time x demand variance + average demand² x lead-time variance)

Take the same product as above (demand 20/day, demand standard deviation 5, lead time 10 days), but now the supplier's lead time swings with a standard deviation of 4 days:

Safety stock = 1.65 x √(10 x 25 + 400 x 16) = about 135 units.

Look at what happened. The same product at the same 95% service level needed 26 units when the lead time was steady and 135 units once the lead time became unpredictable. That five-fold jump is the entire story of 2026 inventory planning: it is rarely your customers that changed, it is your suppliers.

The other side: do not over-hold

More buffer feels safe, but stock is cash sitting on a shelf. Holding inventory costs roughly 20% to 30% of its value per year once you add storage, insurance, obsolescence, and tied-up cash. Overstock is a slower, quieter killer than a stockout, but it kills all the same.

The goal is not maximum safety stock, it is right-sized safety stock: high enough to cover your real variability at the service level you chose, and no higher. If you want to weigh the two failures against each other, our stockout cost calculator puts a dollar figure on running out, and our guide to the carrying cost of inventory puts one on holding too much.

How to tariff-proof your safety stock

You do not need to raise the buffer on everything. Be surgical:

• Raise the service level on tariff-exposed, hard-to-resource SKUs. These are the ones where a stockout could last months. Push them toward 97.5% or 99%.
• Recompute using the variable-lead-time method for anything imported. Steady-lead-time math badly understates what you need right now.
• Extend your planning horizon to 90 to 120 days on key items, so a tariff or shipping shock does not catch you mid-cycle.
• Leave fast, locally-sourced items lean. Do not tie up cash buffering products you can restock in days.

Turning safety stock into a reorder point

Safety stock on its own does nothing until it feeds your reorder point, the stock level that triggers a new order:

Reorder point = (average daily usage x lead time) + safety stock

That is the number you actually act on. For the full walkthrough, see how to calculate reorder points, or use our reorder point calculator.

How Stockria handles this for you

Doing this by hand for one product is easy. Doing it for 500 products, per location, and keeping it current as lead times move is not. In Stockria you set a reorder point and buffer for each item, and the system watches every stock level for you. When an item drops to its trigger, you get an alert and can turn it into a purchase order in two clicks. No spreadsheet, no forgotten reorders, no discovering a stockout after a customer does.

Frequently asked questions

How much safety stock should a small business hold? Enough to cover your realistic worst case at your chosen service level. For most small businesses the average-max method is the right starting point; move to the statistical method for your top items. In 2026, size imported items using the variable-lead-time formula, because supplier delays are the main risk.

What is a good service level for safety stock? 95% is a sensible default for your best sellers. Reserve 99% for products where a stockout is very costly, and accept a lower level on cheap, easily-replaced items.

Should I hold more safety stock because of tariffs? For tariff-exposed and imported items, usually yes, because lead times and prices have become less predictable. For fast, locally-sourced items, no, keep those lean so you are not tying up cash.

How often should I recalculate safety stock? Quarterly at a minimum, and immediately whenever a supplier's lead time changes or you see demand shift. In a volatile year, quarterly is the floor, not the target.

Safety stock is not about hoarding. It is about knowing your real variability and buffering exactly enough to sleep at night without burying your cash in a stockroom.