Par Level Inventory and How to Set Par Levels

A par level is the minimum quantity of an item you want on hand at any time. When stock drops to par, you reorder enough to bring it back up to full. It is the same idea as a reorder point, dressed in the language of restaurants, bars and hospitality, where "par" has been the standard word for decades. Set a par for each item, check against it on a schedule, and you keep the shelf stocked without counting everything every day.

Par level inventory is popular because it is easy for a team to run. A new bartender or line cook can look at a par sheet, see that you keep six bottles of an ingredient, count what is left, and order the difference. The catch is that a par set once and never revisited slowly drifts out of line with real demand. Here is how par levels work, how to set them properly, and where fixed pars go wrong.

How par level inventory works

Every item gets a par: the level you never want to fall below between orders. On each ordering cycle you count what is on hand, compare it to par, and order enough to top back up. If your par for an ingredient is 6 units and you have 2 left, you order 4. The par has to cover expected usage across the time until your next delivery arrives, plus a cushion for a busy day or a late supplier. Done well, par levels turn ordering into subtraction: par minus on-hand equals what to buy.

How to set a par level

A good par is built from real usage, not a guess. The core formula is:

Par level = (Average daily usage × Days between deliveries) + Safety stock

Say a bar goes through 5 bottles of a spirit a night, deliveries come twice a week so up to 4 days can pass between them, and you want a 5-bottle cushion for a busy weekend. The par is (5 × 4) + 5 = 25 bottles. When the count drops to or below 25, you order back up to full. The safety stock term is what stops a normal spike from turning into an empty shelf, so size it to how variable that item's demand really is.

Par level vs reorder point vs min/max

These three are close cousins. A reorder point is the level that triggers an order; par level is the same trigger in hospitality language. A min/max system adds a second number, the maximum you top up to, whereas a classic par sheet usually treats par itself as the "order back up to this" target. In practice a par level often acts as both the minimum and the top-up target: you order the gap between on-hand and par every cycle. The maths underneath all three is the same demand-times-lead-time-plus-buffer calculation.

A worked restaurant example

Picture a kitchen that orders produce three times a week. It uses about 8 cases of tomatoes between deliveries in a normal week, but a catering booking can push that higher, so the chef holds a 3-case cushion. The par is 8 + 3 = 11 cases. On order day the prep cook counts 4 cases left and orders 7 to reach par. The same logic runs across every line item on the par sheet, which is why a well-built sheet lets a junior team member place a sound order without the chef in the room. The quality of that order depends entirely on the usage numbers behind it, so it helps to pull the real quantities off each delivery invoice rather than eyeball how much you think you used.

Where fixed par levels go wrong

The weakness of par levels is the same as any static threshold: demand moves and the numbers do not. A par set for winter trade is wrong by summer. Pars set when you opened are wrong once a dish takes off or fades. Set them once across a big menu and you drift into carrying too much of the items that slowed down and running short on the ones that grew. Pars also assume steady lead times, so a supplier that starts arriving a day late quietly turns a safe par into a risky one.

What par level should you set for an item with no recent purchase history?

Set the par from usage, not from purchase orders. If an item has stock on hand but no recent PO, count what was used or sold over the last four to eight weeks, convert it to a daily rate, and multiply by the days between deliveries plus the supplier's lead time, then add a small buffer. With no usage at all, do not set a par yet: flag the item as slow stock and decide whether to keep carrying it.

Should a par level be in each units or the purchasing unit of measure?

Calculate the par in the unit you use (each, bottle, pound), then round the order up to the unit you buy (case, pack, pallet). Keeping the math in the usage unit stops a par of "2 cases" from hiding a drop from 48 to 30 units a week. The rounding to the case pack belongs on the order, which is how inventory replenishment software builds order quantities from a usage forecast.

Making par levels dynamic

The fix is to treat pars as living numbers that update from recent usage, not a laminated sheet from last year. The modern version keeps the same simple par-minus-on-hand ordering, but recalculates each par from that item's own recent demand and current lead time. Instead of a manager revisiting hundreds of lines by hand, software reforecasts usage and resets the pars, so the shelf level tracks real trade. That live recalculation is the core of what Storekeeper does, alongside per-SKU demand forecasting that keeps every level honest.

Want to know whether your current pars still match how you actually sell? Paste your stock and recent usage into the live stock scan at the top of the site. It shows which items are sitting well above a sensible par and which are already running below one, so you can retune the worst offenders first.