A par level is the minimum quantity of a SKU you keep on hand before you reorder, set from how fast it sells, how long the supplier takes to deliver, and a safety-stock buffer for problems. Multiply average daily usage by lead-time days, then add safety stock for demand swings.
Many operations that get par levels wrong are not bad at math. They are working from a guess instead of a count. Someone remembers that a part "usually" runs out around the third week of the month, sets a round number that feels safe, and never revisits it. Then the supplier's lead time stretches, or a big order hits, and the round number stops covering the gap. Par levels only work when they come from real numbers pulled from your own history, not a feeling about how busy things usually get.
The three numbers behind a par level
Every par level is built from the same three inputs. Get these from your sales and purchasing history, not from memory.
- Average daily usage: how many units of a SKU move in a normal day, calculated from several weeks of sales or production pulls, not one unusually busy or slow stretch.
- Supplier lead time: the number of days between sending a purchase order and having stock on the shelf, measured from your actual PO history, including the supplier who is reliably slow.
- Safety stock: a buffer sized for a late truck, a short vendor shipment, or a week that sells faster than average, set as extra units rather than extra days of guesswork.
The formula is plain: par level equals average daily usage multiplied by lead time in days, plus safety stock. No part of that formula should come from a hunch.
A worked example
Say a distributor sells a 5-gallon pail of primer at six units a day on average. The supplier takes ten days from PO to delivery, based on the last dozen orders. The buyer wants a 20-unit cushion for a vendor that occasionally ships short. The par level is six times ten, plus twenty, which lands at 80 units. When on-hand plus incoming drops to 80, it is time to reorder, not a week after someone notices the shelf looks thin.
Run that same math on a slow-moving part and the answer looks very different. A SKU that sells one unit a day with a three-day lead time and a five-unit buffer only needs a par level of eight. Carrying 80 units of that part ties up cash and shelf space for stock that will sit for months. Par levels fail when one formula gets applied with the wrong inputs, or when a single round number gets used across SKUs that behave nothing alike.
Where par levels break down
- Usage numbers come from a gut check instead of sales history, so the reorder point is set too low or too high.
- Lead time is measured from the supplier's quote instead of what actually happens on the slow months.
- Safety stock gets padded with a round number instead of sized to the SKU's own demand swings.
- Par levels are set once at launch and never checked again, even as demand or suppliers change.
- Each location tracks its own spreadsheet, so the same SKU has three different par levels and nobody can say which one is right.
What this looks like on the floor
Picture three operations with the same inventory headache and three different fixes.
- A fastener distributor carries roughly 400 SKUs across two sales counters. The counter lead used to reorder hex bolts whenever a bin "looked low," which meant some bins sat half full for weeks while others ran dry mid-shift. After pulling six months of sales history, the team set a par level near 150 units for the fast-moving 3/8-inch bolts and closer to 25 units for a slow seasonal size, instead of ordering both in the same case quantity out of habit.
- A small coffee roaster buys green beans from three suppliers with lead times ranging from about a week for a domestic origin to six weeks for an overseas lot. Treating every bean the same way left the domestic roast overstocked in the back room and the imported single-origin chronically short at the worst time, right before a seasonal blend launch. Splitting the par level by supplier lead time fixed both problems without raising total inventory spend.
- A contract manufacturer running two shifts watched a critical fastener run out mid-production because the par level had been set once, at launch, and never touched again as the second shift roughly doubled daily usage. Recalculating average daily usage against current output, not the original one-shift number, caught the gap before the next shortage instead of after a line stopped.
None of these fixes required new software to start. They required pulling real usage and lead-time numbers instead of trusting a bin that "looks about right." Software starts to matter once the SKU count climbs high enough that nobody can hold all three numbers in their head at once, or when the same part lives in more than one building and needs its own count in each one.
When demand swings hard
A garden-supply retailer sells bagged mulch at a near-flat rate from October through March, then volume roughly triples the first week of April and stays high through June. A par level built from the slow months left the loading dock short every spring, with the counter team calling three suppliers a day trying to find anything in stock before a Saturday rush. Splitting the year into a spring par level and an off-season par level, each with its own average daily usage, closed the April scramble without doubling inventory spend for the other eight months.
A regional parts distributor ships to six technicians, each pulling from the same bearing kit on the truck. The branch manager assumed one par level per SKU covered the whole operation, then found the opposite every time the shared part ran out on a Tuesday because Monday's pulls had already emptied the bin for the week. Tracking the kit by truck instead of by branch made the shortage visible before the bin hit zero, not after a technician called in from a job site asking for a part that should have been on the shelf.
A hardware co-op member running a single location watched a fastener supplier's lead time creep from seven days to nineteen over two years without anyone updating the math. The formula still used the old lead time, so the reorder point sat almost three weeks short of what the current supplier actually needed. Pulling the last twelve purchase orders instead of trusting the number from the original setup closed the gap in one afternoon, and the next order landed before the shelf ran dry.
How often to revisit par levels
Par levels are not a one-time setup. Review them at minimum every quarter, and sooner after a seasonal swing, a new supplier, or a lead-time change you notice on an incoming shipment. A par level set during a slow month will undershoot the moment demand returns. A business that reviews par levels once a year is carrying the cost of stale numbers for eleven of those twelve months.
Where Fishbowl fits
Fishbowl is the platform. Fishbowl Inventory tracks on-hand quantity, committed sales, and incoming stock per SKU and per building, so average daily usage and supplier lead time come from real transaction history instead of a spreadsheet someone updates when they remember to.
The AI Operations Specialist is currently in beta and uses live Fishbowl data to prepare purchase orders, transfers, receipts, and adjustments for your approval. When a SKU crosses its par level, the specialist drafts the reorder instead of leaving someone to notice the shelf is thin. Nothing is saved or changed without the operator's approval. Ask, Approve, Done.
Brian Garrett at Telluric Companies saves 45 minutes per purchase order using that prepared-draft workflow, time that used to go into pulling usage numbers and building each PO from scratch.
Setting par levels well takes a system that already has the usage and lead-time history in one place, not a second spreadsheet to keep updated by hand. See how it works on a live SKU: get a demo.
Fishbowl does the work. You make the call.
Frequently asked questions
What is a safety stock buffer and how is it different from a par level?
Safety stock is the cushion inside the par level, not a separate number. The par level is average daily usage times lead time, plus safety stock for a late shipment or a demand spike. Safety stock alone will not tell you when to reorder; the full par level will.
Do I need a different par level for every warehouse?
Yes, if each location sells at a different rate or uses a different supplier lead time. A SKU that moves fast at one site and slow at another needs two par levels, tracked by location, not one company-wide number averaged across both.
How often should I recalculate par levels?
Review par levels every quarter at minimum, and sooner after a seasonal swing, a new supplier, or a lead time change. A par level set during a slow month will undershoot the moment demand picks back up.
How long does it take to get par levels running in new software?
Fishbowl says most customers go live in 2 to 6 weeks, useful from the first live count. Support during setup is in-house phone, live chat, and email support, so questions about a specific SKU's history get answered by a person.