The Complete Guide to Restaurant Par Levels and Reorder Alerts

Most restaurant inventory problems do not come from a lack of effort. They come from a lack of a reference point. A cook opens the walk-in on a Saturday afternoon, sees three cases of chicken thighs, and has no way of knowing whether that is plenty or a problem. A manager writes the produce order on the back of a prep list, guessing from memory what was used last week. Nobody is being careless. There is simply no agreed number that says how much of each item the kitchen should have on hand at a given moment. That number is a par level, and once you have one for every item that matters, almost every other inventory decision gets easier.

This guide is our attempt to lay out the whole subject in one place. We build ParLevelr, a small tool for par levels and reorder alerts, and most of what follows comes from watching operators set up inventory in real kitchens, sometimes with our software and sometimes with a clipboard. We have organized it into seven themes: what pars are and why they anchor everything, how to stand up a counting system from nothing, how to set and tune pars using actual usage, how reorder alerts and order sheets turn counts into purchase orders, how inventory control affects waste and cash, how to run pars across several locations, and how to keep the routine honest over months and years. Each theme links to a deeper article. Read straight through, or jump to the part that hurts most right now.

What a par level is and why it anchors everything else

A par level is the quantity of an item you want on hand at the start of a defined period, usually right after a delivery arrives or right before the busiest stretch of the week. It is not a minimum and it is not a maximum. It is a target. If you count and find yourself below par, the gap between what you have and what par says you should have is your order quantity, adjusted for anything already on the way. If you count and find yourself above par, you have a signal that something changed: sales dropped, a menu item was pulled, or someone over-ordered. Either way the number gives you a reference to react to instead of a feeling.

Operators sometimes confuse par with a min-max system, where you reorder when stock hits a floor and buy up to a ceiling. Min-max works well for warehouses with steady, predictable draw and large storage. Restaurants rarely have either. Storage is tight, demand swings with the day of the week and the weather, and perishables punish over-ordering within days. A single par tied to your delivery schedule is simpler to teach, simpler to count against, and easier to adjust. That simplicity is why it has survived in kitchens for decades, long before any software existed to manage it.

The reason pars anchor everything else is that they convert inventory from a record-keeping exercise into a decision-making tool. A count without a par tells you what you have. A count with a par tells you what to do. Reorder alerts, order sheets, waste tracking, and cash-in-inventory reporting all depend on that comparison. If you read only one article beyond this guide, make it the one on setting pars that hold up, and if you are starting with nothing at all, the piece on building an inventory system from scratch walks through the order of operations.

Standing up an inventory system from nothing

The first mistake most kitchens make is trying to track everything on day one. A full item list for a mid-sized restaurant can easily run into several hundred lines once you include every spice, every dry good, every paper product, and every bar item. Counting all of it weekly, with a team that has never counted before, is how inventory programs die in their second month. Start instead with the items that drive cost and stockout risk: proteins, dairy, produce that turns fast, the handful of high-volume dry goods, and anything that would stop a menu item cold if it ran out. For most independent restaurants that is somewhere between forty and eighty items. Get those right, then expand.

Before the first count, decide three things for every item on the list: the count unit, the storage location, and who owns it. The count unit is the physical thing someone can see and tally, such as a case, a bag, a number 10 can, or a gallon. It should match how the item sits on the shelf, not how it is priced on the invoice. Storage location matters because a count sheet organized by where things live (walk-in, reach-in, dry storage, bar) is far faster to work through than one organized alphabetically. Ownership matters because an item nobody is responsible for will be counted inconsistently or skipped entirely.

Then run the first count as a baseline, not a judgment. The numbers will be messy. Some items will be in two places. Someone will count a half-case as a full case. That is fine. The goal of the first two or three counts is to shake out the item list, fix the units, and settle the walking route through storage. Only once counts are consistent should you start setting pars against them. The article on starting a real inventory system from scratch goes through this sequence step by step, and the piece on faster weekly counts covers the sheet layout and the counting habits that cut an hour-long count down to a fraction of that.

One more thing about the beginning: pick a fixed day and time for the count and protect it. Counting at the same point in the delivery cycle, ideally right before your main order goes in, is what makes counts comparable from week to week. Counting on Monday one week and Thursday the next produces numbers that cannot be compared, which means pars set against them will be wrong before they are even written down.

Setting pars and calibrating them from real usage

There are two honest ways to set an initial par. The first is to compute it from usage: take how much of an item you go through between deliveries, add a buffer for variability, and that is your par. The second is to ask the person who orders it what number feels right and start there. The first method is better in theory but requires usage data you probably do not have yet. The second gets you moving. In practice most kitchens start with the ordering manager's gut number, run it for a few weeks, and then let usage data correct it. Both are fine as long as you commit to the correction step.

Usage is the number that makes pars real. Usage for a period equals what you started with, plus what was delivered, minus what you ended with. If you count weekly and log deliveries, you get a usage figure for every item every week without any extra work. Over four to eight weeks a pattern emerges: some items are steady, some swing wildly, and some are used far less than anyone assumed. Steady items can run with a small buffer. Volatile items need a larger one, or a par that changes with the day of the week. Items with almost no usage probably should not be on the list at all, or should be ordered only when needed.

The buffer is where judgment lives. A par set exactly at average usage will run out roughly half the time, because half of all weeks are above average. A par set at the highest usage you have ever seen will tie up cash and increase waste. Most operators land somewhere in between, tilting toward more buffer for items that are cheap and shelf-stable and toward less buffer for items that are expensive and perishable. The article on setting par levels that hold up goes deeper on how to choose that buffer, and the piece on using usage to right-size pars over time shows how to review pars on a schedule so they drift with the business instead of freezing at whatever felt right in month one.

A word on seasonality and menu changes. Pars are not permanent. A patio opening, a new menu item, a holiday week, or a nearby event will change usage, sometimes dramatically. The failure mode is not that pars go stale, since they always will. The failure mode is having no routine to notice. A monthly par review, even a quick one, catches most drift before it turns into either a stockout or a walk-in full of product nobody is using.

From count to order: reorder alerts and vendor order sheets

Once pars exist, ordering becomes arithmetic. For each item, par minus on-hand minus anything already on order equals the quantity to buy, rounded up to the vendor's pack size. That calculation is simple, but doing it by hand for sixty items across four vendors every week is where errors creep in. A reorder alert is just that arithmetic run automatically the moment a count is entered, surfacing every item that has dropped below par and suggesting an order quantity. The value is not the math. It is that nothing gets forgotten and the person ordering sees the full picture at once instead of walking the walk-in with a notepad.

Alerts work best when they are tied to your actual delivery schedule. An item that is below par on Tuesday but is delivered on Thursday may not need an emergency order; it may simply need to be on Thursday's list. A good alert system knows which vendor supplies each item and when that vendor delivers, so it can group items into the right order on the right day rather than flagging everything as urgent. It also helps to distinguish between below-par and critically low, so a manager glancing at the list can see what needs a phone call today versus what goes on the regular order. The article on how reorder alerts prevent stockouts covers how to set those thresholds sensibly.

Vendor order sheets are the other half. Most restaurants buy from several suppliers: a broadline distributor, a produce house, a meat or seafood vendor, a dairy, maybe a bakery and a beverage supplier. Each has its own item codes, pack sizes, order cutoffs, and minimums. An order sheet per vendor, prefilled with the items that vendor carries and the quantity the par calculation suggests, turns ordering from a memory test into a review. The manager checks the quantities, adjusts for anything unusual coming up, and sends it. The piece on organizing vendor order sheets across several suppliers walks through how to structure those sheets, how to handle items available from more than one vendor, and how to keep order cutoffs from being missed.

Waste, spoilage, and the cash sitting on your shelves

Over-ordering hides well. A stockout is loud: a server tells a guest the special is gone, a cook improvises, the manager makes an angry call to the vendor. Over-ordering is quiet. The extra case sits in the walk-in, gets pushed to the back, and is discovered on a count three weeks later with a date that has passed. Nobody complains at the moment it happens, so kitchens systematically tolerate more over-ordering than under-ordering. Pars, when they are set honestly, push back against that bias by making the target visible. If par is two cases and there are five on the shelf, the count itself raises the question.

Waste tracking closes the loop. Logging what gets thrown out, even roughly, by item and reason (spoiled, over-prepped, dropped, returned by a guest) tells you which pars are set too high and which prep routines are producing more than sales can absorb. The article on cutting food waste through inventory control makes the case that waste reduction and stockout prevention are not opposing goals. Both improve when pars are tuned to real usage, because the same data that tells you an item is running out too often also tells you when another item is being over-bought.

The cash angle is the part operators most often underweight. Every case on the shelf was paid for, or will be paid for on terms, before it turns into a sale. Inventory that sits is cash that cannot cover payroll, rent, or a slow week. Food cost percentage, the number most operators watch, measures what you used against what you sold. It says nothing about how much you are holding. Two restaurants with identical food cost can have very different amounts of money tied up in storage, and the one holding more is more fragile. The piece on why cash tied up in inventory matters as much as food cost explains how to estimate your inventory value from counts and how bringing pars down on slow movers frees working capital without touching the menu.

None of this means running the walk-in bare. A stockout on a signature item costs more than a few cases of excess dry goods. The point is to know which items deserve a generous buffer and which do not, and to stop treating every item as if it carried the same risk. Expensive, perishable, high-volume items get careful pars and frequent counts. Cheap, stable items can carry more slack and be counted less often. That triage is most of what good inventory control actually is.

Running pars across multiple locations

The moment a second location opens, inventory management changes character. What was one manager's routine becomes a question of standards: do both kitchens use the same item names, the same count units, the same storage layout, the same count day? If not, the numbers cannot be compared, which means the owner or director cannot see whether one location is over-ordering, whether usage differs for a reason, or whether a par that works at the original store makes sense at the new one. Standardizing the item list and the units is the unglamorous first step, and it is worth doing before anything else.

Pars themselves should usually not be identical across locations. Sales volume, menu mix, storage capacity, and delivery frequency differ, so a par that fits a high-volume downtown store will drown a smaller suburban one in product. The workable pattern is a shared item catalog with location-specific par values. Each store's manager owns their pars within a review process, and the person overseeing multiple stores compares usage per cover or usage per thousand dollars in sales across locations to spot outliers. The article on managing pars across multiple restaurant locations goes into how to structure that catalog and review.

Vendor relationships also get more complex. Some suppliers will deliver to every location under one account; others treat each store separately. Order sheets need to be per-location and per-vendor, and consolidating purchasing where possible can improve pricing and reduce the number of cutoffs anyone has to remember. The vendor order sheet article addresses how to keep those sheets consistent when the same item is bought at several stores from the same supplier, and how to avoid the trap of one store ordering under another's account by mistake.

Finally, counts across locations should happen on the same schedule so that a roll-up of inventory value or usage represents the same moment in time. If one store counts Sunday night and another counts Wednesday morning, the combined number is meaningless. This is one of the places where a shared tool earns its keep, because it enforces the schedule and the format without the owner having to chase each manager for a spreadsheet.

Keeping the routine honest over months and years

The hardest part of inventory is not setting it up. It is still doing it in month nine. Counts get skipped on a busy week, then two busy weeks, then the sheet is out of date and nobody trusts it, and the kitchen slides back to ordering from memory. The defenses against that decay are mostly about friction: make the count short, make it the same every time, and make the output immediately useful so the person doing it sees why it matters. A count that takes twenty minutes and produces a ready order sheet gets done. A count that takes two hours and disappears into a spreadsheet nobody opens does not.

Speed comes from layout and habit. Count sheets ordered by physical walking route through each storage area, with the unit printed next to each item, remove most of the thinking. Counting in pairs, one calling and one recording, is faster and more accurate than one person doing both. Keeping storage tidy, with like items together and labels facing out, is an inventory practice as much as a cleanliness one. The article on the fastest way to run accurate weekly counts covers these mechanics in detail and includes the small decisions, like whether to count opened containers by weight or by estimated fraction, that add up to a much shorter count.

Variance is the honesty check. If a count says you used more of an item than sales would explain, something happened: waste, theft, a recipe being over-portioned, a delivery shorted, or simply a counting error. Investigating the largest variances each week, not all of them, keeps the numbers credible without turning the manager into an auditor. Over time, consistently small variances mean the pars, counts, and recipes are all roughly aligned, which is the state every kitchen is aiming for.

Then there is the par review itself, which is the mechanism that keeps the whole system tuned. Once a month, or once a quarter for stable items, look at usage over the recent period against the current par and adjust. The piece on how tracking usage helps right-size pars over time describes a lightweight review that takes a manager perhaps half an hour and prevents the slow drift into either constant shortages or bloated stock. Pars that are reviewed stay useful. Pars that are set once and left alone become another number nobody believes.

More guides on this topic

Further reading from the ParLevelr blog, each answering one specific question in depth.

A working inventory system in a restaurant is not complicated, but it is a system, and every part depends on the others. Pars give counts a purpose. Counts give usage a source. Usage gives pars a correction. Reorder alerts and order sheets turn all of it into a purchase order that gets sent on time to the right vendor. Waste logs and inventory value reports tell you whether the pars are honest. Multi-location standards make the whole thing comparable across stores. And a fixed weekly rhythm, protected from the chaos of service, keeps it alive. If you are starting from nothing, begin with a short item list, a fixed count day, and pars set by whoever orders today. Then let the numbers correct them. That is the entire method, and everything else in this guide and the linked articles is detail in service of it.