How to set par levels and stop over-ordering in hospitality

Over-ordering is one of the most reliable ways to quietly destroy margin in a hospitality operation. It does not show up as a dramatic loss. It shows up as products that sit in the store longer than they should, as waste that accumulates gradually, as cash tied up in stock that is not generating revenue. If reducing food waste is a priority for your operation, tightening par level discipline is one of the most effective starting points.

Par levels are the tool designed to prevent it. A par level is the minimum quantity of a product you want to have on hand before the next delivery. When stock falls below par, you order. When it is above par, you do not.

Simple in principle. Routinely mismanaged in practice.

Why most par levels are wrong

Par levels get set once, during setup or at the start of a trading period, and then left in place indefinitely. The problem is that the right par level for a product depends on three things that change constantly: how much you use, how often your supplier delivers, and how much storage space you have.

An operation that sets par levels for its spirits in January, based on winter trading patterns, will be under-stocked every Friday in August if it has not revisited them. A venue that switches from twice-weekly to once-weekly deliveries needs to double its par levels to maintain the same safety stock, and usually does not.

The result is a purchasing process that either over-orders consistently (par levels set high, orders placed regardless of current stock) or under-orders at busy periods (par levels too low, operation runs short mid-week and makes emergency orders at worse prices).

Both outcomes cost money. The over-ordering costs in waste and tied-up cash. The under-ordering costs in emergency purchases, delivery charges, and the operational disruption of running out.

How to calculate a par level properly

A par level has two components: working stock and safety stock.

Working stock is what you use between deliveries. If your supplier delivers every Tuesday and you use two cases of house red between Tuesday and the following Tuesday, your working stock requirement is two cases.

Safety stock is the buffer against unpredictable demand or supply. A busy weekend, a delayed delivery, a product that sells faster than expected. A reasonable safety stock for most products is one to two days of average usage. For high-value or hard-to-replace products, consider a slightly larger buffer.

Par level = working stock + safety stock.

For the house red example: if you use two cases per week and want one day of safety stock on top, your par level is approximately 2.3 cases. Round to three cases to work in whole units.

When stock falls to three cases or below, it goes on the order. When it is above three, it does not.

The categories that need different approaches

Not all products should be managed the same way.

High-volume, low-value products (salt, oil, flour, basic ambient staples): keep a reasonable buffer and order regularly. The cost of running out is high relative to the cost of holding a bit more stock. Par levels can be generous.

High-value, perishable products (fresh fish, premium proteins, specialty produce): tight par levels and frequent communication with suppliers. The cost of over-ordering is spoilage. Order little and often. Build supplier relationships that allow short-notice changes.

Spirits and wine: weekly count, weekly order. Shrinkage risk is real and value is high. Par levels should be reviewed seasonally as cocktail menu changes affect usage rates significantly.

Draught products: set by the keg, factoring in line throughput. Running out mid-service is a revenue and experience problem. Safety stock should be at least one full keg for each high-volume line.

Seasonal products: par levels need active adjustment. A product that sits at the bottom of the menu in winter and anchors it in summer cannot have a fixed par level that works year-round. Flag seasonal products for review when you change the menu.

The ordering discipline that makes par levels work

A par level is only useful if the ordering process checks against it consistently.

The most common failure is ordering by memory or by "the same as last week." This approach ignores current stock levels. If you have four cases of something and your par level is three, ordering the same as last week means you are carrying more than you need. Over several weeks, this compounds into significant over-ordering.

The correct approach: before placing any order, run a current stock check against par levels. What is below par needs ordering. What is above par does not. The order is driven by the gap between current stock and par level, not by habit.

For most operations, this check takes five to ten minutes if the stock data is current. If the last stock take was three weeks ago, the check is unreliable and the ordering reverts to guesswork.

This is why par level management and stock take discipline are not separate practices. One makes the other useful.

Reviewing and updating par levels

A par level that is not reviewed regularly will drift away from accuracy. Build a calendar prompt to review par levels at minimum:

• At every seasonal menu change
• When a supplier changes their delivery schedule
• When a product's usage rate changes significantly (new cocktail menu, new section head, new supplier)
• When storage capacity changes

The review does not need to be exhaustive every time. A scan of the products where over-ordering or under-ordering is a pattern, compared against recent usage data, takes thirty minutes and catches most of the drift.

Operations that actively manage par levels tend to carry less stock on average than those that do not, while running out less frequently. The reduction in tied-up cash and waste pays for the time spent on the review many times over. For operations managing stock across multiple locations, the same principles apply at greater scale -- see managing inventory across multiple sites for the additional considerations.

Par levels in a digital system

In a spreadsheet-based ordering process, par levels are a column that has to be manually compared against a stock count column. The order calculation is done by hand or by formula. If either input is out of date, the calculation is wrong.

In an inventory system, par levels are a property of each product. When a stock take is completed, the system shows you what is below par and what needs ordering automatically. The purchasing suggestion is generated from real stock data, not from memory or habit.

The practical difference is that the ordering process shifts from a mental exercise (what do I think we need?) to a review exercise (what does the data say we need?). The second is faster, more consistent, and less dependent on the individual doing the ordering having a complete mental picture of the stockroom.

If your current ordering process relies on someone walking around and making judgements from memory, Hops was designed to give that person a tool that makes the decision from data instead.