PAR levels should adjust gradually during sales growth using a rolling 3-week average, not manual guesswork. This prevents the most common growth mistake: ordering more of everything when only some items are actually selling faster. The rolling average raises PAR only for items with real usage increases while holding steady on everything else.
Growth is the problem you want. But it is still a problem. Sales go up, you order more, and somehow your COGS percentage gets worse instead of better. The reason is almost always the same: you treated growth as a blanket event instead of an item-by-item shift.
Sales growth does not happen evenly. Three items might be driving 80% of the increase. The other 47 items in your storage room have not changed. But if you bump your entire order up by 20% because sales are up 20%, you just over-ordered 47 items. That waste shows up in next month's COGS report.
PAR (Periodic Automatic Replacement) is the quantity of an item you need on hand to cover usage between orders. If your PAR for flour is 4 cases and you have 2 on the shelf, you order 2. The PAR is the target. The count tells you where you are. The difference is what you order.
During growth, the question is not "how much do I order?" It is "should my PAR change?" If sales are genuinely up on an item, the old PAR might be too low. You will hit red (below PAR) more often, run out, and miss sales. If the PAR does not need to change, raising it just means you order more than you need.
The rolling average looks at the last 3 weeks of usage for each item. If an item's usage is climbing, the average climbs. If it is flat, the average stays flat. If it dropped, the average drops. The PAR adjusts based on where the average goes.
Here is what that looks like for three different items during a growth period:
| Item | Week 1 Usage | Week 2 Usage | Week 3 Usage | 3-Week Avg | PAR Action |
|---|---|---|---|---|---|
| Item A (growing) | 10 | 12 | 14 | 12.0 | PAR rises gradually |
| Item B (flat) | 8 | 8 | 8 | 8.0 | PAR unchanged |
| Item C (declining) | 6 | 5 | 4 | 5.0 | PAR comes down |
Item A is driving the growth. Its PAR goes up. Items B and C are not part of the story. Their PAR levels stay where they are. The total order increases, but only because of the items that actually need more. This is how you grow sales without growing COGS proportionally.
Manual PAR adjustment during growth usually goes like this: the manager sees a busy week, gets nervous, and bumps the order up across the board. Maybe 15%. Maybe 20%. The justification is "we are busier, so we need more of everything."
Three weeks later, sales normalize or the growth shifts to different items. But the PAR levels are still elevated. Nobody remembers to bring them back down. The excess product sits on the shelf. Some of it expires. The COGS report shows the damage.
The rolling average solves this two ways. First, it raises PAR gradually, not all at once. Second, it brings PAR back down automatically if usage returns to normal. Nobody has to remember to undo the increase.
A holiday weekend will spike sales for 3 or 4 days. Without a rolling average, that spike could permanently raise your PAR levels. The rolling average absorbs it. One busy week out of three raises the average slightly, but not enough to dramatically change the PAR. By the following week, the spike is out of the average and the PAR returns to normal.
Sustained growth is different. If usage is genuinely trending up week after week, the average climbs steadily and the PAR follows. The system distinguishes between "one good weekend" and "a real trend" without the manager having to make that judgment call.
The rolling average reacts within the first week of increased usage, but the adjustment is gradual. A 10% usage increase in one week might raise the average by 3-4%. By the third week of sustained growth, the average has caught up and the PAR reflects the new reality. This prevents both under-ordering (missing sales) and over-ordering (waste from jumping the gun).
If you know growth is coming (a new product launch, a marketing push, a seasonal shift), you can use seasonal multipliers to pre-adjust PAR levels. This is different from the rolling average. Seasonal multipliers apply a known percentage increase to PAR levels for a defined period. When the period ends, the multiplier turns off and PAR returns to the rolling average baseline.
Use seasonal multipliers for predictable events. Let the rolling average handle unexpected growth. The two systems work together: multipliers for what you know is coming, rolling average for what you discover is happening.
Yes. The weekly count is what feeds the rolling average. Without the count, there is no usage data. Without usage data, the average has nothing to calculate. The count is the input. The PAR adjustment is the output. The order is the action. Growth just means the numbers are bigger, but the loop is the same: count, calculate, order, confirm, repeat.