A QSR snack brand location saw sales decline 5.7% year over year but kept COGS as a percentage of sales virtually unchanged, moving from 14.94% to 14.97%. The location was using weekly inventory counts, which adjusted PAR levels downward before excess ordering could compound, corresponding with stable margin during the sales decline.
Sales drop. It happens. A mall sees less foot traffic. A nearby construction project reroutes customers. A new competitor opens down the street. The first thing most operators do is panic-order. If sales are down, the instinct says order less. But ordering less without data is just guessing in the other direction.
The problem with a sales decline is not the decline itself. It is the lag. If you do not have weekly usage data, you do not know sales are softening until the monthly report comes in. By then, you have been ordering at the old PAR levels for three or four weeks. The product is sitting on the shelf. The COGS as a percentage of sales spikes because the denominator (sales) shrank while the numerator (what you ordered) stayed the same.
A QSR snack brand location in a regional mall saw sales decline from $839,458 to $791,576 over the year. A 5.7% drop. Not catastrophic, but enough to hurt if COGS does not adjust.
Here is what the numbers looked like:
| Period | COGS as % of Sales | Total COGS | Total Sales |
|---|---|---|---|
| Prior Year (full year-to-date) | 14.94% | $125,388 | $839,458 |
| Current Year (full year-to-date) | 14.97% | $118,475 | $791,576 |
| Change | +0.03 percentage points | -$6,913 | -$47,882 |
COGS as a percentage of sales moved by 0.03 percentage points. Essentially flat. That is the result you want when sales decline. The COGS dollar amount came down by $6,913, tracking with the sales decline rather than staying fixed at last year's level.
When sales drop, most locations see their COGS percentage get worse, not stay flat. The reason is simple: ordering is a habit. You order what you ordered last week. If sales are declining gradually, you do not feel it day to day. You feel it at the end of the month when the numbers come in.
Without a weekly count, there is no early warning. The first sign of a sales decline is the monthly P&L, which means three or four weeks of ordering at the wrong level. Each week compounds the problem. Product accumulates. Some of it expires. The COGS report shows the damage.
This location was counting every item every week. By the second week of softer sales, the usage data showed it. Items that normally turned over in five days were sitting for seven. The rolling 3-week average started to trend down. PAR levels adjusted downward automatically.
The manager did not have to guess whether the decline was real or just a slow week. The data answered the question. If it was a one-week blip, the average would absorb it. If it was a trend, the PAR levels would come down and the ordering would follow.
There is a critical difference between manually cutting an order because sales feel slow and letting PAR levels adjust based on usage data. Manual cuts are emotional. You cut too much, run out, panic, and over-order the next week. The cycle repeats.
PAR level adjustments based on a rolling average are gradual. They come down a little each week if the trend holds. They come back up if sales recover. Nobody has to remember to make the change. The data informs it.
If your sales are softening and your COGS percentage is climbing, the problem is not the sales decline. The problem is that your ordering has not adjusted to match. Without weekly data, you are ordering based on what happened last month, not what is happening this week.
The location in this case study held its COGS margin within 0.03 percentage points despite a 5.7% sales decline. The location was counting every item every week and letting PAR levels adjust to real demand. No guessing. No panic cuts. Just data informing the order.
The rolling 3-week average begins adjusting within the first week of a decline. A single slow week will not dramatically change PAR levels because the average absorbs it. But if the decline persists for two or three weeks, the average trends down and PAR levels follow. This prevents both over-reaction to a single bad week and under-reaction to a sustained decline.
If the decline follows a predictable seasonal pattern, seasonal multipliers handle it. The app adjusts PAR levels based on the known seasonal cycle, so you are not relying on the rolling average to catch something you already know is coming. The rolling average handles unexpected declines; seasonal multipliers handle expected ones.
Yes. The primary waste during a sales decline is ordering product that does not sell before it expires. Weekly counts catch the excess before it compounds. You see the product sitting on the shelf. You see the usage trending down. The PAR level comes down. The next order is smaller. The waste does not happen because the order was right-sized before the product arrived.