You cannot fix what you refuse to count
The first real inventory count is never a number. It is an argument.
By Sovandarapor (James) Kong · · Ventures
There is a particular silence that happens when you ask a team to count something properly for the first time. Not hostility. Something more careful than that. Everybody understands, before a single item is counted, that the number is going to be wrong, and that being wrong will look like somebody''s fault.
That silence is the actual problem. Not the stock. The silence.
Why counting is political
An inventory count looks like arithmetic. It is not. It is a public statement about whether the last twelve months of paperwork was true. When the shelf disagrees with the system, one of two people was wrong, and everyone in the room knows which two people those are.
So counts get soft. The number gets adjusted toward the number that causes the least trouble. Not through dishonesty, usually. Through a hundred small mercies. A broken case is not recorded because it will make someone look careless. A returned crate goes back on the shelf without paperwork because the paperwork is a twenty minute conversation. Each mercy is reasonable. Together they produce a business that does not know what it owns.
Nobody lies about stock. They just decline, politely and repeatedly, to write down the parts that are inconvenient.
Separate the count from the blame
The only thing that made counting work for us was making the first count explicitly consequence-free. Not "we will be understanding about it." Explicitly: this count establishes the truth, nobody is in trouble for what it finds, and the clock starts after.
You get one of these. It is a one-time amnesty and it has to be real, because the moment somebody gets punished for an honest number, you have taught the entire company that honest numbers are dangerous. That lesson takes years to unteach and it costs more than whatever was missing from the shelf.
Then count often enough to be boring
The annual count is theatre. Twelve months of drift, resolved in one exhausting day, by people who want it to be over. What you want instead is counting that happens often enough that no single count carries any drama.
Small, frequent, rotating. A handful of items, this week, done properly. The variance is small enough to actually investigate. You find out that the discrepancy is not theft, it is that two products have nearly identical names and a tired person picked the wrong one at six in the morning. That is a fixable problem. A once-a-year discrepancy of unknown origin across four hundred items is not a fixable problem, it is a write-off.
What the number is really for
People assume the point of an accurate count is control. Knowing what you own, catching what walks out the door. That is part of it and it is the boring part.
The real point is that an accurate count is the foundation under every other number in the business. Your cost of goods is fiction if your stock is fiction. Your margin is fiction. Your view of which products actually make money is fiction, which means your pricing is guesswork wearing a suit.
I have watched businesses argue for months about a pricing strategy built entirely on top of a stock figure that nobody had verified in two years. Everyone in the room was smart. The argument was unwinnable, because it was not really about pricing. It was about a number that was not true.
So count it, take the bad news in one go, and get on with it. The first honest number is always the worst one you will ever have to report, which is a strange kind of comfort.