Stock Discrepancy Explained: What It Means and How to Investigate It

A stock discrepancy is simply the difference between your expected stock count — what your purchase records, sales, and previous counts say should be on the shelf — and your actual physical count. It can run in either direction: you can have less stock than expected (a shortage) or, less commonly, more than expected (an overage, usually from a recording error rather than stock appearing from nowhere). The word itself is neutral. It describes a gap, not a cause. The cause is what an investigation is for.

Legitimate causes — more common than most owners assume

Before any conversation about theft, it's worth ruling out the causes that have nothing to do with dishonesty. In most shops that haven't been counting stock carefully, these explain a larger share of discrepancies than people expect.

• Supplier shortchanging a delivery — a carton marked as 24 units arrives with 22, and nobody counts it against the invoice on arrival
• Damaged or spoiled goods that were pulled from the shelf but never formally written off in the records
• Simple miscounts — easy to do with fast-moving, similar-looking items, especially under a manual, once-a-month count
• Unrecorded returns or exchanges — a customer swaps an item and it's handled informally, off any log
• Items moved between locations (a back-store to front-shelf transfer, or between two branches) without being logged
• Pricing or unit errors — an item logged as a carton of 12 when it was actually sold as loose units, throwing off the count on both sides
• Promotional giveaways, samples, or staff meals/consumption that were approved but never deducted from stock records

Illegitimate causes

The other category is loss without a legitimate explanation — theft by staff, theft by customers (shoplifting), or deliberate fraud such as a staff member logging a sale as a discount or void and pocketing the difference. These are real and worth taking seriously, but they should be a conclusion you arrive at after ruling out the legitimate causes above, not a starting assumption. Treating every discrepancy as theft either wastes time chasing a mistake that has an innocent explanation, or — worse — damages trust with an honest employee.

How to investigate a discrepancy, step by step

1. Recount before anything else. A meaningful share of "discrepancies" are simple counting errors, especially with items that look similar or are counted quickly. Recount the specific item, ideally with a second person, before treating the gap as real.
2. Check the paper trail for that item. Compare the last delivery invoice, any damage or write-off log, and recent sales for that product. Many gaps resolve the moment you actually look at the delivery note against what was signed for.
3. Narrow the window. If you count daily or per shift, you can isolate exactly when the gap appeared. If you only count monthly, you're investigating a 30-day black box — which is a strong argument for counting more often, not just for this one discrepancy.
4. Look for a pattern, not a single event. Is this a one-off on one item, or does it repeat on the same product, the same shift, or the same day of the week? A pattern points toward a cause; an isolated blip usually doesn't.
5. Talk to whoever was on shift, but ask before you conclude. Frame it as "help me understand this gap" rather than an accusation. Staff often have a legitimate explanation — a return that wasn't logged, a customer complaint that led to an exchange — that resolves the discrepancy immediately.
6. Only escalate to a theft investigation once legitimate causes are ruled out and a pattern holds across multiple counts. At that point, follow a fair, evidence-based process before any accusation or disciplinary action.

The single biggest lever in all of this is how often you count. A shop that only reconciles stock once a month is trying to investigate discrepancies with a month's worth of sales, deliveries, damages, and shift changes all tangled together — by the time you notice a gap, there's no way to know which of those seven possible causes actually produced it. A shop that counts daily, or per shift, narrows that window to hours, which makes almost every step above faster and more conclusive.

This is the specific problem Shelfie is designed around. An attendant photographs the shelf and Shelfie's AI counts what's visible and reconciles it against expected stock the same day, flagging any gap immediately rather than at a month-end audit. Because each count is tied to a specific shift and time-stamped, narrowing down when a discrepancy appeared — step three above — stops being guesswork and becomes a quick look at the record.

When a discrepancy is small enough to ignore

Not every gap needs a full investigation. A small, one-off variance on a low-value item, with no repeat pattern, is normal in any retail operation and not worth the time it would take to chase. The line worth watching for is repetition: the same item, the same shift, or the same magnitude of loss showing up again and again. That's the signal that separates ordinary noise from something that actually needs to be run down.