Skip to content
ANANTATECH HUB

Worked scenario

Twenty pieces missing and no day where it shows

Imitation jewellery is not a smaller version of a gold business. The risk is not theft of one valuable item, it is slow leakage nobody can see.

2 min read

Not a small gold shop

People assume imitation jewellery works like gold with cheaper stock. It does not. There is no hallmarking, no making charge, no weight-based valuation and no customer who will notice if one piece is missing from a tray of forty.

What there is instead is volume. Thousands of designs, bought in mixed lots, sold singly, at prices where a single piece is a rounding error in the day's takings.

Why the loss is invisible

Twenty pieces at ₹300 is ₹6,000 of stock. Spread across a month in a shop turning over reasonably, no single day looks wrong. The cash reconciles every evening because the cash is fine. The stock is what moved.

It surfaces at the annual count, as a number the owner cannot explain and cannot attribute to any period, supplier or member of staff. By then it is eleven months old and there is nothing useful to do with it.

The queue test

One thing decides whether this works in a jewellery counter, and it is not the stock model. It is whether billing survives a Diwali-week queue.

A system that adds four seconds per line is fine on a Tuesday in June and unusable when there are eleven people waiting. When that happens, staff bill everything against a generic line to keep the queue moving, and the stock record silently becomes fiction while appearing to work perfectly.

Time it at peak before committing to it. Not at the demo, not in the morning. At peak.

Counting a slice, weekly

The annual stock-take is the wrong instrument. Counting a rotating slice each week, perhaps forty or fifty designs, catches a discrepancy within days of it happening, when it can still be traced to a delivery or a shift.

It also takes less total time than one annual count, which is the part owners do not expect.

What it will not do is tell you whether a piece walked out or was miscounted at receipt. It narrows the window to a week, which is usually enough to change behaviour. The system reports a difference. It does not explain one.

What changes

  • Design-level counts that make a discrepancy visible while it is still small
  • Counter billing fast enough to survive a festival queue
  • Movement data separating designs worth repeating from designs worth clearing
  • Purchases tied to designs rather than to a supplier's lot description

Questions about anything here, or a situation this does not cover? contact@anantatechhub.com