Managing Jewellery Stock Through Dhanteras, Diwali and Wedding Season Without Losing Count

On a normal Tuesday a showroom might move 30 to 50 pieces across the counters. On Dhanteras the same shop can churn through 400 to 600 pieces in a single day, with three customers waiting per salesperson and trays moving between the safe, the display case and the polishing desk every few minutes. Akshaya Tritiya and the October to December wedding window stretch that pace over weeks rather than hours.

Shrinkage in this period is rarely a dramatic theft. It is a tray of bangles that came out of the locker and went back one piece short, a lightweight chain swapped for a heavier-looking but lower-karat one, a returned item logged against the wrong barcode, or a piece sent for sizing that never came back. None of these get noticed on the day because nobody has time to weigh and reconcile while the shop is full.

The financial exposure is real. A single 22K gold bangle of 20 grams at roughly Rs 6,500 per gram is about Rs 1.3 lakh of stock. Lose three or four pieces across a festival fortnight and you have wiped out the margin on hundreds of genuine sales. The whole point of festival-season discipline is to make verification fast enough that it survives the rush instead of being abandoned the moment things get busy.

You cannot verify against a count you do not trust. The single most useful thing you can do two weeks before Dhanteras is a full physical stock take while the shop is still calm, and lock that result in as your festival baseline. Every piece gets a barcode, a recorded weight to two decimal places, and a clear reference photo. This baseline is what every later check is measured against, so the half day it costs now saves you from arguing with your own records in December.

Do this section by section rather than all at once, so a counter is never fully closed during business hours. Count the locker stock first thing in the morning before display trays are pulled, then reconcile display against the system before lunch. If you import your stock via CSV with barcode and weight as the baseline, this is the moment to clean that file: remove sold items still showing as in stock, fix weights that were entered as full grams when the piece is in fractions, and flag pieces with no photo.

Fresh stock arrives in bulk before every festival, often as last-minute karigar deliveries the night before Dhanteras. This is the highest-risk moment because tired staff are tempted to put new pieces straight into the display to fill gaps. Resist it. Create a simple receiving station, even if it is just one table near the safe, and rule that nothing moves to a counter until it has been weighed, photographed and added to the system.

For a delivery of, say, 120 new pieces, a two-person team can intake roughly 40 to 60 pieces an hour: one person scans or assigns the barcode and captures the photo, the other weighs and reads out the figure. Cross-check the karigar's challan weight against your recorded weight piece by piece. A 0.3 g gap on a 15 g pendant is the difference between an honest rounding and a quiet underweight delivery you would otherwise pay full price for.

During the rush, ad-hoc checking collapses because everyone assumes someone else did it. The fix is to assign verification tasks the way you assign sales counters: by name, by counter, by shift. The person responsible for the bangle counter on the morning shift owns the open-to-close reconciliation of that counter's stock, and the system shows exactly which items are on their list.

Use role-based permissions so the assignment also protects you. A junior salesperson can scan a barcode, photograph a piece and record its weight, but only a manager can clear a flagged discrepancy or edit a baseline weight. That separation matters most in a crowded shop, because it means a count can never be quietly 'corrected' to hide a shortfall by the same person who created it.

Keep the unit of work small. Instead of asking a person to verify a whole counter at once, assign trays: tray of 24 rings, tray of 30 chains. A tray takes a few minutes between customers, so verification happens in the gaps of the day rather than as an exhausting end-of-day marathon nobody completes.

The verification step itself has to take seconds, or it will not happen on a busy day. The practical loop for each piece is: scan or enter the barcode, photograph the item, place it on the scale, and submit. The recorded weight is compared against the baseline and the captured photo is matched against the baseline image. If both agree within tolerance, the piece is marked verified and the staff member moves on. If either disagrees, the manager is alerted immediately.

Two checks together are what make this trustworthy. Weight alone catches an item that has been replaced with something lighter or heavier, but a clever swap might match the weight while being a different, lower-value piece. Image matching catches that. A barcode scan alone, or an RFID tag, only tells you a tag is present; it cannot tell you the piece under that tag is the same gold you photographed two weeks ago. Weight plus image is what turns a count into genuine accountability.

Set a sensible weight tolerance. Cleaning, polishing and normal handling can shift a reading by a few hundredths of a gram, so a tolerance of around 0.05 to 0.10 g for small items avoids false alarms while still catching a real swap, which will usually be off by half a gram or more. Tune it to your scale's accuracy rather than chasing a perfect number.

Festival selling generates an unusual number of returns and exchanges: a gift that did not fit, a bangle swapped for a larger size, an item sent for sizing during the rush. Each of these is a moment where stock leaves and re-enters your system, and each is a common point where the count quietly drifts. The rule that saves you is simple: a returned or exchanged piece is treated exactly like new incoming stock and must be re-verified before it goes back on display.

Repairs and sizing jobs need their own visible list. When a piece leaves for the karigar bench, it should be logged out against its barcode and recorded weight, and logged back in with a fresh weight when it returns. Comparing the before and after weight catches the rare but expensive case where a piece comes back lighter than it left. Without this log, a piece 'out for sizing' is the easiest thing in the shop to lose track of for a week.

End each peak day with a short, structured close rather than a vague glance at the trays. The goal is a clean three-way match: physical pieces on hand, system count, and the day's sales and returns. Because trays were verified in small batches through the day, the close becomes a quick confirmation that the verified count plus sales equals the morning count, not a fresh stock take at 10 pm when everyone is exhausted.

When the season ends, run a full reconciliation against the festival baseline you locked two weeks earlier. This is where a tamper-proof audit trail earns its keep: for every piece you can see who verified it, when, the recorded versus baseline weight, and the captured versus baseline image. If something is short, you are not guessing; you can trace the last point at which the piece was confirmed present and by whom.

Treat the numbers as a learning tool for next year. If most of your discrepancies clustered on the bangle counter during the evening shift, that tells you where to add a second verifier next Dhanteras. If most flags were false alarms from a tight weight tolerance, loosen it slightly. The shops that lose the least are the ones that review this honestly each season instead of being grateful nothing 'big' went wrong. If you want to see how this whole loop works end to end on a single platform, it is worth booking a demo before your next peak.