Cutting Theft and Shrinkage in Indian Factories with Cameras
Shrinkage — stock that vanishes between receiving and dispatch — quietly drains an estimated 1–2% of throughput at a typical Indian plant, split between employee pilferage, external theft, and paperwork error. Camera analytics attacks the theft share directly: it watches docks, stores, and gates for unauthorised removal and flags it in near-real time.
Most owners treat shrinkage as a cost of doing business — a number the auditor reconciles once a year. That is exactly why it stays expensive. This article puts India-specific figures around the problem, separates the losses you can actually catch on camera from the ones you can't, and works the ROI against the one line item every plant already pays for: guards.
How big is factory shrinkage in India?
Hard manufacturing-specific shrinkage data for India is thin, so the honest starting point is retail, where it has been measured. The Global Retail Theft Barometer put India at the highest shrinkage rate in the world, 2.38% of sales, with the loss split 45.2% shoplifting (external), 23.3% employee theft, and 22.6% administrative/process error (reported via IFSEC Insider). That figure is from the 2011 barometer (2010–11 data) and the study was later discontinued, so treat it as directional, not current.
More recent signals come from listed Indian retailers' own disclosures: in FY24, Trent reported shrinkage rising to 0.41% of sales (from 0.22%), and V-Mart to 0.5% (from 0.4%) — both blaming volume growth and a mix of shoplifting and employee action (Business Standard, June 2024).
Factories differ from stores in one important way: the "shoplifter" isn't a customer. On a factory floor, external theft means contractors, transporters, scrap dealers, and after-hours intruders — and the internal share (staff, supervisors, dock crews) tends to be higher than in retail, because insiders know the blind spots, the shift gaps, and the reconciliation lag. As a planning assumption, 1–2% of material value lost to shrinkage is a realistic indicative band for a mid-size Indian plant handling loose, portable, or high-value stock (metal offcuts, copper, yarn, fasteners, pharma packaging, FMCG cartons).
Employee vs external vs paper: what cameras actually catch
Not all shrinkage is a security problem, and cameras only help with part of it. Splitting the loss by cause is the first step to sizing what an analytics layer can recover.
| Loss type | Typical share (indicative) | Where it happens | Can camera analytics catch it? |
|---|---|---|---|
| Employee / internal theft | ~25–35% | Stores, dispatch, scrap yard, tool cribs, shift changeover | Yes — unauthorised removal, after-hours store access, loading unlogged goods, tailgating through gates |
| External theft | ~30–50% | Perimeter, gates, loading docks, parking, transporter interface | Yes — intrusion, loitering, perimeter breach, vehicle in restricted bay, dock activity outside schedule |
| Supplier / vendor fraud | ~5–10% | Inbound dock, weighbridge, receiving | Partly — short-loads and swap-outs are visible on dock cameras but need cross-check against GRN/weighbridge |
| Administrative / process error | ~15–25% | Counting, data entry, reconciliation | No — this is a systems/ERP problem, not a camera problem |
The takeaway: cameras realistically address the theft slice — roughly half to two-thirds of total shrinkage. Marketing that promises to erase all shrinkage with cameras is overselling; the paper-error slice needs process and ERP fixes, not lenses.
What the analytics module looks for
Modern warehouse and dock analytics — offered in India by vendors such as Agrex AI, viAct, and Katomaran, running on your existing RTSP/CCTV feeds — cluster around a handful of theft-relevant detections:
- Intrusion and perimeter breach — a person crossing a fence line or entering a yard after hours.
- Loitering — prolonged, unusual presence in a restricted or sensitive zone (scrap yard, finished-goods store, dock at 2 a.m.).
- Unauthorised removal / dock anomaly — goods moving off a dock outside the expected workflow or schedule.
- Tailgating and gate events — a second person or vehicle slipping through on one authorisation.
Indian logistics-analytics vendors advertise case-study figures such as a 62% cost reduction for warehouse-and-logistics deployments (Agrex AI). These are vendor-reported, not independently audited — treat them as indicative and demand a measured before/after in your own pilot.
The INR case: what's recoverable, and the guard-cost offset
Two numbers make the ROI concrete for a plant head: the shrinkage you can claw back, and the guard spend you can either offset or make more effective.
Recoverable shrinkage. Take a plant moving ₹50 crore of material a year. At an indicative 1.5% shrinkage, that's ₹75 lakh/year lost. If cameras address the theft slice (say 55% of it) and a monitored deployment cuts that theft slice by even a third, the indicative recovery is ₹75,00,000 × 0.55 × 0.33 ≈ ₹13.6 lakh/year — from software layered on cameras you likely already own. The exact numbers are yours to measure; the point is that a fraction of a percent of throughput dwarfs the cost of the analytics.
The guard-cost offset. Manned guarding is the default anti-theft control at Indian plants, and it is not cheap once you count statutory costs. A legally compliant unskilled guard runs roughly ₹24,000–₹25,000/month in Delhi including PF, ESI and statutory bonus (Knighthood, 2025); a single round-the-clock post needs ~4–4.5 guards (an industry relief factor) to cover shifts, weekly-offs and leave — so one 24/7 gate or yard post costs on the order of ₹1.0–1.2 lakh/month, ~₹12–14 lakh/year (indicative). Analytics doesn't replace guards, but it lets fewer guards cover more ground: cameras watch the perimeter and blind spots continuously, and guards respond to alerts instead of patrolling blind. Deferring even one additional 24/7 post pays for a multi-camera analytics deployment on its own.
Where to point cameras first (the money follows the exits)
Theft leaves through a small number of choke points. Prioritise:
- Loading/unloading docks — the single highest-risk zone; goods and vehicles meet here.
- Finished-goods and high-value stores — after-hours access and unlogged removal.
- Scrap yard and weighbridge — a classic collusion point (under-weighing, swap-outs).
- Gates and perimeter — tailgating, intrusion, off-schedule vehicle movement.
Getting which camera watches which exit, from which angle right is the hard part — a helmet needs a face, but a dock theft needs the goods, the person, and the vehicle in one frame. This is the gap Mama closes: you record a short phone walkthrough of the floor, docks and yard, and it returns a floor plan plus a camera-placement plan — which zones, which sightlines, which theft events each camera should cover — without waiting on a site survey.
Do it legally: worker video is personal data
Anti-theft monitoring points cameras at your own staff, and that carries a duty. Under India's Digital Personal Data Protection Act, 2023, video of workers is personal data — you need a defined purpose (loss prevention/safety), signage and notice, a retention limit, and restricted access (official text, MeitY). Separately, any new cameras you buy in India from 1 April 2026 must meet BIS/STQC Essential Requirements (ER-01) for security (BIS CRS guidelines, crsbis.in) — so specify compliant hardware for any expansion. Do the compliance groundwork before you deploy, not after an incident.
FAQ
What shrinkage rate should an Indian factory assume? Manufacturing-specific data is scarce, so use retail as a proxy: India's measured retail shrinkage peaked around 2.38% of sales in the Global Retail Theft Barometer, while recent listed retailers report 0.4–0.5%. A 1–2% band of material value is a realistic indicative planning figure for a plant handling loose or high-value stock — measure your own reconciliation gap to get a real number.
Is most factory theft internal or external? Both matter. In Indian retail the split was roughly 45% external (shoplifting) and 23% internal (employee), but on a factory floor the internal share tends to be higher because staff know the blind spots and reconciliation lag. Docks, scrap yards and finished-goods stores are the common collusion points between the two.
Can cameras really recover the money, or just record it? Analytics recovers money by shifting detection from after-the-fact audit to near-real-time alert — an unauthorised removal or after-hours store entry is flagged while it's happening, so it can be stopped and investigated. It addresses the theft slice (roughly half to two-thirds of shrinkage), not administrative/process error, which needs ERP and counting fixes.
Will analytics let me cut security guards? Usually it reallocates rather than eliminates them: cameras cover the perimeter and blind spots continuously, so guards respond to alerts instead of patrolling blind. Given a 24/7 post costs ~₹12–14 lakh/year (indicative), deferring even one added post can fund a multi-camera analytics deployment.
Is it legal to monitor my own workers on camera in India? Yes, for a legitimate purpose like loss prevention, but worker video is personal data under the DPDP Act, 2023 — you need clear signage/notice, a defined purpose, a retention limit, and restricted access. Cameras bought new from 1 April 2026 must also be BIS/STQC ER-compliant.