Running a counterfeit detection programme
Buying verification software is the easy part. What determines whether counterfeits actually reduce is everything around it: who photographs what, how often, how evidence is handled, and what happens when a distributor is implicated. Most programmes fail operationally rather than technically.
A counterfeit detection programme has five components: a market survey process that samples channel stock systematically, trained field officers with a defined capture routine, disciplined evidence handling, geographic aggregation of verdicts, and a defined escalation path to distributors and legal. The technology answers whether a pack is genuine; the programme decides what happens next.
Where the samples come from
| Source | What it tells you | Frequency |
|---|---|---|
| Planned market surveys | Baseline prevalence by district | Quarterly |
| Field-team spot checks | Continuous low-cost coverage | Every visit |
| Consumer and trade complaints | Where damage is being felt | As received |
| Warranty and service returns | Product that already failed | Continuous |
| Invalid-code scan clusters | Where fakes are circulating | Weekly, if you serialise |
| Competitor and distributor tip-offs | Often the earliest signal | Ad hoc |
That last row is worth taking seriously. Honest distributors lose margin to counterfeit product in their own territory and are frequently the first to know. A programme that treats distributors as suspects rather than allies loses its best sensor network.
Market surveys that produce usable data
- Sample systematically, not opportunistically. Fixed number of outlets per district, chosen the same way each cycle, so quarter-on-quarter numbers mean something.
- Buy the product. Do not ask the retailer for a sample — you want what a customer would actually be sold, and you want a receipt.
- Record the purchase context — outlet, area, price paid, what was said. A price 20% below market is itself evidence.
- Photograph before anything else happens to the pack.
- Keep the physical sample. Photographs support investigation; the object supports a legal case.
- Verify against the correct batch reference, since packaging is revised over time.
Field officer capture discipline
Almost all disappointing verification results trace back to photography rather than analysis. What to standardise:
- A fixed shot list — full face, logo close-up, label and batch panel — so every comparison is like for like.
- Quality gates before capture, with the app refusing to arm the shutter until sharpness, glare and steadiness pass.
- Manual capture. The user composes and presses; auto-shutter fires on frames nobody chose.
- Diffuse light, never direct flash on varnished or laminated packs.
- A clip-on macro lens where print structure matters to your category.
- Never forward images through messaging apps. Re-encoding destroys the forensic layers.
Evidence handling
Treat every verification as potential evidence, because a small proportion will become exactly that:
- Retain every image, untouched, with its capture metadata.
- Store the verdict with its full layer scores and the itemised differences, not just a headline result.
- Record who captured it, when, and where, with device details.
- Keep the physical sample, labelled and linked to the digital record.
- Keep the purchase receipt.
- Preserve the chain of custody from purchase to storage.
A brand-protection team that can produce that package moves much faster with both distributors and lawyers than one arriving with a phone photograph and a recollection.
From incidents to a map
The strategic output is not individual verdicts but geography. Once findings are geolocated, patterns emerge that no single incident reveals:
- Clusters around a specific distributor's territory — the most common and most actionable finding.
- Corridors along transport routes, indicating where the product enters.
- Concentration in price-sensitive districts, which tells you the counterfeit is competing on price rather than deceiving on quality.
- Timing patterns around festive demand peaks, when genuine supply tightens and fakes fill the gap.
- SKU concentration — counterfeiters pick the most profitable, most recognisable items, not the whole range.
If you also run a serialised programme, invalid-code scan clusters give you the same map at far higher density and far lower cost — see QR program fraud prevention. The two datasets corroborate each other well.
Escalation, carefully
- Confirm before accusing. A wrongly accused distributor is a relationship destroyed and, occasionally, a defamation exposure.
- Distinguish complicity from victimhood. Most distributors handling counterfeit stock were sold it, and treating them as partners in the investigation usually produces the source.
- Escalate patterns, not instances. One pack proves nothing about a distributor; a cluster over two quarters does.
- Involve legal early on anything you intend to act on, so evidence is collected the way they need it.
- Close the loop with the field team. Officers who never hear what happened to their findings stop reporting them — the fastest way to kill a programme.
What to measure
| Metric | Why |
|---|---|
| Prevalence by district, quarter on quarter | The only number that shows the programme working |
| Verifications per officer per month | Coverage and engagement |
| Inconclusive rate | Capture discipline — rising means training has slipped |
| Time from finding to action | Operational responsiveness |
| False-positive rate on known-genuine | Field trust |
| Share of findings traced to a source | Whether investigation is actually working |
Frequently asked questions
How do you run a counterfeit market survey?
Sample systematically rather than opportunistically — a fixed number of outlets per district chosen the same way each cycle — buy the product rather than asking for a sample, record outlet, area and price paid, photograph before anything else happens to the pack, and keep both the physical sample and the receipt.
What is the most common cause of poor verification results?
Photography, not analysis. A fixed shot list, quality gates that refuse to arm the shutter until sharpness and glare pass, manual rather than automatic capture, diffuse light on varnished packs, and never forwarding images through messaging apps together resolve most of it.
How should counterfeit evidence be handled?
Retain every image untouched with capture metadata, store the verdict with full layer scores and itemised differences, record who captured it and where, keep the physical sample labelled and linked to the digital record, keep the purchase receipt, and preserve chain of custody.
What patterns should a counterfeit map reveal?
Clusters around a specific distributor's territory, corridors along transport routes showing entry points, concentration in price-sensitive districts, timing spikes around festive demand when genuine supply tightens, and concentration on the most profitable and recognisable SKUs rather than the whole range.
Should distributors be treated as suspects?
Usually not. Most distributors handling counterfeit stock were sold it themselves and lose margin to it in their own territory, which makes them the earliest and best sensor network. Treating them as partners in the investigation typically produces the source; treating them as suspects loses the network.
What metric shows a brand protection programme is working?
Counterfeit prevalence by district, measured quarter on quarter from a systematically sampled market survey. Verification counts, findings and enforcement actions all measure activity; only prevalence measures outcome. Track false positives alongside it, because field trust determines whether the programme continues.