Owl · Brand Protection

Finding counterfeits
is finding a needle
in a very large haystack.

Millions of new listings go live every month across Shopee, Lazada, and TikTok Shop. A handful are fakes of your product. Manual detection finds maybe 1%. Owl finds them all.

Book a demo See how it works

The problem

The haystack is enormous.
And it grows every minute.

$500B+

Annual global cost of counterfeiting

U.S. Chamber of Commerce, 2024

$185B

ASEAN ecommerce GMV in 2025

Projected to exceed $1T by 2030

52%

ASEAN GMV on Shopee alone

Shopee · Lazada · TikTok Shop — three platforms to monitor

At manual speed

Listings checked per hour ~30
10,000 listings reviewed in 333 hours

With Owl

Listings scanned per minute 1,000s
10,000 listings reviewed in < 10 minutes

Manual speed assumes 2 minutes per listing for a trained analyst.

The solution

Three automated steps replace manual brand protection

Owl runs the full detection pipeline automatically — from scraping to risk score to takedown export — without a human in the loop until the list is ready to act on.

Collect -->
01 · COLLECT
Collect
Keyword intelligence and automated scraping across Shopee, Lazada, and TikTok Shop. Up to 5 brand keywords per cycle, refreshed weekly. Anti-bot infrastructure operating since 2020.
→
02 · DETECT
Detect
The Owl Assessment Engine analyses every listing using four signals: price anomaly, description heuristics, review sentiment ('palsu', 'kw', 'fake'), and merchant profile. Each signal feeds a risk score.
→
Act -->
03 · ACT
Act
Risk-scored listings triaged by priority. Takedown-ready exports for platform reports, legal proceedings, and enforcement workflows. Your team reviews confirmed cases only — not the full haystack.

Multi-signal scoring

A keyword alert sees the name.
Owl sees all four signals.

A listing using your brand name is not necessarily counterfeit. But a listing using your name, priced 62% below your floor, with three reviews containing 'palsu', from a seller with no certification — that is a different signal entirely.

Price signal

Anomalous pricing vs. authenticated price bands per marketplace

Description heuristics

Missing warranty claims, non-original language, vague product descriptions

Review sentiment

Scans for 'fake', 'peke', 'palsu', 'kw' across Bahasa, Filipino, Thai, Vietnamese

Merchant profiling

Seller badges, reputation scores, trusted seller status, account age

Case study · Live deployment

Confidential manufacturing brand

Automated scanning of Shopee, Lazada, and TikTok Shop to identify unauthorized sellers and counterfeit industrial parts for a Japanese manufacturing brand with no prior ecommerce brand protection capability.

77% flagged

916 of 1,189 listings

916

Suspected counterfeit listings identified

1,189

Total products scanned across 3 platforms

99.5%

Platform uptime across all scan cycles

4 wk

Requirements to go-live onboarding time

The difference

Manual processes vs. Owl

Dimension Manual Owl
Detection speed Hours per day Minutes per scan
Scale 100s per week 1,000s per month
Accuracy Human error prone AI precision, brand-tuned
Language coverage Team language only Bahasa, Filipino, Thai, Vietnamese
Signals checked Visual inspection 4 signals per listing
Cost per takedown High Lower enforcement cost

Why Owl

Built on infrastructure that's hard to replicate

Shopee's anti-scraping infrastructure is among the most aggressive in global ecommerce. We've been operating a reliable, high-volume pipeline through it since 2020. Owl inherits every year of that engineering investment.

Six years of Shopee scraping

Anti-bot defenses, rate limits, IP rotation, API schema changes — handled continuously since 2020. The infrastructure that breaks competitors is the infrastructure Owl runs on.

Weekly scanning, 24/7 uptime

Standard cadence is weekly deep scraping. 99.5% platform uptime in live deployment. For high-risk categories or active enforcement periods, frequency can be increased.

Southeast Asia language model

Purpose-built for ASEAN. Counterfeit signals in Bahasa Indonesia ('palsu', 'kw'), Filipino ('peke'), Thai, and Vietnamese — the languages counterfeiters actually write in.

Battle-tested with real brands

Live deployment with a confidential manufacturing brand. Part of the Revenue Scaler group serving Panasonic, Nestlé, Kraft Heinz, and 30+ FMCG brands across Southeast Asia.

Deployment

Live in 4 weeks

From your first conversation to first enforcement export. No existing ecommerce brand protection setup required.

1

Requirements

Week 1

Brand data, SKU catalogue, and authenticated price bands per marketplace. This is all we need to calibrate the engine to your products.

2

Model training

Week 2

AI calibration with your products. The engine learns your authentic price bands, typical seller profiles, and description patterns to minimise false positives.

3

Testing

Week 3

Accuracy validation, false positive reduction, threshold calibration. Your team reviews early results and we adjust scoring before go-live.

4

Go-live

Week 4

Dashboard activation, team training, first enforcement export. Weekly scanning begins. Your team receives prioritised takedown lists — not the haystack.

Protect your brand

The haystack is too big
to search by hand.

Contact us to discuss how Owl can protect your brand portfolio across Southeast Asia's marketplaces. Live in 4 weeks.

Book a demo

FAQ

Common questions

Which platforms does Owl monitor? +

Owl currently monitors Shopee, Lazada, and TikTok Shop. Tokopedia and Blibli are on the roadmap.

Does Owl handle takedown requests automatically? +

No. Owl surfaces and prioritises listings for your enforcement team. It exports takedown-ready packages for platform reports and legal proceedings — the submission is your team's action.

How often does Owl scan? +

Weekly deep scraping is the standard cadence. For high-risk categories or active enforcement campaigns, frequency can be increased.

What makes Owl more accurate than a keyword alert service? +

A keyword alert sees the brand name. Owl sees four signals simultaneously — price anomaly, description heuristics, review sentiment, and merchant profile. A listing that passes one signal often fails three others. That combination is what separates a counterfeit from a legitimate reseller.