In the evolution of retail analytics, the demand has shifted from basic footfall numbers to deeper insights on visitor behaviour, store engagement, and journey mapping. Re-Identification (Re-ID) technology is a significant advancement that meets this demand—offering highly accurate, GDPR-compliant tracking of individuals throughout a store or shopping venue.
This blog explores how Re-ID works, the types of metrics it enables, and how it upholds strict data privacy requirements.
What Is Re-ID Technology?
Re-ID (Re-Identification) is a vision-based AI technique that allows a people counting system to recognise the same person across different camera zones—without using facial recognition, mobile signals, or personally identifiable information.
How It Works
- Person Detection: A visitor is detected using computer vision. Visual Feature Extraction: AI extracts non-biometric attributes like:
- Clothing colour and texture
- Accessories (e.g., bag, hat)
- Body shape and silhouette
- Movement patterns
- Embedding Generation: These attributes are encoded into a unique embedding (an anonymised numerical vector).
- Cross-Camera Matching: Embeddings are compared across cameras to determine if the same individual has been seen before.
- Session ID Assignment: A temporary, anonymous ID is assigned per visit and discarded once the session ends.
⚠️ Important: No images are stored. No biometric or personal data is collected or retained.
Re-ID Metrics and Applications in Retail
By enabling the system to follow individual visitors across cameras, Re-ID unlocks a new level of analytics:
1. Site-Level Dwell Time
- Metric: Average time each visitor spends in-store
- Advantage: Unlike Wi-Fi tracking (sample-based), Re-ID provides full-coverage census-level data
2. Unique Visitor Count (Deduplicated)
- Metric: Accurate daily unique visitors
- How: Identifies returning individuals to avoid double-counting across multiple entries/exits
3. Pass-Through Traffic Filtering
- Metric: Ratio of passers-through vs actual visitors
- Use case: For stores with multiple entrances or mall walkthroughs, this identifies non-engaged traffic
4. Zone Dwell and Journey Mapping
- Metric: Time spent in each store area, including transition paths
- Use case: Measure customer engagement in specific zones (e.g. promotions, fitting rooms)
5. Staff Exclusion (Without Wearables)
- How: Detects repeated behavioural patterns and recognises uniform appearance
- Advantage: No need for badges or manual check-ins; improves conversion accuracy
GDPR Compliance: Built into the Core
Re-ID technology is designed from the ground up to comply with GDPR and global privacy regulations. Unlike facial recognition or biometric tracking, Re-ID does not collect or process any personal or biometric data. Instead, it relies on anonymised appearance-based features—such as clothing color, body silhouette, and accessories—which are encoded into mathematical representations known as embeddings. These embeddings cannot be traced back to an individual and are used solely for the purpose of recognising movement patterns during a visit.
All tracking is session-based: once a visitor leaves the store, their temporary ID is discarded, ensuring that no persistent identifiers are retained. This means Re-ID does not profile, re-identify across days, or store any personally identifiable information. Furthermore, all processing is performed locally on the device, meaning video footage does not leave the site or require cloud transmission. Retailers can optionally enable short-term video recording for operational auditing, but this is strictly under their control and is not necessary for Re-ID functionality.
This privacy-by-design approach ensures that retailers can gain deep behavioural insights into visitor traffic without violating data protection laws or requiring customer consent, making Re-ID a compliant and future-proof solution for physical space analytics.
Re-ID transforms how retailers measure store performance—not just counting how many people came, but understanding who they were (anonymously), what they did, and how long they stayed.
With GDPR-compliant, AI-based technology, retailers can now access:
- True visitor counts
- Customer journeys
- Behavioural segmentation
- Accurate conversion and engagement metrics
For more details on how Re-ID can be integrated into your retail analytics ecosystem, contact our technical team or schedule a live demonstration.
#reidtechnology #retailanalytics #peoplecounting #gdprcompliance #privacymatters #storeanalytics #customerjourney #aiinretail #retailtech #shopperinsights #dataprivacy #smartretail #behaviouralsegmentation #conversionrate #footfallanalytics #storeperformance #anonymoustracking #ai #futureofretail
