Retail & Consumer Goods case study
Measuring Product Visibility in Advertising and Retail Displays
A US consumer goods company needed to measure how prominently its products appeared in video advertising and on store shelves.
- Project type
- Computer vision API
- Industry
- Retail & Consumer Goods
- Location
- United States
- Services
- AI Development, API Integration
At a glance
- Detection accuracy
- 97.2%
- Manual auditing costs
- 87%lower
- Delivery
- 5.5months
Overview
The Summary.
A US consumer goods company needed to measure how prominently its products appeared in video advertising and on store shelves. VARTEQ built a Product Recognition API using convolutional neural networks (CNNs) to detect placements automatically in real time. The resulting data helped the company assess marketing activity, allocate ad spending, and plan product positioning.
About the project. A consumer goods company that measures product visibility in advertising and retail displays.
Engagement length. 5.5 months to delivery; development began 3 weeks after initial discussions. Original plan: 6 months.
The challenge.
- Client need: The company wanted automatic measurement of product visibility during ad slots and in retail displays. Manual checks were slow and inconsistent, which weakened campaign analysis.
- Discovery findings: Existing recognition tools did not detect products precisely enough or supply timely data about their visibility and shelf position.
The solution.
- A computer vision API that uses CNN models to analyze images and video and track products in real time.
The system had to process media quickly, support growing workloads, and connect to the client’s analytics platform. Development covered:
- Model training: CNN models were trained to identify products in images and video frames, reaching 97.2% accuracy.
- API and backend: Web API services and TensorFlow handled media analysis and data processing.
- Cloud deployment: GPU-accelerated cloud infrastructure supported the processing workload.
- Business intelligence integration: The API supplied live product visibility reports to the client’s BI system and supported automated reporting workflows.
Services on this project
Technologies used
- TensorFlow
- Convolutional neural networks (CNNs)
- Web APIs
- GPU-accelerated cloud infrastructure
The outcome.
- Development began 3 weeks after initial discussions.
- Planned implementation: 6 months.
- Actual delivery: 5.5 months.
- Brand visibility insights improved by 40%, supporting more precise ad placement.
- Shelf monitoring supported a 25% improvement in in-store product positioning.
- Automation reduced manual auditing costs by 87%.
- Product detection accuracy: 97.2%.
- Processing capacity: more than 100 video frames per second.
- Marketing teams reported 30% more effective advertising campaigns after using the data in their decisions.
| Metric | Before | After |
|---|---|---|
| Product Visibility Tracking | Manual review, prone to errors | Automated, 97.2% accuracy |
| Ad Performance Analysis | Delayed, incomplete insights | Real-time reporting; 40% improvement in ad placement |
| Shelf Placement Monitoring | Spot-checks, no automation | Continuous tracking, 25% better positioning |
| Processing Speed | Slow, manual verification | 100+ video frames analyzed per second |
| Auditing Costs | High labor costs for monitoring | 87% cost reduction via automation |
More Client Stories
All case studiesEdTech
2014
Partner since
Follett Software Development Case Study
Rebuilding Follett's school library and eContent platform and its student data platform, as a long-term engineering partner since 2014.
Read case studyIT Services
4weeks
Delivery
Scheduling, Service Reports, and Timesheets in One Place
An IT service provider was using Excel to coordinate site visits and track work. Technicians submitted reports in different formats, and supervisors could not see field activity as it happened. VARTEQ built a platform for scheduling visits, recording services, and generating timesheets, replacing manual spreadsheet entries with a shared operational record.
Read case studyManufacturing
60%less
Lead engagement time
Turning Permit Records into Timely Sales Outreach
A US kitchen manufacturer used permit activity to find sales opportunities, but manual searches and scattered contact records delayed outreach. VARTEQ built a system that monitors permits, sends records to HubSpot, and triggers SMS, email, and phone follow-up. Lead engagement time fell by 60%, and the number of qualified leads grew by 25%.
Read case study
Start the conversation
Facing a Similar Challenge?
Tell us what you're building or fixing. You'll meet the engineers who would work on it before you sign anything.