Innovating Category Management: Disrupting Traditional Approaches

Differentiate solutions in grocery by using computer vision and AI
Georges Mirza Headshot
Category Management
Computer vision is the critical technology that enables retailers to innovate and enhance their operations.

The retail industry increasingly relies on computer vision and artificial intelligence to address various business challenges, such as identifying on-shelf availability and creating cashierless stores. Computer vision is the critical technology that enables retailers to innovate and enhance their operations. Improving scalability, speed and accuracy can boost customer satisfaction both in-store and online. Furthermore, it has the potential to streamline processes, reduce labor costs, and enhance the overall shopping experience by facilitating product identification and understanding shopper behavior.

Evolution

Various approaches to computer vision have been implemented in the past decade, and most of the effort has focused on building robots that navigate aisles. Some used fixed cameras to collect images for processing, identify on-shelf availability and measure compliance. This has proved challenging, with limited success in reaching accuracy levels and scale deployment. 

For shelf-based solutions, many still rely on handheld devices and labor to capture images of shelf conditions periodically and get a snapshot in time of the shelf. Other uses extend to in-store aisle traffic and monitoring to reduce shrink. Much experimentation is underway on cashierless stores, with several trial locations deployed by Amazon.

Reflection

There’s an abundance of solutions and startups claiming to solve on-shelf availability. The field is growing crowded, and solution viability with a clear ROI is proving to be challenging. As product recognition becomes commoditized with the availability of Google and Azure computer vision services, differentiation is critical for acceptance and growth.

Differentiators

Some solutions have the potential to solve multiple store-level business problems and generate insights accurately, repeatably and reliably at scale. Let’s face it: We shouldn’t need to use the same type of hardware sensors from different vendors to solve various problems at a store. There will be a time when we must simplify to a one-to-many approach, with one sensor solving many issues.

Savvy companies are building solutions with the potential to expand and address several use cases for the retailer. We repeatedly see computer vision at the heart of solutions for these use cases at scale.

Computer vision is the key differentiating solution, from shopper behavior to shelf conditions. Understanding retail store traffic and shopper behavior while complying with the General Data Protection Regulation (GDPR), is a challenging task, and something retailers are highly cautious of when introducing new tracking technologies to their stores. Using computer vision to identify accuracy and validate scanned items at checkout is a $100 billion-a-year inventory problem for retail.

An abundance of images and videos shared in social media has proved to be a great resource to identify trends and generating forecasts, and is the most likely successor to traditional product research methods. This approach is powered by computer vision capabilities as well.

On-shelf availability and planogram compliance solutions are proliferating at the store level, using various methods to capture imagery. Cameras, fixed or moveable, operated by humans or robots, are getting much closer to recognizing on-shelf availability and reporting on planogram compliance deploying computer vision as the engine to make this traditionally laborious and costly activity possible.

Potential

These solutions stand out from the others by successfully disrupting traditional retail approaches, starting with a laser focus on mature computer vision capabilities and planning early on for scale. They then expand their reach at the store level, solving multiple problems based on learnings from the imagery captured.

With these differentiators established, such solutions are positioned to grow the market and expand portfolios under the retail solution umbrella. They will play a key role in advancing the retail computer vision and AI solution field by continuously bringing about real value from innovative approaches. 

X
This ad will auto-close in 10 seconds