Photo by Alex Knight on Unsplash
How Will Computer Vision Change the Retail Experience?
Artificial intelligence (AI) has become a buzzword in the retail industry, and retailers like Sam’s Club and Home Depot are stepping up their game, particularly in one area of AI: computer vision.
Computer vision is the technology that allows computers and systems to understand visual data from images, processing it to derive insights. According to Statista, the market size for computer vision is projected to reach a whopping $26.26 billion by 2024, with the retail sector holding a 4.55% share of this market in 2022.
At its core, computer vision is a specialized field within the vast realm of artificial intelligence. It grants machines the ability to interpret and make decisions based on visual data, transforming raw information into actionable insights. It’s akin to giving retailers a sixth sense, rooted not in intuition but in empirical data.
By emulating human visual perception, it can capture, process, and analyze images and videos with astonishing accuracy. It goes beyond mere object recognition; it can decipher emotional states, facial expressions, and even body movements. For marketing teams and sellers, these insights are invaluable.
From the moment a customer sets foot in a retail store, their every move becomes part of this intricate tapestry of data. Every interaction with products, every second spent in aisles, even facial expressions — all are ripe for analysis.
One notable player in this field is Sam’s Club, which recently unveiled a pilot project spanning 10 locations to deploy AI and computer vision technology. This innovative application aims to streamline the checkout process by eliminating queues at exit areas. Traditionally, customers had to wait for a Sam’s Club employee to review their receipt before exiting, causing frustration among shoppers.
In these pilot stores, a combination of computer vision and digital technology at the exit captures images of shopping carts and verifies payment for all items in a member’s basket. This not only speeds up the process but also enhances the overall shopping experience. Developed in-house over the course of a year, this technology reflects Sam’s Club’s commitment to addressing customer pain points.
The benefits are already evident, with a 20% improvement in overall satisfaction scores at the pilot locations. Moreover, more than half of the members are experiencing quicker exits, while employees have more opportunities to engage with customers throughout the store.
Similarly, sports stadiums like Lincoln Financial Field and Lumen Field are leveraging computer vision to enhance the fan experience. By deploying cashierless express stores and checkout-free concession stands, these venues are revolutionizing retail operations and cutting down wait times for fans.
At Home Depot, computer vision is being utilized to optimize inventory management. Through a mobile app called Sidekick, store associates can prioritize restocking tasks based on real-time data analysis of shelf inventory. This ensures that customers always find what they need, leading to higher satisfaction rates and increased productivity.
Additionally, retailers like SpartanNash are deploying autonomous inventory robots equipped with computer vision technology. These robots traverse store aisles, scanning shelves to track inventory levels and identify out-of-stock items or misplaced products. This not only improves in-stock rates but also frees up employees to focus on providing better customer service.
This technology extends to various aspects of retail operations, reducing the need for extensive human staffing. Retail heat map analytics monitor customer movements within the store, pinpointing high-traffic areas and bottlenecks. Queue management becomes more efficient, with computer vision-powered cameras predicting queue lengths and waiting times.
Traditional checkout systems, often plagued by issues such as long lines and manual scanning, are also being replaced by automated, cashierless stores like Amazon Go. Using computer vision alongside sensor fusion and deep learning algorithms, Amazon has introduced the “Just Walk Out” shopping concept. Customers enter the store, scan their smartphones, and freely select items from the shelves. Cameras with computer vision technology track their movements, seamlessly adding items to their virtual carts.
In the realm of security, computer vision adds a layer of protection against theft and fraud. Machine learning technology monitors behavior patterns, identifying potential instances of shoplifting. Retailers like Walmart use computer vision-powered cameras to detect suspicious behavior and unscanned items at self-checkout lanes.
By tracking customer behavior and shopping patterns, computer vision also enables personalized marketing strategies. Retailers can create detailed consumer profiles and send targeted offers based on individual preferences, fostering loyalty among customers.
Discussion Questions
How can the retail industry maximize computer vision’s impact, moving beyond traditional uses like inventory and checkout, to enhance customer experiences in-store and online while addressing privacy concerns?
As retailers adopt computer vision to better understand customer behavior, how can they ensure transparency in data usage while maintaining customer trust and loyalty, especially amidst heightened privacy concerns and regulations?
Given AI and computer vision’s rapid advancements, how can retail leaders integrate these technologies effectively, ensuring they align with consumer expectations and deliver ROI while also fostering a culture of innovation and digital transformation within their organizations?
Poll
Loading …BrainTrust
John Hennessy
Retail and Brand Technology Tailor
Mark Self
President and CEO, Vector Textiles
I think this has some solid applications, if only because it is a more sophisticated version of technology that already exists such as customer tracking. That said, in areas like automated checkout and stock checking it is going to compete with other solutions like RFID. The winner will be determined by effectiveness and the cost and simplicity to deploy.
There are, as ever, potential issues that need to be worked through. The first is cost and the return on investment. The second is privacy concerns if the technology is being used to scan and track customers. The third is accuracy: you can imagine the problems that would arise if shoppers start being flagged as potential shoplifters when they are nothing of the sort.
Computer vision seems promising, helping retailers with availability, loss prevention, store design and staffing. Consumers gain speed, ease, in-stock items and personalized communications.
Next steps could include integrating all retail systems for a seamless experience: computer vision, loyalty data, retail media and store tech like digital screens. Computer vision could also monitor employee activity to spot opportunities to redeploy workers as consumer traffic fluctuates throughout the day.
Analyzing data as granular as shoppers’ facial expressions will raise questions and eyebrows. Retailers can communicate exactly what data they’re collecting, who has access to it and how it will be used to alleviate consumers’ privacy concerns.
Wait, so we like Self-Checkout again? This is so confusing!
I hope Computer Vision captures my frustration at seeing 10 cashiers and no one in the aisles to help me at “any” DIY hardware store. DIY really means DIYWY – Do It Yourself With YouTube! And maybe Computer Vision will catch my frustration with Carvana’s Sebastian, a chat service in bad need of any kind of intelligence.
Technology that isn’t deployed for technology’s sake but that actually addresses a problem can be profitable. Bring on the self-checkout, but please use the savings to employ more people in the aisles (but choose people that actually like people – steal them from In-N-Out or Trader Joe’s if you have to).
Sam’s Club implementation appears to be a great example of tech improving the customer experience. It sounds like SpartanNash is onto something too. I’m sure more than a few of us have had to search endlessly for the elusive employee just to ask them to “run to the back” and see if they have the item in stock. (There’s already tech to solve that problem but let’s not ruin Computer Vision’s parade just yet.) The good news is that the “runner” can now focus on helping people and when people buy from people they buy more (higher AOVs, higher conversion rates).
Unfortunately, Computer Vision is all for naught! One of my startups is ready to deploy the killer of all retail tech – YBWITYALI (pronounced You Be Witty Ali). Okay, so the name needs a little work but it stands for “You’ll Buy What I Tell You And Like It.” People will receive targeted subliminal messages while they’re still in the parking lot guaranteeing ridiculous amounts of purchases. It’s kinda like eating Goldfish crackers but for the brain – people will just inexplicably want more!!! Stay tuned!
John,
Great commentary this morning. Both with and without tongue in cheek. That is a good point about companies using tech for tech’s sake and not for solving challenges.
A Ferrari in LA rush hour is moving as fast as any other car! Sometimes, it’s just not the tech.
Thanks, Gene!
There was nothing I saw mentioned that really seems likely to raise privacy concerns. Alternately, I saw a lot of “hope for”s that will likely take a long time to perfect…if ever. Improved security seems like an obvious area – largely b/c of how much publicity shrink has received lately – it’s not just Amazon, “Just walk out” has been introduced everywhere!! – but how well it will work remains to be seen (no pun intended). Which means, by default, inventory management seems destined for the best results (at least of the areas mentioned). Regardless, I expect incremental improvements, largely unnoticed, and after the hoopla has died down…just like most any other technology, ever.
As an aside, the new Beijing airport is entirely biometric/facial recognition. Not just for checkouts at the shops and restaurantsbut also for ticket dispensing, bag checking, security passes through, and boarding.
Many use cases here and we are obviously in the early stages. These applications will just keep getting stronger and more user friendly. Cost/ROI will put the brakes on with regard to quick rollouts, however interest and innovation will continue here. Inventory will be the first to be improved, then security. I am eager to see what happens next here!!
Today’s imagination can’t even see where this is going.
While at Alert Innovation, I was fortunate to have many interactions with our talented vision team. It’s a fascinating and complex discipline that integrates a variety of other disciplines. At its core it adds another sense to your data collection making it easier to have a more complete picture (pun intended) and reach accurate conclusions more quickly. I can’t talk about much of what was shared at Alert, but I can say that vision technology is in its early days. Every aspect of computer vision is rapidly improving as the technology gets better, practitioners gain more experience and and as more talented engineers enter the field expanding use cases.
Like any new technology, not everything is going to work perfectly in the beginning. But as this evolves, the tools will continue to be refined and add additional value in all of the areas mentioned. There are real use cases here!
The sophistication and accuracy of computer vision has accelerated in the past 5 years. Cashierless checkout technology has become more cost effective and is now showing positive ROIs for some retail applications. Robotic inventory taking has proven successful and computer vision to check the accuracy of self-checkout is helping to reduce errors and theft. There are many other use cases for computer vision and it will be exciting to see what becomes a reality in the next 5 years.
The key to effective “decisioning” is line of sight. This is what makes Computer Vision (or image recognition so powerful). From visual monitoring inventory levels to evaluating the patterns in foot traffic (like air traffic control monitoring) to understand the human response by way of body language. There’s nothing like “blind spots” for what the human can’t see or discern. CV has already made inroads into creative testing with companies like Realeyesit or creative performance optimization with Xpln.ai or audience supply path optimization like Lumen/TVision.