April 2, 2025
Mojahid_Mottakin/Depositphotos.com
Can Videotaping Store Shelves Reduce Out-of-Stocks?
Instacart is introducing a program to pay “shoppers,” or its gig delivery workers, to take videos of store shelves to reduce out-of-stocks and the need for substitutions.
In a blog entry, Daniel Danker, Instacart’s chief product officer, noted that a single grocery store carries over 30,000 products and pointed to a study from ECR Retail Loss Group that found that as much as 60% of inventory records are inaccurate. He stated, “Traditional inventory checks are slow, manual, and prone to error, leading to out-of-stocks that frustrate customers and hurt sales. On the flip side, excess inventory leads to inefficient use of retail space and waste.”
Danker added that while customers are satisfied with 95% of shoppers’ substitutions, “our goal is to make substitutions unnecessary in the first place.”
Under the “Store View” program, Instacart’s shoppers will be offered a “new type of earning opportunity” involving taking videos of store shelves “one aisle at a time.” Utilizing AI and computer vision, the videos will identify products that are out of stock or running low in quantities.
Store View will further support Instacart’s predictive models to know when certain items are likely to be out of stock at a particular store. Danker wrote, ”If you place an order at night for early next-day delivery, we can predict whether the item is likely to be back on the shelves when the store reopens, and then verify that information in the morning.”
The information will be combined with the over 10 million data points collected daily by Instacart’s approximately 600,000 shoppers to gain a “more accurate, real-time understanding” of inventory levels across its grocery partners.
Store View will launch with select retailers in the coming weeks, with further rollout planned throughout 2025.
Instacart is additionally introducing a “Second Store Check” program to also reduce substitutions. If a shopper finds an item out of stock at the store they’re shopping, a second shopper at a nearby store will check if it’s available and deliver it to the customer “for no additional service or delivery fee,” according to The Verge. Second Store Check will roll out across the Instacart marketplace in the coming months.
Details weren’t available on how shoppers would be compensated to support Store View or Second Store Check. One shopper in Pennsylvania in a trial of Store View told Business Insider that she was paid $12 to record a display of Dove body care products with the task taking about 10 minutes to complete, a rate higher than handling a customer delivery that can take an hour to complete.
Danker also said that Caper Carts, Instacart’s screen-equipped smart carts that are rolling out at grocers, already use advanced sensor fusion and edge AI for item recognition and location-based recommendations and will soon add outward-facing cameras to provide “inventory insights in some stores as frequently as every hour throughout the day.”
BrainTrust
Brad Halverson
Principal, Clearbrand CX
Shep Hyken
Chief Amazement Officer, Shepard Presentations, LLC
Allison McCabe
Director Retail Technology, enVista
Recent Discussions
Discussion Questions
What do you think of the potential of gig delivery workers video recording grocery store aisles to reduce the need for substitutions in grocery delivery?
Do you see Store View and Second Store Check as workable at the store level and offering enough of a benefit to adequately compensate shoppers?
Poll
Loading …
Given the products at grocery stores are fast moving (heck, they’re called FMCG – fast moving consumer goods) I am not completely sure how accurate taking videos is. It’s not exactly real time. Nor does it capture what is in-stock behind the scenes. Sure, it’s partially helpful for alerting when something may be getting low, but it’s not a substitution for real time inventory data, driven by something like RFID.
It is an interesting concept, but it is not the Holy Grail. Having Instacart shoppers video shelves may help reduce substitutions by adjusting inventory available on the app, but it doesn’t solve the real problem – out of stocks. Retailers need to move to real-time inventory tracking, which has been a goal for many years, but is expensive to achieve. Camera vision on shelves or on robots scanning aisles may be a less expensive option for identifying out of stocks, but eventually we need to achieve real-time inventory tracking.
Video? Why not just put sensors on the shelves? The “data” the sensors share can give instant insights to inventory, what’s hot, what’s not, and much more.
There’s an idea here, but the time and steps to video capture, interpret and make future assumptions about products on shelf could take too long or not really be relevant for when the customer is ready. Grocery item turns vary, as do pricing changes, competitive offerings, with variables in in ordering and receiving from distributors/wholesaler.
The alternatives to help customers make educated decisions are in real-time inventory programs increasingly being launched at store level, as are other available tech monitoring features.
Having Instacart shoppers take videos of store shelves doesn’t seem like it will be much help. As Neil pointed out, the info won’t be completely accurate. And while I like that Instacart shoppers will earn extra money taking the videos, as a customer, I want the shopper focused on my order.
Second Store Check is a nice customer perk, but I have to wonder how cost effective it will be.
I also wonder about whether lots of people taking videos becomes annoying for regular shoppers. Having pickers in store is already a bit disruptive. Having people trying to take videos may add to the annoyance.
Plus, if I can make more taking videos than I can make filling Instacart orders, I’m making videos.
Once customers complain to store managers about encountering this activity 2-3 times a visit, in waiting around to shop, grocers will frown on the practice. At least the customer-centric stores will.
Many things can work, but none of them will work, unless there’s follow thru; that’s always been the weak link in ideas we see floated about (both here on RW and elsewhere) and while I don’t want to be a killjoy, I still have to ask “would it be different in this case?”
Everyone asks, “Can we?” and no one asks, “Should we?”- Dr Ian Malcom, Jurassic Park.
Wait! You mean if I take a video of a fire it will put the fire out? Brilliant!
If 95% of customers are satisfied with substitutions and the goal is to make substitutions unnecessary then it sounds like there’s probably “time” to address the root cause. However, I understand that addressing the root cause of this problem is probably neither easy nor quick so a Little Dutch Boy magic is probably not a bad way to buy more time. Unfortunately, the video itself is not the solution; it’s just a finger in the dike.
The Store View and Second Store Check provide innovative solutions in contrast to traditional store methods, which often rely solely on signage and staff assistance.
The use of these technologies enhances the shopping experience by providing real-time inventory updates and personalized recommendations. In the modern retail environment, they are valuable tools because they can improve customer satisfaction and potentially increase sales.
In spite of this, retailers may face a number of challenges when implementing these technologies. Integration of advanced systems and training staff to use them effectively can be costly initially. Furthermore, these technologies often require the collection and processing of customer data, resulting in concerns about data privacy and security.
I like the idea of using people who are frequently in aisles to create new data that helps address out of stocks. There remains the issue of linking this new, analyzed video data with a system that will, 1) verify inventory in store available to restock, 2) re-order if needed, 3) get the shelf restocked and 4) clear the out-of-stock condition in whatever system triggered it.
Interesting that Instacart is working to solve the out-of-stock problem for its online customers. Stores live with it.
Instacart quotes a 95% shopper satisfaction with substitutions. They don’t talk about how many orders are incomplete due to out of stocks and no substitution. That second number has to be driving their interest in solving this problem.
At least someone’s trying to keep the shelves full, shoppers happy and capture the full value of the order as ordered.
There are certainly far more efficient options out there than translating a picture with AI. A video is outdated the moment its taken.
Bizarrely convoluted and inefficient.
This feels like a hammer looking for a nail. In my time, I’ve seen many iterations of shelf sensors that track shelf capacity and the amount of inventory on the shelf in real time. They have all failed for a number of reasons, most importantly, ROI. These solutions, when rolled to all stores and shelves, even in a moderately sized chain, are expensive and only address one, relatively small piece of the out-of-stock issue, in-store fulfillment. As noted by Neil and others in the comments below, this doesn’t do anything to address available inventory, supply chain issues, or forecast inaccuracy, to name a few things.
The other issue, from an adoption perspective, is that this issue can be easily handled by simply implementing processes in the store to have team members walk the aisles from time to time. Today, between to normal flow process that most fast-moving items demand and the increase in store team members or professional shoppers in the store aisles for the various omnichannel fulfillment strategies, there are a number of ways to monitor shelf stock without deploying technology. Which leads back to ROI. If a store is having significant issues with out-of-stocks, instore fulfillment is one of the easier issues to address. In reality, that’s usually only a small part of the problem.
This is only the technology portion of a larger process. The retailer must be committed to ‘react’ to the signal(s) that items are out of stock and to refilling those ‘holes’ as quickly as possible. The technology alone is not going to fix the out-of-stock problems.
While I am sure the intentions are good, this seems like far too slow of a process with too many steps to be effective. Addressing out of stocks is the real issue that needs to be solved here.
More Discussions
Would an Amazon Shift of Prime Day to June Be a Good Move?
Would it be a wise move for Amazon to move Prime Day ahead of its current…Should Amazon Be Banning AI Shopping Agents From its Platform?
Amazon is clamping down on access to its platform from competing AI agents.Does Macy’s Own Celebrations?
Macy’s has made a name for itself in terms of high-profile celebrations. Will that drive cross-over…
Given the products at grocery stores are fast moving (heck, they’re called FMCG – fast moving consumer goods) I am not completely sure how accurate taking videos is. It’s not exactly real time. Nor does it capture what is in-stock behind the scenes. Sure, it’s partially helpful for alerting when something may be getting low, but it’s not a substitution for real time inventory data, driven by something like RFID.
It is an interesting concept, but it is not the Holy Grail. Having Instacart shoppers video shelves may help reduce substitutions by adjusting inventory available on the app, but it doesn’t solve the real problem – out of stocks. Retailers need to move to real-time inventory tracking, which has been a goal for many years, but is expensive to achieve. Camera vision on shelves or on robots scanning aisles may be a less expensive option for identifying out of stocks, but eventually we need to achieve real-time inventory tracking.
Video? Why not just put sensors on the shelves? The “data” the sensors share can give instant insights to inventory, what’s hot, what’s not, and much more.
There’s an idea here, but the time and steps to video capture, interpret and make future assumptions about products on shelf could take too long or not really be relevant for when the customer is ready. Grocery item turns vary, as do pricing changes, competitive offerings, with variables in in ordering and receiving from distributors/wholesaler.
The alternatives to help customers make educated decisions are in real-time inventory programs increasingly being launched at store level, as are other available tech monitoring features.
Having Instacart shoppers take videos of store shelves doesn’t seem like it will be much help. As Neil pointed out, the info won’t be completely accurate. And while I like that Instacart shoppers will earn extra money taking the videos, as a customer, I want the shopper focused on my order.
Second Store Check is a nice customer perk, but I have to wonder how cost effective it will be.
I also wonder about whether lots of people taking videos becomes annoying for regular shoppers. Having pickers in store is already a bit disruptive. Having people trying to take videos may add to the annoyance.
Plus, if I can make more taking videos than I can make filling Instacart orders, I’m making videos.
Once customers complain to store managers about encountering this activity 2-3 times a visit, in waiting around to shop, grocers will frown on the practice. At least the customer-centric stores will.
Many things can work, but none of them will work, unless there’s follow thru; that’s always been the weak link in ideas we see floated about (both here on RW and elsewhere) and while I don’t want to be a killjoy, I still have to ask “would it be different in this case?”
Everyone asks, “Can we?” and no one asks, “Should we?”- Dr Ian Malcom, Jurassic Park.
Wait! You mean if I take a video of a fire it will put the fire out? Brilliant!
If 95% of customers are satisfied with substitutions and the goal is to make substitutions unnecessary then it sounds like there’s probably “time” to address the root cause. However, I understand that addressing the root cause of this problem is probably neither easy nor quick so a Little Dutch Boy magic is probably not a bad way to buy more time. Unfortunately, the video itself is not the solution; it’s just a finger in the dike.
The Store View and Second Store Check provide innovative solutions in contrast to traditional store methods, which often rely solely on signage and staff assistance.
The use of these technologies enhances the shopping experience by providing real-time inventory updates and personalized recommendations. In the modern retail environment, they are valuable tools because they can improve customer satisfaction and potentially increase sales.
In spite of this, retailers may face a number of challenges when implementing these technologies. Integration of advanced systems and training staff to use them effectively can be costly initially. Furthermore, these technologies often require the collection and processing of customer data, resulting in concerns about data privacy and security.
I like the idea of using people who are frequently in aisles to create new data that helps address out of stocks. There remains the issue of linking this new, analyzed video data with a system that will, 1) verify inventory in store available to restock, 2) re-order if needed, 3) get the shelf restocked and 4) clear the out-of-stock condition in whatever system triggered it.
Interesting that Instacart is working to solve the out-of-stock problem for its online customers. Stores live with it.
Instacart quotes a 95% shopper satisfaction with substitutions. They don’t talk about how many orders are incomplete due to out of stocks and no substitution. That second number has to be driving their interest in solving this problem.
At least someone’s trying to keep the shelves full, shoppers happy and capture the full value of the order as ordered.
There are certainly far more efficient options out there than translating a picture with AI. A video is outdated the moment its taken.
Bizarrely convoluted and inefficient.
This feels like a hammer looking for a nail. In my time, I’ve seen many iterations of shelf sensors that track shelf capacity and the amount of inventory on the shelf in real time. They have all failed for a number of reasons, most importantly, ROI. These solutions, when rolled to all stores and shelves, even in a moderately sized chain, are expensive and only address one, relatively small piece of the out-of-stock issue, in-store fulfillment. As noted by Neil and others in the comments below, this doesn’t do anything to address available inventory, supply chain issues, or forecast inaccuracy, to name a few things.
The other issue, from an adoption perspective, is that this issue can be easily handled by simply implementing processes in the store to have team members walk the aisles from time to time. Today, between to normal flow process that most fast-moving items demand and the increase in store team members or professional shoppers in the store aisles for the various omnichannel fulfillment strategies, there are a number of ways to monitor shelf stock without deploying technology. Which leads back to ROI. If a store is having significant issues with out-of-stocks, instore fulfillment is one of the easier issues to address. In reality, that’s usually only a small part of the problem.
This is only the technology portion of a larger process. The retailer must be committed to ‘react’ to the signal(s) that items are out of stock and to refilling those ‘holes’ as quickly as possible. The technology alone is not going to fix the out-of-stock problems.
While I am sure the intentions are good, this seems like far too slow of a process with too many steps to be effective. Addressing out of stocks is the real issue that needs to be solved here.