segments

The Automated Supermarket

by Bill Bishop

Next >>

Recommend This Post 632

Automated supermarketJohn Lert’s vision of the automated supermarket is a concept that we think has the potential to reinvent the grocery store, reduce the cost of operations, and double sales per square foot thanks to increased customer appeal.

In the first installment of this two-part interview, Lert describes the store from the shopper’s point of view. In Part 2, he talks with us about how retailers could convert existing stores, why it would be a good investment, and how this concept differs from the multi-channel model that combines conventional self-service stores with an ecommerce component.

BMC: Describe your concept for the automated grocery store.

LERT: In an automated store, packaged goods will be picked robotically in response to electronic orders placed by customers via personal digital devices, either in-store or online. Customers can self-select their fresh goods (or choose to have these picked also). When finished, customers will make payment, typically via smartphone and without any other checkout procedure, then leave the store and receive their orders at drive-through pick-up stations. For an additional fee, a home-delivery option will be available for all orders, whether generated online or by a store visit, most likely provided by a third party (like Instacart).

What motivated you to apply your talent as an inventor to designing what you hope will be the first successful automated supermarket?

It began with a thought experiment almost 20 years ago: How would the “perfect” supermarket work for the customer? I started imagining a store in which items are picked not by customers, but by machines in response to customers’ electronic orders – even in real-time while the customers are in the store. The checkout process would be eliminated, and customers could order online before going to the store if they wanted, especially packaged goods. I’ve always felt that most customers prefer to self-select their fresh goods, so that option would have to be available.

This was really just a pleasant fantasy about customer experience until I started thinking about how much more efficient and effective the operation of such an automated store would be, and then it started looking like the perfect store for the retailer as well. I came to believe that this concept could have major business potential if an affordable and practical automated order-picking technology could be invented, and I spent a great deal of time thinking about that very challenging problem. I began focusing on mobile robots as a solution more than a decade ago and am now convinced it’s solvable. Meanwhile, the extraordinary development of the internet has made the automated-store concept even more compelling than I originally imagined.

At BMC we put a lot of emphasis on understanding what shoppers want and are looking for. What will the automated store do for shoppers that they’re not getting today in terms of value for money, time-saving opportunities, and other benefits?

Initially, I expect operators of automated stores will price at parity with self-service stores, but automated stores will represent a far greater value to the customer because of the superior shopping experience. Over time, as more automated stores appear and compete with each other, I expect prices will drop because retailers will be able to generate acceptable profit margins even at lower retail prices.

For customers, I believe this model combines the best aspects of both brick-and-mortar and online retailing in a manner and to a degree not possible today. It offers the time savings, convenience, schedule-flexibility, and information-richness of online shopping, while also providing the immediacy of fulfillment, personal service, and opportunity to examine and select products that only brick-and-mortar stores can deliver. It will also enable brick-and-mortar retailers to transcend the physical constraints of the store and offer customers the vastly larger product assortments that are economically feasible only with online retailing.

If you had to choose just one, what is the most important benefit you see for shoppers?

The automated grocery store will make it much easier for customers to fit grocery shopping into their busy lives, because time efficiencies and modality choices will provide a degree of flexibility that is simply not available today. On any given shopping occasion the customer can choose to shop entirely in-store, entirely online, or a combination of both. Similarly, the customer can choose to collect the order at a drive-through pick-up station or have it delivered from the store to the home, whichever is more convenient at the time.

A scenario I like to imagine is a mother who needs to shop on a day when she also wants to attend her son’s afternoon baseball game. So she orders her packaged-goods the night before as usual, but she visits the store in the morning to select her fresh goods before work and leaves the entire order at the store during the day. After the game she picks up the order on the way home, or if the game runs late, she can decide to have it delivered instead. In a very real sense, this degree of flexibility will represent the “mass-customization” of the shopping experience: The store will work the way each customer needs it to work at any given moment.

How do you expect the introduction of the automated supermarket to impact the grocery shopping patterns of customers?

The most important behavioral shift I expect is what I call “bimodal shopping”: Shoppers will typically find it easier and faster to order their packaged goods online prior to the store visit – especially the night before – but will prefer to self-select their fresh goods at the store.

The good news is that bimodal shopping will be highly beneficial to retailers in two ways. First, it will improve capital efficiency by mitigating the previously intractable problem of peaky demand. In a self-service store, the front-end checkout lanes are a critical bottleneck limiting the store’s service capacity during peak hours. In the automated store, there is no front end.

The automation system is a potential bottleneck, but bimodal shopping effectively decouples order-fulfillment from the store visit, so that the item-picking workload can be shifted to off-peak hours. The system will enjoy a much more efficient capacity utilization than checkout lanes (which are idle most of the time), so fewer robots (and hence less capital) will be required than if all demand had to be satisfied in real time. Pick-up stations could also be a potential bottleneck, but since they are exterior to the store and relatively inexpensive, their number should be highly scalable. As a result of all these factors, automated stores will have a much greater service capacity than self-service stores of comparable size.

Second, bimodal shopping will enable virtual expansion of assortment by retailers, because items ordered with sufficient lead-time (the night before a store visit, for instance) can be picked at a remote facility and delivered to the store prior to the customer visit. It will be economically feasible to stock many times more slow-moving SKUs in a single facility covering a large trading area than would ever be practicable in any single store. I anticipate that over time this will enable retailers to operate smaller stores, and also to expand their in-store assortments into more categories by stocking only faster movers within each category and relying on virtual inventory stocked at the remote facilities to satisfy customer demand for slower-moving SKUs.

I’ve heard you talk about how this concept would change the 'shape' of grocery stores as we know them. What do you mean by this?

One important implication of all of these changes is that packaged-goods assortment will become a much less significant point of competitive differentiation between retailers operating automated stores. Other than prices and proximity, the dominant factors driving store-preferences among customers will then be the quality of execution within the fresh market and personal service. Both are fundamental, value-added components of supermarket retailing.

In other words, the automated store model will tend to level the playing field with respect to what we now call center-store and focus competitive differentiation on customer service and what we now call the perimeter departments.

I believe this focus will likely be reflected in store layouts as well, resulting in an inversion of conventional references to store geography. The fresh market will become the central shopping area and the packaged-goods display section will tend to be placed in peripheral locations.

Special thanks to John Lert for participating in this interview. John is an inventor and entrepreneur who lives in Boston. He has spent his career developing technologies in various industries, including media research, marketing research, and retail. He founded CasePick Systems, LLC (now known as Symbotic) where his team developed state-of-the-art automation technology for distribution centers. He is now pursuing the development of the "each-picking" technology that will enable the automated supermarket concept to be realized.  

2018 update: Vist the Alert Innovation site to learn more about Alphabot and Nova store systems.

Related post

The Automated Supermarket: Part 2

Recommend This Post 632

 

Comments RSS

Mohamed Amer said:
Now that’s an attention grabber and a real thought-provoking article!

The headline “automated supermarkets” made me think of the Tesco (Homeplus in Seoul, South Korea) and Woolworths (in Sydney and Melbourne, Australia) virtual store pilots launched last year. And Kroger has also tried a vending machine on steroid (or an automated mini C-store) with a robotic refrigerated kiosk aimed at the college crowd. These have been about using technology and trying out new business models that add value and convenience to a specific customer segment.

Lert’s vision is a disruption to the existing supermarket format and traditional supply chain networks, vendor relationships, trade promotions, unions, and so on (but I get ahead of myself as this may be addressed in part 2). From the consumer perspective, there would be novelty in trying something new like that, but it’ll really come down to whether or not – in comparison to existing options – the automated supermarket saves me time and money, is easy to use, and if it’s accurate and timely on selections/deliveries.

From my current vantage point, I find it difficult to imagine how you can divorce the idea of multichannel (maybe we call it multi-point retailing?) considerations even from the automated supermarket model as suggest here (e.g., to deliver this: “On any given shopping occasion the customer can choose to shop entirely in-store, entirely online, or a combination of both” requires a technology platform that understands all about each specific customer regardless of when, where, and how she shops).

Whether or not such a format and model will come to fruition in the near future is not as important as the nudge it gives current supermarket operators to push the envelope, to imagine future possibilities and innovate the customer experience.

BlackBeltMichael Sansolo said:
This article, following up on a recent BMC blog posting about future competition with Amazon, is exactly the kind of discussion the industry needs to engage in pondering the future of the shopping experience. There is simply no way any of us can know if these types of "supermarkets" will dominate, become niche players or even have an impact. But the topic simply must be addressed.

As Mohamed Amer wrote in his terrific response to this article, the "automated supermarket" does make you think about current efforts like Tesco Homeplus in South Korea. Clearly the technology exists to allow the customer to create their personal virtual supermarket, filled with just the goods they buy on a regular basis. Using a tablet, phone or whatever to place their order, they "shop" virtually. The opportunity for the store to build on their sales comes from making it easy for them to collect the order and possibly entice them into the store for last minute items, fresh goods, menu ideas...whatever. Great merchandising will be at a premium.

The issue that might least matter--that is, after the novelty wears off--is how the order is filled. Shoppers don't care about our logistics, they care about value. If the automated store permits quick order fulfillment it could win. But so could orders picked in advance by workers.

The bottom line is that the future of retail is likely to look extremely different than the present. And given that Amazon is already a virtually automated supermarket, this is a discussion that must be had.

Wish I knew the answers.
BlackBeltDave Carlson said:
I’ve had the rewarding opportunity to know John Lert since he developed a key chain scanner for IRI panelists about 20 years ago. There’s more to know about his work. He also pioneered the Stop & Shop cart-mounted “Shopping Buddy” and its current handheld “ScanIt” descendant, as well as the DeliVision automated deli ordering system. Not a tinkerer, John is a talented and practical engineer who also deeply understands the shopping experience from his real world deliveries in the grocery industry.

His vision for an automated supermarket is sound and well-articulated. The technology fundamentals have been proven in his recent work with Casepick. Take note progressive retailers, this concept will become reality in short years and John Lert will have been at the helm. As Mohamed Amer points out, this is a disruptive opportunity. It will accelerate the pace of market leaders choosing to partner and help navigate.
BlackBeltTom Lemke said:
Great comments everyone, and thanks Bill for bringing this to our attention. In the early 90's I was at a conference where we were all told, "brick and mortar" stores are dead and will be gone from the landscape within 5 years. Clearly that did not happen and maybe never will. However, over time internet shopping has and continues to grow by leaps and bounds. It has been very disruptive to the traditional retail landscape. Why? As we've learned, some customers like to shop via internet and some still like shopping by store...and many do both. Mr. Lert's vision and thought process clearly provides the customer with a choice to do some, part or all of their grocery shopping through an automated process. It's clear that he's put the customer at the center of his thinking and offers alternatives for various types of customers, lifestyles and geographies that can be ideal for an automated approach. Convenience is always one of the central concerns in the shopping experience...and many customers are willing to pay for convenience. Would we ever have imagined that we would pay extra for an aisle seat on an airline? Fifty years ago, we could not have imagined sending a package from the US to Australia overnight and paying extra to do it. Delivery solutions continue to improve and the grocery industry needs to think "outside the grocery box." Mr. Lert is certainly doing that. Automation may not be the right solution for every customer, but lots of highly successful businesses have been built by carving out a niche for a segment of customers. It may not be an overnight sensation but there is a large audience looking for this type of solution. Great thinking! Keep it up.

BlackBeltMike Spindler said:
Two years ago when John first made the pitch the concept of attaching local, intelligent, robotic picking to omni-channel shopping and a endless assortment chain seemed pretty brilliant. It still does.

The omni-channel capabilities exist, as does the endless assortment chain. Refinement is necessary, but is occurring rapidly.
No doubt this robotic picking system will be done, and will be done in a manner efficient enough to use in a store while still providing a substantial ROI for the investment..

Some questions loom....
How does the current labor force react?
What happens to all those folks displaced in the process?
BlackBeltDan Raftery said:
John Lert's concept is intriguing and certainly disruptive, as noted. Two quick comments: His point about grocers eventually not being competitive on packaged goods implies two scenarios. The first is distributed fulfillment, where the store does not hold the CPG inventory. It is shipped from warehouses, even from supplier warehouses, directly to the home, or the store, segregated for the specific consumer order. Why have so many robotic picking locations?

And let's not forget private label. Store brand packaged goods could be a big point of differentiation for retailers in Mr. Lert's future supply chain. This means they will be more than me-too imitations of the national brand SKUs, which is where the smart retailers are going now.

The second scenario is some evolved version of automated replenishment at the consumer level.. We now have some cool shopping list applications, but I think that the portion of our population that can pre-plan as Mr. Lert sees "the night before" is pretty small. There's a bigger chunk of the population that can't do that, for many reasons. Technology will eventually help them. Personally, I can't wait.
BlackBeltMarc de Speville said:
As the range of views already expressed here shows, it will take time for people to fully comprehend and get used to a concept as as innovative as John's automated store. However, the attractions for both customers and retailers are clear. The part of the shopping experience which is boring, labour intensive and can easily be done on a screen will be automated, freeing up resources to improve the part which is more enjoyable and service-intensive - choosing one's own fresh food, with the help of well-informed, paid and motivated in-store advisors/helpers. For store staff this means no more constantly re-stocking shelves and scanning items at the checkout , for customers it means no more pushing large, heavy trolleys full of packaged goods around the store then standing in line to unload all the items onto the checkout conveyor in order to have them scanned, before repacking the whole lot and carrying it all out to the car (not very efficient really when you think about it...!)

The other main attraction is the range of options given to customers. If you are not a natural pre-planner and prefer to get inspiration in store even for packaged items, you can stroll along the virtual aisles and simply scan whatever takes your fancy. Only you won't have to walk past 12 feet of Campbells soup to get to the tinned beans , maybe just a one foot screen (which could tell you a lot more about the product than 12 feet of physical product).

This concept goes very very far indeed beyond Tesco Korea or Peapod's so-called virtual stores in subway stations (actually just a large screen showing a limited number of commonly purchased items). Personally I think It could potentially mark as significant a paradigm shift as the switch from counter to self-service that occurred some 70-90 years ago. It seems the keys to success will be 1) a step-change in single-item product to person picking technology (which Amazon - for all its technological prowess - has still yet to deploy); and 2) partnership with a progressive food retailer in order to validate the economics and operational aspects. I look forward to hearing more next week.
BlackBeltDavid Lubert said:
Disruption is in the eyes of the beholder! From a shopper perspective they do not care about omni-channel or supply chain, trade promotions or end caps In the end consumer behavior will have to transform and while I think it might take some time to get consumers to plan ahead with enough scale the real question is how to handle Millennials who I can see placing orders via smart phones and expecting to pick up orders in two to three hours. The ability to fulfill those expectations will make or break this concept in my opinion. The millennials should not be discounted, because those are your future shoppers in grocery terms. Interested in Part 2 and kudos to Bill and team for bringing this topic to us!
Bill Bishop said:
A lot of great builds on John’s vision. One that caught my attention was Dan Raftery’s concept of distributed fulfillment which, if I’m tracking with him, suggests a way traditional grocers can maintain their competitive position in this brave new world by pushing the order selection process back to the DC to reduce capital costs at the store and/or perhaps even to the DC of major suppliers if logistics can be worked out.

This seems to me to be an option worth more attention and is just one more good reason to apply activity-based costing to evaluate the economic feasibility of these new retail distribution options.
Gerald Lewis said:
Here are some possible problems I see with this concept.
1. Will it reduce impulse purchases? By requiring non-perishable purchases to be pre-planned in advance of the visit, shoppers will not be exposed to new offers and promotions in the grocery and non-foods aisles.
2. If the store is to provide the option of shopping the grocery/non-foods aisles, won't it require more, rather than less space, so that customers can "share" the aisles with the robots?
3. If there are to be a number of exterior pick-up points, how will the system know where to deliver each order without causing significant waiting time?
4. Won't the required traffic access lanes to the pick-up points significantly increase the size of the lots required, in order to accommodate both traffic flow and parking requirements?
5. How will the payment and pick-up system accommodate customers who just want to pick up bread and milk and get out thru the fast checkout lanes in between their regular shopping trips?

Add a Comment *