Who will seize the opportunity to turn stores into fulfillment centers?
Photo: Ahold Delhaize USA

Are store robots cute, creepy – or nearly useless?

Giant Foods president Nick Bertram revealed at the NRF Big Show in January that his company and sister business Stop & Shop, both part of Ahold Delhaize, were in the process of rolling out Marty,  a 6’3” robotic assistant, to all the stores operated by the two chains.

The Marty robots were initially being used to roam the aisles of stores identifying hazards such as liquid spills and alert associates of messes that needed to be cleaned up.

When speaking at NRF, Mr. Bertram said reactions to Marty ranged from some customers being freaked out to others who “actually loved” the robot, which has “googly eyes” added to make its appearance more humanizing.

Nicole Gallucci, writing for Mashable, can be counted among those who have been puzzled by Marty’s presence. In a column yesterday, Ms. Gallucci reports that each Marty costs $35,000 and goes around stores identifying potential hazards, and that’s it. Marty, if Ms. Gallucci is correct, doesn’t distinguish between spills that pose an immediate serious risk and those that are less urgent. 

“I’ve only seen Marty ‘go off’ once in the produce section,” she writes. “When an employee came to the rescue she couldn’t seem to find the hazard, so in an attempt to quiet the robot she scanned the floor and began picking up any fragment in sight — a questionable crumb, a plastic bread tag, a shred of corn husk. To this day I have no idea what minuscule object Marty was trying to warn me about, but after causing quite the scene it eventually resumed floor duty.”   

In Marty’s defense, the unit is intended to do more than shadow customers around the store beeping away and calling for associates when it finds something/anything on the floor. The robot’s role, as reported earlier this year, is expected to expand to include planogram compliance and monitoring of out-of-stocks on store shelves. 

While some such as Ms. Gallucci may question the utility of tools such as Marty, retailers appear firmly committed to putting robots to work in stores. Walmart announced in April that it planned to add robotic floor scrubbers (Marty, are your reading this?) to 1,500 stores. The world’s largest retailer also announced plans to deploy robotic shelf scanners to improve its in-stock efficiency and pricing accuracy. 

BrainTrust

"While the example of Marty may leave many wondering how the ROI could ever make sense, what Giant Food is doing is learning – and that’s very valuable. "

Mark Ryski

Founder, CEO & Author, HeadCount Corporation


"Right now, robots like Marty are a novelty … a way for companies to flaunt their technological prowess (and in some cases, their wealth)."

Jasmine Glasheen

Content Marketing Manager, Surefront


"Are grocery stores awash in a sea of broken glass and spaghetti sauce? Do grocers need a $35,000 robot to tell a store associate to clean up the mess?"

John Karolefski

Editor-in-Chief, CPGmatters


Discussion Questions

DISCUSSION QUESTIONS: Are robots currently being deployed in stores ready for their primetime performances? What roles will robots play in big box stores, at least those operated by large chains, 10 years from now?

Poll

18 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Mark Ryski
Noble Member
4 years ago

Robots still have a long way to go – but they’ll get there. These are still very early days of the deployment of robots in retailing, and while the example of Marty may leave many wondering how the ROI could ever make sense, what Giant Food is doing is learning – and that’s very valuable. There is no question in my mind that robots will take over many of the mundane and repetitive tasks which are being done by humans today. From floor washing and stock checking to helping customers navigate the store, the possibilities are endless. However, the cost of acquiring and maintaining this technology is still prohibitive and the practical value negligible.

Neil Saunders
Famed Member
4 years ago

Frankly, I do not see the point of Giant Food’s robot. It identifies a problem and then a human has to intervene to clear it up. What’s the point? Just have employees keep an eye on the shop as they go about their other tasks. Walmart’s robots are different as they actually clean the spill which takes away work from store teams, saving them time. I have no doubt that robots will do more work in stores and that it will save retailers costs. But Marty – cute googly eyes and all – is pretty useless!

Evan Snively
Member
4 years ago

My local grocer uses “Tally” to monitor shelf stock. And while I can’t speak first-hand to its effectiveness as well as some in this group, my kids and every other kid I have seen absolutely love seeing it. So patrons being “freaked out” is certainly not a concern. Once their role tries to transition to interacting with customers though — that could be another story.

Bob Amster
Trusted Member
4 years ago

The easiest task for which to deploy in-store robots is inventory replenishment and cleaning the floors.

Camille P. Schuster, PhD.
Member
4 years ago

The use of robots will increase for roles like cleaning or identifying hazards or maybe even doing some inventory tasks. How will retailers integrate these robots? Will robots replace employees, will they be additional “employees,” or will they work alongside employees who will now need skills to manage robots? The use of robots will increase and become more sophisticated so retailers need to start thinking about how to incorporate them.

Steve Montgomery
Steve Montgomery
Member
4 years ago

My first impression of Marty was that he looked a lot like Gumby. On reading about his current duties, I sincerely hope he will be learning more tasks such as those mentioned in the article. If not, Marty is an expensive attraction.

Cathy Hotka
Trusted Member
4 years ago

Until these things can perform a useful function (like checking inventory and cleaning the floor) they’re only an annoyance. How many more years will we have to see Pepper the robot at NRF?

Ryan Mathews
Trusted Member
4 years ago

That depends on what their, “primetime performances,” are. If they are to create buzz, amuse kids, and perform a few tasks, then yes. Will they grow in terms of “skills” and effectiveness? I think so. A decade from now is really fairly unpredictable, but we can easily assume some form of companion function, inventory management, wayfinding, suggested selling, maybe even onsite recipe printing. Who knows? One thing is for sure, they aren’t going away.

John Karolefski
Member
4 years ago

The reported benefit of robots in supermarkets initially is to alert store associates to breakages and spills in the aisles. I’ve always considered that to be a joke of some sort, or something for unsuspecting folks to believe. Are grocery stores awash in a sea of broken glass and spaghetti sauce? Do grocers need a $35,000 robot to tell a store associate to clean up the mess? Can’t the store associate spot the mess himself/herself and clean it up?

So the end game is for robots to check for voids on the shelf and mispriced goods. Can’t a store clerk be paid $35,000 to do the same? Oh, wait a minute. The human may call in sick one day, take a coffee break, or ask for a pay raise. So the real issue is saving money. Invest in sophisticated tin cans instead of people. Yeah, that’s a good policy. I might add that the supermarkets around the country that make customer service a top priority would never install self-checkout terminals or robots. They value people.

One more thing before I conclude my rant. My sister-in-law posted a photo on Facebook of Marty in a Stop & Shop supermarket in Connecticut. The comments were revealing. Here is a sampling: Too creepy, give the job to a human, sooo creepy, too weird, what a waste, not the future I want, etc.

Those are the shoppers speaking. Clear-thinking grocers who value their customers and human employees should pay attention.

Heidi Sax
Member
Reply to  John Karolefski
4 years ago

Humans are good at a great many things. But let’s be honest here: can we really expect store personnel to be able spot out-of-stocks and planogram noncompliance by scanning thousands of SKUs? A typical grocery store carries 30,000-50,000 SKUs! That said, robots are not the answer. Grocers who want to capitalize on AI technology should give their store managers and associates access to apps that help them spotlight and resolve costly issues quickly. Googly-eyed robots, however, are not only unnecessary, but potentially problematic in terms of employee morale, customer experience, and cost.

James Tenser
Active Member
4 years ago

Inventory-scanning robots that can take a reliable read of stock on hand and planogram compliance are coming, but it’s not yet clear to me how soon the tech will be accurate enough and cost-effective enough.

It comes down to machine vision technology that is extremely complex. I worked with pioneering company in the Silicon Valley more than 10 years ago that tried to tackle this challenge. Here’s some stuff they learned:

  • The cameras need very high image resolution and very fast image analyses to accurately identify an individual item by its shape, size and label information. The system must match each item with its shelf label and/or planogram location to determine on-shelf availability status.
  • It’s very hard to distinguish between packages that contain the same item but in different flavors.
  • Flat boxes and round cans arranged up to 12 deep on the shelf are “eclipsed” by the item in front, so robot eyes are challenged to deliver an accurate item count.
  • It’s even harder to identify and count items in bags and pouches that have shiny, uneven surfaces that reflect light.
  • Soft line products are very tricky to identify by style, size and color. Imagine a rack full of sweaters that have been handled by shoppers.
  • An array of dozens of cameras is needed to view the merchandise from many angles at once so as to capture enough information for the system to positively identify items on the shelf.That’s why a bulky six-foot tall machine is required to scan a grocery aisle.

All of this means the device “on the edge” – the robot itself – must carry considerable processing and machine intelligence power on board. It has to work accurately while it is in motion. Data for thousands of items must be collected rapidly and communicated to an even more awesome central computer that can parse the data into actionable form.

So … Store inventory robots are not a trivial challenge. In supermarkets, terror about out-of-stocks is driving experimentation even though the economics are not yet proven. The hardware and AI are certainly getting better, but I think the length of the (machine) learning curve has been underestimated by stakeholders.

Georganne Bender
Noble Member
4 years ago

Robots that serve a purpose, like the ones utilized by Walmart, make perfect sense. Marty, however, is mostly an expensive novelty.

There’s a robot at the hotel where I am staying in Las Vegas. Priscilla brings items requested by guests to their rooms. It was interesting the first time I saw it in action, but after that it became just another fixture. Kids love it though, they follow it around and take selfies. Priscilla is kind of like Gen Z’s free version of Sandy the Wonder Horse, the ride we used to see in front of grocery stores.

Doug Garnett
Active Member
4 years ago

If robots are going to be useful in the store, then they need to be designed to look like they’re useful. So far, it appears that given resistance, stores decide that cartoon characters are the best option. Except they aren’t. Am I going to trust my serious shopping to a clown?

I’m reminded of a case where we worked with consumers around a water purifier for coffee makers. A clever agency decided it should be made in the shape of a cartoon fish. What was clear from consumers was that they didn’t believe it was useful — because it looked like a cartoon. (Removing impurities from water is scientific — not a cartoon action.)

Stores need to take a lesson from this and make robot forms imply their function — assuming there really is a valuable role for them to play. (And on that I’m skeptical.)

Jasmine Glasheen
Member
4 years ago

Right now, robots like Marty are a novelty … a way for companies to flaunt their technological prowess (and in some cases, their wealth). There is definitely a future for robotic assistants in stores, but right now the technology isn’t up-to-par and robotic assistants are more of a headline-grabber than anything else.

Craig Sundstrom
Craig Sundstrom
Noble Member
4 years ago

It should be remembered that all responses aren’t equal: the damage from someone being “freaked out” is likely greater than the gain form someone else falling in love.

That having been said, we should remember that robots don’t have to be used during store hours: many of the planned uses are for stocking and inventory and wouldn’t come in contact with customers.

Paco Underhill
Paco Underhill
4 years ago

At my local Stop & Shop, I find it creepy and the employees do too. It creates work rather than solves a problem. In our grocery store work in Mexico, we found that having ongoing cleaning that the customer can see (rather than after hours) added to the customers perception of hygiene. Like the concept of robot floor scrubbers — like the robot vacuum at home.

Sterling Hawkins
Member
4 years ago

I give credit to Giant Foods for trying out something new (especially a robot!), but I’m not quite sure what the value is, the measurable ROI. Unless this becomes an effective platform to manage out of stocks, planogram compliance, etc., it’s an awfully expensive way to make sure aisles are clear of “hazards.”

Robots are best doing simple, repeatable tasks where there’s a real need. Walmart deploying robotic floor cleaners is a perfect example. I think we’d probably all agree that robots will be increasingly included in our lives, it’s just a matter of having them start in manageable and sensible places to match their costs with current capabilities.

John McIndoe
John McIndoe
4 years ago

There is a lot to like about the Marty project. I like that it has started out with a limited scope to gain insights as to what performance characteristics of Marty work and which need rethinking. I like that there is an expansion plan in place to increase Marty’s role to include important activities such as planogram compliance and out of stocks. And, I like the attempt at innovation. Will Marty pencil out over the long term? That’s hard to say, but if this iteration of Marty does not, I’m confident a future innovation will.