Whole Foods Seeks to Kinect With Customers

I have to admit that I’m usually tickled pink when I get to a store and find a cart that isn’t filled with soaked flyers and has wheels that don’t wobble. That’s good enough for me. But for some time, inventors of various sorts have seem preoccupied with the idea of coming up with a smart cart to improve the shopping experience and, typically, create a new stream of advertising revenue.

While smart shopping cart ventures have failed to establish a significant presence at retail up to this point, Microsoft, Whole Foods and Chaotic Moon, a third-party developer, have begun testing a prototype called the “Smarter Cart” that has the internet buzzing.

According to reports, the new cart is motorized and comes with a Windows 8 tablet attached to the handlebars. Using Microsoft’s Kinect, it will follow a shopper, identified with a loyalty card, around the store. Items are automatically scanned as the shopper places them in the cart. The cart then handles the checkout at the end of the shopping trip.

The cart can also check off items from a shopping list as it goes. A video on the GeekWire site shows the cart alerting the shopper that a pasta with gluten was added when a gluten-free item was on the list.

[Image: Kinect Grocery Cart still]

BrainTrust

Discussion Questions

Discussion Questions: Is there a need for high-tech shopping carts in grocery stores? What will it take for smart or “smarter” carts to become common at retail?

Poll

21 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
David Dorf
David Dorf
12 years ago

I have to say this is a terrible idea. Talking to a cart in a busy grocery store just isn’t going to work. And expecting the cart to follow me is very error prone. I can just imagine how much fun teenagers could have walking around randomly giving carts commands and coaxing them away from owners.

On the other hand, these are just the types of ideas that really push us to consider what’s possible. And from these experiments there will be a few good ideas that make to production.

I don’t need a smarter cart — I need faster checkout.

Max Goldberg
Max Goldberg
12 years ago

If a smart cart can save consumers time and money, while being easy to use, and does not require a team of IT specialists to keep it operating properly, it could gain consumer acceptance. Those are 3 big “ifs.”

Rather than add expense to retailers to buy and service these carts, why not develop an app, which can perform the same functions, that consumers can download to their own devices?

Gene Hoffman
Gene Hoffman
12 years ago

This dude thinks the real need for high-tech shopping carts is not there. And what will it take for “smarter” carts to become common at retail? Luck.

While it’s possible that in the future, technology will produce something consumers will want more than squeezing fresh produce. The driving force will likely be the cash flow from advertisers more than consumer demand.

If a trip to a grocery store is worth its salt, then the food and groceries should deserve our affection over the ever-increasing number of new, alleged consumer bells and whistles. But then I’m more of a foodie, not a techie.

Dr. Stephen Needel
Dr. Stephen Needel
12 years ago

Here’s the problem with smart carts — you end up looking at the cart instead of where you are going. This was one of the causes of Videocart’s demise in the ’80s and continues to be a major problem. We’ve also had discussions about the reduction in customer contact due to self-checkout — this will continue that problem. Assuming shoppers have a choice and choose this option, it’s hard to see how this will ever pay itself back. It’s going to have to drive a lot more purchases over the life of the cart or bring lots of people into a store — don’t see that happening.

Ed Dunn
Ed Dunn
12 years ago

Kinect Sensor for Windows cost $250 on Amazon.com. I do not think this is economically feasible nor can the ROI be justified, no matter how “nifty” this looks or sounds.

Dan Frechtling
Dan Frechtling
12 years ago

I give this one out of three stars.

Does this meet a compelling need? Yes. As Modiv has demonstrated, scanning items in-aisle saves time for shoppers and labor for grocers. But a cart that follows me, repeats what I put in and risks my bumping others is a turnoff (silent mode? apology mode?).

Does this do a better job than existing technologies? No. Scanning via handhelds, carts, and mobile phones already exists. Scanning by swiping items at the bottom of the cart works until the items pile up and block the scanner. As demonstrated, this turns a push cart into a pull cart, which is awkward.

Is this economical? No. Kinect’s specialized camera and sensors plus a tablet will be expensive to install and even more expensive to repair and maintain.

This is a nice showcase of Microsoft technologies — Kinect, Windows 8 tablet, TellMe — but together the whole (foods) is not greater than the sum of the parts.

Marge Laney
Marge Laney
12 years ago

I shop at Whole Foods every week and I can’t imagine how this will work in the always crowded store. Talking carts won’t do anything for the ambience except add confusion.

The only way I see this maybe working is if the customer is issued a wireless headset to communicate with their cart. But, even then, it’s already difficult to navigate the aisles with everyone talking on their phones. I can’t imagine what it will be like with the carts talking to them in their other ear.

And what about speed? The carts probably don’t move very fast, and I’m sure they won’t cut to the chase too often and will say the same things over and over. I think this is a lousy idea.

Verlin Youd
Verlin Youd
12 years ago

I am always a fan of innovation and applaud those who take the risk to try new things and determine how and where they provide real value to the shopper and thus to the retailer. However, in this case I have to ask what is new and innovative? There have been numerous experiments with tablets mounted on shopping carts, starting with VideoCart in the early ’90s and more recently with both Fujitsu and IBM offerings among others. Each has had some different twist, however, none of these experiments have resulted in broad rollouts for some pretty basic reasons; challenges that still exist in this latest attempt.

Why not use what is already in the shopper’s pocket instead of adding less capable and more intrusive technology to the cart?

Ed Rosenbaum
Ed Rosenbaum
12 years ago

The consumer can be convinced they need anything, whether true or not. Even in the case of a “smart cart,” which I doubt will make the shopping experience easier based on the demo shown.

The smart cart is going to be expensive. So two things will happen: we will be charged a fee to use them; or a standard fee will be added to each bill at checkout. I am not interested in paying for someone else’s folly. Nor am I interested in a cart following me and smart aleck teenagers playing around with them.

David Slavick
David Slavick
12 years ago

Classic line, it knows that I completed my shopping list and “hopefully I don’t go wild buying additional things in-store.” Hmmmm … you think the grocer would be happy that the customer had three items on their list and walked out without picking up an extra item or two? Talking to the shopping cart, really? This is technology gone wild. One of my unique skills developed from years of working at an IGA grocery store was picking produce. Can the cart tell which melon or pineapple is ready to eat today, or will be ripe and sweet in 3 days? My nose knows … the cart has no clue. Give me personal service anytime. By the way, no kids allowed in these carts … otherwise a scanfest!

Joel Rubinson
Joel Rubinson
12 years ago

This is exactly as I predicted in November … tablets rather than smartphones might be the way to revolutionize grocery shopping. Here is the RetailWire discussion on the topic based on my blog.

James Tenser
James Tenser
12 years ago

After viewing the short video of Microsoft’s Smarter Shopping cart, I can’t avoid the scary/funny image of hundreds of talking bumper-carts following shoppers around the supermarket.

This is an interesting conception of how the very cool Kinect technology might be applied to an everyday task. But this interaction is quite limited and crude, really. I’m left with more questions than answers:

Why doesn’t the Kinect camera recognize each item without need for a separate camera? Shouldn’t the cart create a blue-tooth “pair” with the shopper’s mobile phone? Can it accept coupons within the integrated transaction? Will it roll my items to my car and return itself after I load up?

Or look at this from the alternate perspective: The Kinect device seems to be deployed solely to enable the motorized cart to follow the shopper around. I’m not quite certain how it avoids running over small children in the process.

Cathy Hotka
Cathy Hotka
12 years ago

Retail keeps doing this — getting excited about a silver bullet solution, rather than doing the basic blocking and tackling that results in customer satisfaction. I’ll settle for a store that is largely in-stock and doesn’t have contradictory prices on the items.

Ken Lonyai
Ken Lonyai
12 years ago

First, full disclosure: I’m a founder of Hum Ex.org which is focused on promoting Human Expression as a natural computer interface for use in these sorts of scenarios.

Having said that, this is somewhat of a solution looking for a need. In other words, it’s somewhat technology for technology’s sake. It’s clearly conceptual and early stage and like anything, in need of many iterations. Think of the Apple I PC iterating over many years, and consumer’s slow adoption to ultimately progress to today’s iPad being considered desirable/useful and popular for many applications.

Clearly dozens of carts, talking, navigating, observing, advertising (of course!), up-selling (of-course), etc., etc., require a lot of refinement to make them feasible and not an annoyance or hindrance.

Then there’s the cost issue, the stolen cart issue (it will need GPS to report its whereabouts), the consumer adoption/learning curve, and more. So beyond hype, it’s a great concept that is very unlikely to materialize in some sustainable fashion for a long time to come. But chances are, one day it will!

Lee Kent
Lee Kent
12 years ago

Another piece of equipment and technology for the store to manage. These carts would have to be tested regularly to make sure they are working properly, especially if payment is involved. Don’t know what all that would entail, but I hardly think it would be worth the effort. I am much more pro about the ‘smart’ technology coming into the store with the customer.

Ronnie Perchik
Ronnie Perchik
12 years ago

I’m always a fan of emerging technologies. But indeed, they should only be utilized if they truly improve the customer experience.

In the case of the “smart cart,” there were some features that were useful. In particular, the store map. Upload your list and discover where each product is located in a couple seconds? I like it. And moreover, tell me if my product, or one like it, is discounted to take advantage of promos. The voice activation and self-steering cart features are cool, but not all that feasible.

But this begs the question, what about browsing? Retailers depend in part on shoppers browsing products, coming upon deals, and investing a bit of time in the store. Maybe though, it’s more about offering new technology to get new shoppers through the doors, which ultimately means higher sales.

This is a fascinating technology to ponder because there are concrete pros/cons. Thanks for posting.

Lee Peterson
Lee Peterson
12 years ago

Do you remember that book High Tech/High Touch? It did a great job of pointing out how new high tech items work best with high touch solutions. Apple has made a mega-brand out of the same idea. Point is, if a shopping cart can help, like product location, that’s a really good thing. But if it starts following me around the store and talking when my kids says “check out,” it becomes a real problem.

Can you imagine a grocery store, packed on a Saturday, filled with carts all talking at once? Bye bye. A bit surprised that Whole Foods, the masters of high touch, wouldn’t suss that one out right away.

Anne Bieler
Anne Bieler
12 years ago

While the promise of a better shopping experience is there, this version of smart cart raises more issues than answers. The navigation would be more difficult as described by previous panelists. Biggest value would be scanning all items before checkout, but alternate solutions are available. And surely an app could be developed for in-store shopping assistance that would quietly help loyalty card members?

Interesting that another discussion for today is abandoned shopping carts, which are another major area of expense for retailers. Doesn’t seem to make sense to add to that cost with higher-tech connected carts and tablets.

Marc Funaro
Marc Funaro
12 years ago

Put the scanner on the side of the cart; eliminate the talking, instead putting a simple display on the cart showing current contents and total cost; eliminate the motorization/follow-me feature, I can push my own cart faster and more accurately (who and where are these carts going to get charged up?); keep the “I can bag these products and walk out, since the cart has already charged my card” optimization; figure out how to prevent an increase in shrinkage, when nobody’s checking “BOB” or interacting with the customer anymore, and everyone’s just “walking out with stuff.”

Pine for the days when someone bagged your groceries, and showed appreciation for the fact that you were even in their store.

Rick Grossman
Rick Grossman
12 years ago

Version 1.0 is a good start. V2.0 has some improvements to be made if it’s going to be economically feasible (if the price can come down).

I do NOT want people around me knowing what I’m buying and choosing. Either the volume should be off, or it can interact with my cell/bluetooth. Shopping is not for the public. I want a little privacy. There’s a comedy sketch in the making. (“Did you want the EXTRA Strength Adult Diapers … you went through the last batch very quickly!”)

Also, a self-driven cart is going to create chaos in the a store. It’s enough to deal with carts parked in a store and navigating around them, let alone one that appears to move by itself. Very cool, but not practical at all.

The scanning demonstrated here is more awkward than hand scanning. Have the scanner installed at a fixed point on top. It’s hard to bend to the bottom to scan. This way items can be placed in bags.

I hope this can interact with shopping list apps such that the customer can plan their trip ahead of time, choosing what they need. A screen on the cart can light up with an arrow pointing to the item needed.

IF the cart does not need to be pushed, let’s rethink the entire shape of the cart. Why not adjustable heights? Why not sides that can open? In an aging society, lifting a bag up and over the edge of the cart to put it in a trunk is difficult. The shopping cart has not changed in a hundred years with minor exceptions.

Philip Wheat
Philip Wheat
12 years ago

Phil Wheat here from Chaotic Moon Labs. Love the comments and the various views. We’re in no way retail experts — we started this project as a step forward in computer/human interfaces. Whole Foods was kind enough to let us test it in their flagship store and learn about how people react to it and how they view it both good and bad.

We’ve been THRILLED with the attention from the press and from Microsoft, especially since we are now about 3 weeks into the project. We’ve heard lots of feedback on the challenges (navigation, power, durability, etc) but we’ve also heard lots of feedback from excited people who are looking forward to us overcoming those challenges. We hope we can help move the state of the art forward. Even if it never gets into every cart, we’d love to ensure that the motive systems help those with shopping challenges and the monitoring of ingredient fine prints should be helpful for most people. And that doesn’t even address the scenarios we have for recipes and localvores. The options are endless once we connect the shopper to the product database through an intelligent agent.

I’ll be happy to address any questions – just remember this is a Labs project; we haven’t started looking at product releases yet.