Hacker News new | past | comments | ask | show | jobs | submit login
Amazon Go (amazon.com)
1247 points by mangoman on Dec 5, 2016 | hide | past | web | favorite | 982 comments

I worked at a grocery store for several years, and one thing I recall is customers CONSTANTLY putting items back in a random aisle, rather than where they found it.

I wonder how this tech deals with that? Maybe they figured that out, too. But I was amused in the video when I saw the customer putting it back where it belonged, because that's not how I remember that going...

All that said, this is fantastic and exciting.

Edit: I also hope they're already thinking about EBT cards and WIC.

One interesting benefit of the “detect removal from shelf” concept is that there might finally be a time stamp associated with the removal of the item.

Right now, if you see a random perishable item sitting on a shelf, you HAVE to throw it away because it could have been there for a long time. On the other hand, if you can see that some Frozen Peas were only taken off the shelf 2 minutes ago, you can just put them back and they’ll be fine.

Although, what I’d really want is not only the time stamp but the customer. I’m sorry but if you cost the store $25 by leaving a damned ROAST in the cereal aisle, I would be perfectly happy to never let you in the store again.

> I'm sorry but if you cost the store $25 by leaving a damned ROAST in the cereal aisle, I would be perfectly happy to never let you in the store again.

Wouldn't a better solution be to charge the customer for the roast and if they complain, you explain: "sorry, you didn't put it on the proper shelf, the technology considers that as a purchase", and possibly eat the cost in the form of some incentive to come back to try to keep them. The ones that don't complain either didn't notice, or they don't care enough to stop shopping, or they won't come back like you suggest.

The ones that complain get it taken care of, the ones that don't don't cost you anything. Win-win-break even?

Aside from putting up the odd sign to discourage certain behavior, no, I do not think that stores have any good reason to bend over backwards to keep individual customers.

Retail is far too accommodating already. Some people figure that being a “valued customer” means they can be an unending source of sunken costs in time, effort and stress, among other things. And those costs can be multiplied across the other customers waiting in line, too.

If a “customer” is destroying your inventory, annoying other customers, or commanding far more of your time than warranted, there is no reason to put up with them. Protect the larger investment, which is: all your other customers, your store, and your employees.

Stores bake that into the cost of doing business, it's like dropping something and breaking it, in the vast vast majority of cases you won't be charged for it, just be clean up on aisle 3...

Heck I've dropped jars or similar things during bagging after paying for them several times and every time I was offered a replacement.

Mistakes happen, it's often considerably more expensive to deal with customer complaints especially in the age of social media than it is to replace an item.

It's also important to note that this is baked into the cost of doing business all along the supply chain, if items are not sold they will be often returned by the store to the distributor which would chuck them as a loss, or more often than not sell them for other uses other than human consumption.

Some perishables are thrown away others are then sold to other industries e.g. the roast that was left over might end up as dog food...

For a dog food company it's cheaper to buy discarded meat produce the dog food take samples and while it's being shipped do the cultures to ensure that there are no contaminants or bacteria and if something fishy is found just do a recall upstream for specific batch than it is to buy "fresh" meat and ingredients which are fit for human consumption.

Supply chains are huge and complex and all these little annoyances don't really count for much, it only really bothered very small stores that have to buy everything almost up front and they aren't leasing effectively shelf space for distributors.

> > Some people figure that being a “valued customer” means they can be an unending source of sunken costs in time, effort and stress, among other things.

> Stores bake that into the cost of doing business


> "I return half of what I buy," says 30-year-old Alex Demetri, who spends £500 to £700 on clothes each month.

> She also admits to wearing some of her clothes first before returning them.

> It is customers like Ms Demetri who are causing problems for shops, which are "struggling to cope" with the number of items returned, new research suggests.

> So-called "serial returners" are bringing back items which have been used, are marked or have parts missing, making a quarter of it unfit to resell.

Occasional mistakes happen, but some people deliberately do this kind of stuff.

In Australia if you break something in store, they typically wont charge you, but if they do, they have to charge at cost. I wonder if this another reason many stores dont charge. As any store with a larger markup would have to let you know that $55 vase they were trying to sell you cost them $6.

I've never heard this "rule" before, do you have a source? "Cost" is a pretty flimsy concept.

I think the rule is that, when a business charges a consumer a penalty fee of any kind, the fee must relate to actual costs that the customer's actions imposed on the business. A business could estimate the cost of bananas, but they couldn't use the sale price if that was usually ten times the cost. This applies to things like bank overdraft fees and hotel cleaning charges as well. The idea is that consumer contracts can recover costs, but aren't allowed to punish people.

I don't have a source for that.

Wait til they apply music piracy concepts!

"sir, you ate 4 grapes in store. Our grapes average 2seeds/grape, so we've calculated lost sales in the range of $800.

You see, those seeds could each grow into a vine that will produce an estimated $100 worth of grapes over the lifetime of the plant. You're basically stealing that money from us!"

I may be wrong... I read this previously that was very clear about the customer was due to pay the supplier cost. But I did a google now to find the article and the best I could find is this is a civil case and not "not covered by the Australian Consumer Law" and it seems to be at discretion of the court for value lost and how at fault you were.


I thought "cost" was one of the simpler concepts - the price at which the store bought the item from the supplier.

Even that fluctuates frequently. When I worked at a grocery store, when you ordered something it would show you what the item would "cost" the store.

This could change from day to day (e.g. fruits, veggies), vary based on quantity ordered, etc.

Then you get into private label (store brand) products, where the "cost" was usually either $0 or some ridiculously low number.

This was a national chain. At the store level at least, we wouldn't have been able to find the value of an item "at cost" with any confidence.

It'd be nice, for the stores, if they didn't have to bake so much "attrition" into their accounting.

The attrition is going to be high anyhow, tons of stuff gets damaged during shipping, handling and stocking, tons more is never sold.

Most large stores take out nearly everything off the shelves at night even if it's still within it's "use before" date which on it's own is utterly nonsensical to begin with as most items don't expire for days, if not weeks, months and even years from the "use/best before" date.

People drop stuff, people mishandle goods, how many people squeeze a vegetable or a fruit to check if it's ripe damaging it? how many apples get a mushy spot because they banged around in the crate?

You are literally scraping the bottom of the barrel of inventory attrition when you are talking about perishable misplaced items, compared to everything else they are a rounding error.

There is a lot of loss baked into every supply chain, unless you are going to change it in it's entirety really don't bother with the end, the loss at the point of sale is minimal compared to everything else.

I get that there's a lot of attrition, but less attrition would be nice.

> I do not think that stores have any good reason to bend over backwards to keep individual customers.

Funny; want to know how Amazon CS became so popular with people? They gave complainers what they wanted, and only shut down repeat abusers. You want to exchange or return this TV for no reason at all? Go ahead. You bought this a year ago and want to return it now? Get a rep on the phone and it's done.

It was fascinating to watch from the inside. CS reps became easier to hire (no need for independent thought when 95% of the calls can be answered with binary flags determined by a "follow the prompts" wizard), Amazon's CS approval rating skyrocketed, and they're still making money to this day.

So, yeah, they have proven that there is a really good reason to keep all but the most abusive individual customers.

Perhaps my example was a bit too extreme. But someone else replied to me with a better example of notifications on an app, or warnings that items are still in their cart but did not pass the store threshold, etc.

Seems like the maximized value could be somewhere in the middle, maybe without trying to sneak in replacement costs for items left on other shelves..

I agree, especially you see those videos on youtube where the kids are intentionally 'slipping' and throwing multiple jugs of milk onto the floor.

The store will have video footage of everything... maybe their sensor AI will eventually distinguish between intentional and accidental spoilage.

IMO, any messaging in the format "sorry, $thingYouDid, the technology considers that as $notWhatYouDid" is a recipe for customer loss.

I agree with your point, but at the same time, you should be charged the price of the food if you grab perishable food and place it in an area that renders it unsafe. I worked in retail and at theme parks, and I'm personally of the opinion that it's better for everyone in the long run if you fire entitled customers.

> if you fire entitled customers.

That would be fine in the ages before social media. Now everyone "takes to twitter" and tries to organize a social pitchfork campaign. The victimhood mentality is real. People get off on the celebrity from being wronged by a big bad corporation. It wouldn't be long before some jerk posts a video of himself leaving a roast on the cereal aisle and being '86'ed from the store, then post it straight to Youtube for all the delicious karma points. Corporate image is a big deal.

Unless you make it "your thing." Take for example the Alamo Drafthouse, a chain of movies theaters that are aggressive in removing patrons that disrupt other viewers' experience. In the few times that people have tried to complain, the company has generally come out of it for the better.

I will admit this strategy won't work for everyone. Most corporations are not willing to respond to a complaint by using social media to (accurately) call the complainer an obnoxious asshole.

The real problem, though, is that one time when the company is in the wrong over a bug in the AI. Then you basically have a faulty AI, and by extension, the store, falsely accusing the customer of vandalism or some such thing. Not a good way to go.

Maybe this is different, but my experience with recent tech innovations in brick-and-mortar payment systems haven't been positive overall. More trouble than they're worth.

This could very well be different, but the minute the store starts valuing the AI over the customer, I think the store is in for some trouble public relations-wise.

Agreed. On a related note, I wonder with the extra efficiency gained from no checkout lines, how much it would offset lost revenue. In other words, if, say, you are able to serve 20% more customers, even if there's 4% more loss from tech bugs (not considering shoplifting), the fact that you're moving more people through the store might make up for it.

If someone buys 45 items, but 3 dont ring up, as long as the 3 were relatively cheap items like a can of beans, as opposed to a $15 jar of spices, does it really matter ? Over time, the system will learn which items "go missing" most often and focus on them specifically for better inventory mgmt.

who's got more profit: Alamo Drafthouse or Regal Cinemas ? Or rather, if you own X-thousand shares of XYZ corp, do you want them to "aggressively wage morality war" at the cost of $1.00/share, or maximize profits ?

From what I've read Alamo Drafthouse has over double the per-screen revenue of Cinemark. Their strict, pro-viewer policies really engender customer loyalty.

Sounds like that place would have lower prices since the costs of customers like that isn't spread around to considerate people, I think I would like to shop there.

, you should be charged the price of the food if you grab perishable food and place it in an area that renders it unsafe

That'd constitute a massive retooling of consumer behavior and logic and would require quite a bit of conditioning over many years, no?

That sounds like exactly what Amazon are trying!

Not really. Customers already pay more if they don't have the company card, aren't "loyal" etc... Instead of it being framed as a fine or punishment or way to "fire customers" it is simply referred to as a discount and a way to get coupons.

> retooling of customer behavior

One or more hits to the wallet would be strong negative reinforcement.

Getting rid of bad customers is exactly what is being suggested.

Not at all in this case. The benefits of this technology and the disincentives for bad behavior are strong enough that this would work.

According to the cupcake example in the video, this is all in real-time, right? A much better experience would be warning the user as they go - the ham remains in their "cart" and is clearly flagged as abandoned (maybe with a push notification if they don't have the phone out, though I imagine most customers will be double-checking).

Why not just take it out of the cart then if it is clearly flagged as abandoned ?

> Wouldn't a better solution be to charge the customer for the roast and if they complain, you explain: "sorry, you didn't put it on the proper shelf, the technology considers that as a purchase", and possibly eat the cost in the form of some incentive to come back to try to keep them. The ones that don't complain either didn't notice, or they don't care enough to stop shopping, or they won't come back like you suggest.

One could simply retrain the customer by literally charging them the moment they take it off the shelf, refunding them if they put it back. It would look crazy on your bank/credit bill but i think a legion of micropayments maybe the future anyways. There might be an opportunity there to turn a stream of 100s of micropayments into meaningful data for the end user.

There are many potential reasons that an item could be put back in the improper spot on a shelf that does not induce a loss for the store nor is malicious by the shopper. Things very often get put back into the same shelf but a different position.

But what about this: What is to stop someone from pulling items out of other people carts?

There are no carts, just bags. This doesn't appear to be a replacement for traditional grocery stores but rather corner stores or bodegas, so there isn't a need for them.

>Wouldn't a better solution be to charge the customer for the roast and if they complain, you explain: "sorry, you didn't put it on the proper shelf, the technology considers that as a purchase",

It might be legal federally, and in red states, but this sounds like a class action lawsuit in a liberal state with strong consumer protection laws just waiting for hungry lawyers.

UPC codes are not uniquely identifiable to the unit, only to the product. You could narrow it down to all of the people who took that item off the shelf, but you'd probably be into surveillance reviews after that.

An option is to label all the items with a unique identifier, but that is typically seen as too costly, which is also the reason Amazon hasn't fully fixed the FBA counterfeit problem.

Amazon is boasting it involves "computer vision, sensor fusion, and deep learning" so there is definitely quite a bit more going on than the UPC code. It also doesn't appear any UPCs are actually being scanned in the first place.

I think it might be purely rfid based. All major rfid vendors have theft prevention devices which can detect in real-time products crossing a given line.

Decathlon, the world largest sport goods retailer already use that technology.

They still have checkout lines but technically speaking they would be able to charge you when you leave the store.

Last time I spoke with them they were using Embisphere hardware, but any vendor, Checkpoint for instance, could be used to similar effects.

If all products are rfid tagged (which is entirely possible given the current price of metallic ink) then this store is at most state of the art.

If they actually use vision techniques then it is actually quite a feat. Current vision techniques used in retail are either too crude (when based on the store cctv cameras) or too costly (another French company, IVS, has demonstrated a self service buffet style automated checkout but AFAIK it is still prototype).

Disclaimer: I have links with investors in both Embisphere and IVS.

The last time I checked, the price of a printable RFID tag was around 10 cents. That's far too expensive for a 15oz can of beans.

What sort of prices are you seeing?

The ballpark of 10 cents per printed RFID is right.

I have reasons to believe several ink manufacturers/printers are working on an order of magnitude less per tag.

But are you so sure that 10 cents is too much?

What do you think the gross margin on that can is ? ;-) What do you think the net margin on that can, once the checkout lines (and their personnel) are removed, would be?

If a stitched RFID chip on a 2€ thsirt is currently cost-effective for Decathlon, I see no reason why printed RFID tags would not be cost effective in the very near future, if it is not yet the case (and once again, I have reasons to believe it is almost already the case).

I share your optimism - the future of retail belongs to self-service with the checkout process integrated into the shopping cart. Grab and Go.

However, I don't think that end game is the first use case for an RFID solution - inventory monitoring/LP/loyalty can demonstrate ROI long before the checkout lines are removed. For those use cases, the unit price needs to be much cheaper, but an order of magnitude improvement might be right.

You are absolutely correct. Inventory monitoring and LP are where money is to be made (or rather "not lost" :-) ).

I have skin in the game on the loyalty part so I will abstain from speaking on that.

Coming back on your thoughts about inventory, the birth of Embisphere, the RFID company I spoke about earlier (and the reason why Decathlon started to use them), was solely the invention of a "racket" for fast instore inventory


Other uses (checkout, LP) were almost an afterthought. The sole gain on speed and accuracy of in-store inventory was enough to decide Decathlon to add RFID chips on all its inventory and Decathlon inventory is massive ! They are in the range of 30-70k active SKUs, with sometimes hundreds of thousands units of stock. Sell that in hundreds of stores worldwide, add the warehouses in every country and the manufacturing facilities in China, and you end up with millions of euros of investment just for that damn inventory ;-)

Nice! It looks like Nuukik is well positioned to take advantage of that new infrastructure!

It would be lovely if you developed an in-store navigation capability. It's so frustrating to run into a store to pickup something, and not be able to immediately find it.

The results of periodic scans should provide a decent point cloud that could be used to determine shelf/aisle geometry without a blueprint. Foursquare uses this sort of approach for its interior mapping process, but they can't tell me where to find the bean dip. There are multiple obvious ways to monetize that dataset including simply selling it to Foursquare.

We may or may not already have developed proofs of concept of in-store navigation systems for european retailers ;-)

In-store location of products is deeply linked with very complicated discussions between retailers and product manufacturers.

Moreover, facing is a very strategic part of retail and I doubt retailers would be happy to release their facing strategies to outsiders or competitors.

Even inside a retailer's organization, several opposing views exist, between maximizing breadth of product range, giving prime exposure to the private label, etc....the equation they have to solve is very complicated and I don't think there's an ideal solution to this. A retailer facing strategy is linked to its core values. It has a direct impact on its bottom line and an indirect one: the consumer's unconscious perception of facing "strategies" is probably very significant.

It's certainly buzzword compliant...

No plan survives first contact with the customer's kids.

Good luck with that. The customer is always right.

You start doing stuff like that it will bite you back 10x.

The customer is always right just means that if the customer wants a bright pink top with bring pink shorts, they are right.

It doesn't mean every stupid thing the customer does is right.

> The customer is always right.

This adage only works when marginal value of a customer is high, and monetary preferences aren't utterly dominating customer's thought process. Business dealing with necessities, or ones where demand outweights supply, aren't like that - that's why in a grocery store, customer is trash. There's plenty more where he/she came from.

Have you ever worked in grocery?

The customers tend to be brand loyal, and the lifetime value of a customer is very high (family with 2.4 kids and a dog is minimum $10k/year in gross sales), so when you start banishing customers for doing things that they may not even realize that they did, they will loudly tell everyone they know what a bunch of assholes you are! The $25 roast will cost you $500.

If I stop shopping at a local grocery that uses a loyalty card for two weeks, they will immediately begin sending coupons worth 10-20% of my average transaction value to get me back. The ROI of giving away $20 at a pretty low margin implies a high value.

Only briefly, helping with inventory. But my SO did, and well... I eat food, and so does my family. And so I learned that with grocery, the one consideration that literally trumps all others is... price. Other significant factors are geography - people tend to shop closest to home or their commute path, and assortment - the more you can buy in one place, the better. There's very little a grocery store can do to chase away customers living in the area except having prices higher than the shop next door.

Now my experience is of course limited to shops in urban Europe, servicing low- and middle-income populations. Maybe high-income people can afford to vote with their wallets, but with all the people I know, the ability to save $100+ / month by just going to the cheaper store of the few nearby is enough to make them not mind grumpy cashiers.

> they will loudly tell everyone they know what a bunch of assholes you are!

I have never in my entire life seen this behaviour impact a single company. Even though I'm first to badmouth asshole businesses and praise the nice ones. Even in tech, I'm yet to see a single company seriously impacted by people's reaction to bad behaviour. I mean, how is Uber still around? Or how is Lenovo still selling laptops?

This is a tangent, but I think that's better stated as "the customer is never wrong". That frames it as a customer support challenge rather than a surrealist exploration of what your customer may claim as their desire.

Way back when I worked retail we had a cohort of particularly ignorant customers who would get fresh meats from the deli counter, wander about the shop then decide they didn't want said meats and stash it behind random products in whichever aisle they happened to be in at the time.

> I worked at a grocery store for several years, and one thing I recall is customers CONSTANTLY putting items back in a random aisle, rather than where they found it.

If the incentive is reducing the chances of being charged by the system for an item they didn't buy, you bet people will put stuff back where they found it.

If they were trained on normal shopper, how long does it take for the first exploits to appear, like some ML-equivalent of tag switching.

Like I heard from HEB Central Market employee, that they had a hard time with their bulk self-portion coffee beans, where the price range is quite large yet the beans look pretty identical.

Like most bigger shops, I think amazon will just tolerate the loss without enforcing it too much, if it stays manageable and under enforcement costs, which includes a too negative impact on the general shopping experience.

Like self checkouts, these technologies will not work well in high crime areas. By requiring customers to have a cell phone and an amazon account they can avoid a lot of the higher risk customers at least.

I'm curious, what about self checkouts does not work in high crime areas? Usually stores with self checkouts still have employees and cashiers in the store monitoring for theft.

What about the presence of self checkouts makes theft more likely?

True. And maybe they also figured out a technical solution, too. If not, I'm sure they will eventually.

But I'd also have been more impressed with the video if they showed stressed-out parents with crying kids and their hands full as they've got their cell phones tucked between their heads and shoulders, rather than young people quietly grabbing a single item and leaving.

This can actually HELP with those problems because a lot of those problems happen while in a line. Maybe they expect their customers to be like the ones in this video, but certainly my store was a little more chaotic. They should design for that chaos -- and maybe they did, but the video doesn't show it, is all. Presumably because they wanted to stress how easy it was, but to me that comes off as alien to the real world, based on my experience.

> But I'd also have been more impressed with the video if they showed stressed-out parents with crying kids and their hands full as they've got their cell phones tucked between their heads and shoulders, rather than young people quietly grabbing a single item and leaving.

It's an 1,800 square foot convenience store in a yuppie area stocked with what appears Whole Foods like take and go food, not a Wal Mart Supercenter. It will be quite a bit different if they open a large store out in the suburbs.

That's fair. Although yuppies have kids, too!

It's not the kids that cause trouble--it's the carts with 250 items, the arguments over price, and the paying by check.

If that risk exists it will turn into a disincentive for people to even pick up items and look at them closely. If this gets in the way of people making impulse purchases it could significantly reduce the store's sales volume - after all few people stick to their shopping lists and retailers know this. I'm looking forward to seeing the results!

This. That 80$ champagne bottle over there? Better get my hands off it!

Dont take the kids to that store :)

I can tell you've never worked in QA :)

you must assume you're users will do whacky things and they will have no idea how the system works.

Would it be feasible solution to have drop-off basket at the counter for items that you do not want to buy? Also there could be refrigerated and non-frigerated baskets, along with a nice sign saying that it's ok to not change your mind.

I really wouldn't trust 'picked up off the shelf' detection, not without the whole thing looking like a giant vending machine. RFID tags on products probably works better.

I don't think the system works like that.

From my first quick take of the video, the app+turnstyle is used to identify you to the store. The video system then tracks your position as you walk around.

When you walk out, the items are recognized and tallied by a large RFID sweep. Funneling you back out through a turnstyle makes sure the vision system knows it's you. Notice that you don't need to barcode yourself on the way out, and the exit system is phone agnostic (it's not checking for an NFC or Apple Pay tag or anything).

The whole "tracking individual items as they come on and off the shelves" task is a very complex thing. But tracking bodies as they walk around a 1,500 square foot room isn't that hard.

Not as crazy as it sounds. With RAIN Rfid you can get the xyz location of a tag from ~30ft away accurate to 6" in realtime.

Examples: https://www.youtube.com/watch?v=p1LykdRWTfk https://www.youtube.com/watch?v=-hCA8L7v-R8

I don't think that helps in the use case presented here.

What is really needed is a way to talk to multiple RFID tags quickly without any crosstalk. Think of a handbag with a dozen candy bars in it. How do you scan them all quickly without missing any?

The goal has always been to scan something large in one pass, say a shrink-wrapped pallet of items that could number in the hundreds or thousands. Obviously this is technology that Amazon could benefit from as well, I wouldn't be surprised if there's a lot of technology overlap here.

> How do you scan them all quickly without missing any?

Most RAIN Rfid readers have a read rate of 800+ tags/second. Door portals are a solved solution and now phased arrays are on the market which scan all items in realtime. Take a look at the impinj marketing material for more information. For example the RS2000 chip: http://www.impinj.com/products/reader-chips/indy-rs2000/

That's a definite plus. I was just saying that the long-range position tracking might not be as important.

Also, since the app is still on your phone there is no reason they can't be using it to gather data as well.

They did say sensor fusion. I'm not sure if they have indoor location tracking in the app but it seems reasonable.

> When you walk out, the items are recognized and tallied by a large RFID sweep.

But the video clearly shows the items being recognized as someone takes them from a shelf (and puts them back). The items don't need to be recognized at the time you walk out, the store just needs to know that you've walked out.

You mean that floating list in the air?

I think that's just visually giving you an idea of what's happening. Like I said earlier, I'm giving a simple naive presentation of how the system might work. Or, at the least, how I would design it without dealing with finicky shelf sensors. Ask anyone that has ever worked in a hotel with in-room minibars how well those things pan out in real life.

Unless that store is just a mockup and not how it really looks and works, I see nothing on those shelves or in the sky above it that is watching you put that cupcake back on the shelf.

The real way to find out is if someone in the demo video can pull their phone out of their pocket mid-trip and see their current inventory and total price. But I didn't see that in the video.

The girl in the video at 1:26 looks at her list and total, but after she's been through the exit turnstyle.

Agree with you, would be much easier to just checkout once when you leave.

The video makes it look like it's constantly up to date, but that sounds quite complicated (despite dropping the ML, DL buzz words).

> You mean that floating list in the air? I think that's just visually giving you an idea of what's happening.

A friend at Amazon said he tried to trick it by taking two items at a time and it got it right every time, so yeah it does appear to track when you take the items.

Did the system know before checkout, or after?

I can't help but picture Indiana Jones swiping the golden idol from the pedestal with one hand, and quickly replacing it with a bag of sand using the other hand.

RFID probably works better if you control the packaging, for instance, for deli counter products. But for packaging that you don't control, you either have to slap an RFID tag on when the items get to the store, negotiate with suppliers to start to include RFID tags in their packaging for items shipped to your store, or arbitrarily limit yourself in terms of which products your store may sell. Not to mention that, if your store slaps on RFID tags itself, it raises the OpEx of running the store as well as needing to come up with tamper-resistant tags, because somebody would be able to steal a product by discreetly removing a tag and then walking out of the store.

In essence, RFID tag technology already exists, and there's a reason why it hasn't taken off in grocery stores.

> because somebody would be able to steal a product by discreetly removing a tag and then walking out of the store.

That's already handled in grocery -- people can mess with regular barcodes if they want. it's a known cost of business

And if they know who you are because of your phone, and they are doing inventory on a really regular basis (every 6 hours, maybe even less), they might actually be able to track down who it was.

I would assume RFID would be more accurate to begin with. Seems like there's a lot of room for errors with their proposed technology.

BUT what if their goal isn't to make razor thin margins at grocery stores, but to test out their tech on human tracking. You've got to believe that they have a more long term plan than making a better self-checkout.

They can now tie in your Amazon Go app with you as a person. Most likely all the sensors and cameras in the store will be both tracking you via your app with Bluetooth, but also perhaps body heat sensors and the like that track you based on your unique body pattern.

Then they can tie this in with your Amazon Prime account to better improve recommended purchases to you, they can tie this in with your Prime Now account and your home address and more immediate delivery needs. Drones, Kindle, Echo as well.

The goal is deeper analytics and tracking on individual humans. That's how Facebook makes their money, and Amazon can perhaps have deeper data on individuals than any other company tracking browser cookies.

RFID has some privacy concerns, since the tags continue work when you are outside the store and somebody can scan what you are carrying.

Using these as EAN barcode replacement also has the problem that the system can't tell if you bought the product from Amazon GO or if you just happen to be carrying it with you when you enter the store.

Someone can already walk around a grocery store and see what's in my cart or they can just watch me checkout. Buying groceries isn't exactly a private experience unless you buy your stuff online. Securing tags in things like a passport or credit card is important since identity theft is real and hardware to read the tags is cheap. Knowing what things I buy in the grocery store might help some marketers know what sells but I can't really think of any nefarious use of that kind of data.

Maybe you could use it to make some guesses about a person. An older guy buying frozen dinners, beer, and not much else is probably single or divorced. A woman buying prenatal vitamins is probably pregnant. Someone buying both Special K and Fruit Loops might have a family. But someone can already make these determinations from using their eyes. RFIDs might make it a bit easier but don't really add any new infosec facet to this grocery store experience.

> Someone can already walk around a grocery store and see what's in my cart or they can just watch me checkout.

The problem with electronic surveillance is never the surveillance part. It's the part where we can surveil billions of people at once without the need for billions of spies.

Of course it's trivial to tell. When individual items are tracked, it's exceptionally clear if the item is from the store or not.

Bring a Faraday's cage, or rather Faraday's bag. Use it once you've paid to store your goods.

I think RFID tags are too expensive to put on every product. I may be wrong on this.

Quick search on Alibaba shows that on volume these are just $0.01 each (and probably lower when you buy larger volumes). Of course sticking these to each product is inconvenient and I would assume the tags are not yet built in in most grocery product packaging.


Alibaba sellers will list almost anything for $0.01 each, though, regardless of actual selling price.

Also the cost of labor and/or machines to stick them on.

Also never underestimate Jeff Bezos' willingness to lose money in the short term.

Most libraries RFID all their books now so you can just drop them on a scanner that detects all the books you have, swipe your library card and go.

I don't think that's actually what Amazon Go is doing though.

I have a feeling you're wrong -- I see RFID tags on almost everything these days, even tiny cheap products.

Yep, in smallish bulk (low 1000s), you can get them for about 10-20c a tag (and these are "self printed" ones you can print in-store as needed via dedicated RFID printers) (e.g. https://www.atlasrfidstore.com/rfid-label-4x2-for-the-zebra-...)

I am sure that amazon/walmart/tesco/whoever will need more than 1000 at a time and so obviously get a larger discount. Probably down to well under a penny or two - starts to make a lot of sense when you factor in the other benefits (lower staff, easier stock-tracking/taking etc)

RAIN Rfid tag is < 10cents.

1% is a decent average margin in a grocery store (http://smallbusiness.chron.com/industry-standard-gross-margi...).

So, it would require a $10 item to break even on a €0.10 tag.

I don't think you can make that conclusion. Presumably this store will be saving money on cashiers by instead using RFIDs.

The article I quoted claims 1% is the _gross_ margin, in which case my argument is valid. I now think that is incorrect, as other high-ranking Google results claim it is the _net_ margin.

But if using RFID saves them more than the cost of the technology its a win.

Basically if Savings from reducing sales persons > Cost of New Technology, you'll add the new tech.

It doesn't have to be strictly cheaper, as it also adds convenience. Not having to wait at the checkout will be worth something to customers, possibly quite a lot in the case of a shop targeting richer customers.

Given the turnstiles that the users entered/exited from I'd have guessed there is RFID in play too.

Another thought I had was how it will handle multiple people filling the same "cart". Often I'm shopping with my girlfriend and, being the time-strapped millennials that we are, we'll split up the grocery list to divide and conquer and meet back up at the checkout aisle. How does this reconcile that? Would both of us have to scan in when we enter the store, or link our accounts in some fashion?

Exactly that. Or what if you go to shop with a child and he puts candy bar to the "cart"?

Avoiding getting charged is an incentive to put it back in the right place that doesn't exist currently. Maybe not even where you picked up just a discard area

Honestly, that's an "incentive" that should be illegal. You should only get charged for something if you actually take it home or use it. Anything else is a cop out to push the costs of a deficient system onto the user/customer.

It's the store's fault when you selfishly and silently put perishable goods in non-refrigerated areas so they can spoil? IMO the store would have the right to charge you with destruction of property, if not the cost of the goods they're now out entirely due to your actions.

The number of people that never learned "put stuff back where you got it" from their parents is astounding.

> It's the store's fault when you selfishly and silently put perishable goods in non-refrigerated areas so they can spoil?

It's incorrect and misleading to talk like that malicious use case is the only one at play here.

Malicious is probably not the right word, more like thoughtless and lazy. The same mentality that leads people to abandon carts willy nilly in the parking lot.

It sounds like the careless actions of the general public really bother you. Let me tell you something: grocery stores are low margin retailers who make their money by volume of sale, not enforcing conformity. The moment you open your doors to the public, you are going to get all kinds of mentally ill, aloof, high/drunk, distracted, disabled, elderly/senile, and (literally) retarded people in your store. It's futile to judge their actions using your idealist looking glasses.

The moment you start kicking people out over subjective "rule-violations", you are eating into your own profits, pissing off people, and projecting your own morality onto strangers. Aunt Minny may have set down that roast because it hurts her hip to walk across the store, and she realizes she already bought a roast yesterday. But over the last 10 years she's spent $25,000 shopping there. Some guy with Crohn's disease may literally shit on your floor if he doesn't drop his perishable item and run home/to the rest room. If both of those people are regular shoppers, sure the lost perishable item eats into your bottom line, but in the long run you are making a profit off them.

Waging an unnecessary morality war can only impede your ability to run a profitable business.

Edit: why not just give employees handheld IR thermometers, and if the temperature of the product is < $MEAT_MAX or $VEGGIE_MAX degrees then allow them to restock it.

It sounds like the careless actions of the general public really bother you.

That's accurate. Ask me how many times I wish I could find the guy that carelessly tossed a cart out in the parking lot (given easily accessible corrals) which subsequently bashed into the side of my car. Or found a packet of hamburger (with accompanying drippy juices) in with the toys. Or...

The moment you start kicking people out over subjective "rule-violations", you are eating into your own profits, pissing off people, and projecting your own morality onto strangers.

Leaving aside the misuse of the word subjective, this is why we avoid the "kicking out" part altogether and use financial incentives instead. It's no different than, say, ALDI charging you a quarter to get a cart (that you get back if, and only if, you put the cart away). Put stuff back where you found it like a civilized person, and there's never a problem.

The number of people who think that people had parents growing up is astounding! -_-

> It's the store's fault when you selfishly and silently put perishable goods in non-refrigerated areas so they can spoil?

I find it really annoying when people put words into somebody's mouth.

Misplaced perishable goods are certainly a problem, but not the one they're discussing right now.

We are talking about people misplacing goods in a store, are we not?

Perishable goods are a subclass of goods in a store, are they not?

Why? If it's understood that the way the system works is that it's in your 'virtual cart' as soon as you take it off the shelf, why isn't it the customer's job to put it back if they don't want it anymore? There are other checkoutless systems that use the customer scanning a barcode as they pick up each item; same deal there.

Why is it the customer's responsibility to figure out and comply with the rules of your (nonstandard) system to avoid getting charged for something they don't intend to buy and didn't in fact use or take out of the store?

There are all kinds of reasons why that is unreasonable:

* different stores might have systems with different rules and policies, causing confusion;

* people may not remember where to return the item;

* the magic machine-learning system might glitch and not recognize the item was replaced, and you probably won't notice since there's little feedback from the "virtual cart" since you're not interacting with it directly;

* another glitch could put another customer's item in your "virtual cart", so you have nothing to return;

* etc.

Systems like this should fail in customer-friendly ways, and "item returned to the wrong place" is a kind of failure.

> Why is it the customer's responsibility to figure out and comply with the rules of your (nonstandard) system to avoid getting charged for something they don't intend to buy and didn't in fact use or take out of the store?

Because that's the store's policy. If you don't like it, nobody is forcing you to shop there.

Interesting considering the case of hotel mini-bars. In some cases, you get charged regardless of use it. If it senses that you removed the item, you will be charged.

if the costs of the products are reduced along with the operating and spoilage costs then its a net + to the customer, which aligns with Amazon's basic value propisition

Yeah, or it's an incentive to not shop there at all. You can't just use "incentive" as a reason why you expect to succeed with a UX that punishes your users for small mistakes.

incentives can be positive or negative

If that happens the customer would just complain and in time we'd lose the trust in the store. So, no, the incentive is most definitely on Amazon to get it right. If they don't, they won't last long. Just imagine news stories with people that later noticed that they "purchased" items that they didn't leave the store with. It would be a huge mess.

It's actually an incentive to not take the product of the shelf (i.e. not buying it).

I imagine having people put items in their original position when they change their mind won't be an issue. I know that if it were me I wouldn't risk being accidentally charged for something simply because I couldn’t be bothered to put it back in the right spot.

If all merchandise in the Amazon Go store have been RFID tagged, this is a non-issue for checkout, as the turnstile tag scanner should be able to sense the items you take out of the store. As far as randomly put back at a different place of the items, the staff can use the tag scanner to detect that and fix it up even easier as oppose to physically locate the item by human vision which is arguably harder and less efficiently.

We don't know the details yet but RFID may solve this issue completely.

Ehh, there'd probably just be a "don't buy" bin that you out the items you don't want in.

Faraday cage bags are already a thing.. I can see them being a lot more popular..

RFID along cannot solve it - it's Computer vision + RFID. Don't forget cams are everywhere in the store that will capture the moment you pick something up...

If the system detects tags disappearing while inside the store, it can trigger an alarm and tell approximately where they were (to be combined with surveillance footage).

I guess tech like this could collect this information and then create an optimized "garbage collection" path for a single employee to put all the stuff back on the shelves once in a while.

The big store chains nowadays doesn't even have someone to put products in their place. Their product provider personal that put the product in the correct place. Now they will have a way to make their own customers to work for them. This is a store owner dream.

How does this handle produce? The assumption seems to be that everything is packaged.

At a lot of grocery stores in Europe, you weigh and tag your own produce. Basically you put your bananas in a plastic bag, set them on the scale in the produce section, and push the number for bananas or select them on the touch screen. It prints out a barcode and price, you stick it on the bag, and the cashier scans it.

Perhaps this could be similar, but instead of printing the barcode, it automatically adds the item to your Amazon cart.

It seems like everything is already prepackaged in the store according to what they have for sale. I know my local grocery store has certain things prepackaged into bundles instead of allowing people to bag them.

It's an 1,800 square foot store, I doubt it has a produce section other than for individual fruits and the like.

It could be flagged as "must restock"... one cool thing is it could then tell you exactly where it needs to go back to, finding that was always a pain (source: I was a cashier in a large grocery store, once upon a time)

To this day (20-odd years later), I always put grocery items back where I found them (and bag my own groceries). Don't be a dick.

Meh, cheap rfid, or funny barcode positionning to ensure their sensors can always track stuff.

I'm not that thrilled, somehow I'm not in love with todays tech and progress (that's on me). Moreover I wonder what people living on cashier jobs (it's an easy target for unqualified and hurried people) will feel.

Goodbye profession.

All the supermarkets I use already push you to self check-out stations....

True, but these often require one or two hosts to unlock issues.

I was for the test but I find the self checkout annoying. It's less efficient than the usual kind. You have less space to unload you stuff; less space to rebag them, and the cashier is lighting fast at scanning and grabbing the money because of 7h/day of doing so. I'm not pro human cashier, I don't think lots of cashier really like it either. But self checkout is not as good as I thought.

Yes. My main supermarket (a former Safeway in the UK) has one person looking after 20 self-checkout stations. I guess most of us have learned the system by now.....

Self-checkout works well for half a dozen items in a hand basket. If you have a trolley full of a week's shopping, you're going to use a cashier. In fact, in this supermarket, you can't even get a trolley into the self-checkout area.

My local M&S store has three different checkout areas. (1) traditional, for people with trolleys. (2) Self-checkout area for people with baskets. (3) Really fast (but very narrow) lane for people just holding a few items in their hands. You almost never see people using the "wrong" area.

What's interesting is that I usually pick the one with the longest queue, ie line 3. The cashiers are really fast so you don't have to wait long.

One other factor is that a lot of us in line 3 are using contactless (Wave & Pay) cards, so the payment process doesn't slow things down. People who don't usually use the waiting time to get their cash ready.

In our local stores in the USA, its one person looking after 4 or 6 stations, 2 or 3 of which are perpetually out of service.

Same. It seems that this is just a temporary and fragile step toward something else. Amazon Go or else.

Maybe an early adopter problem, unless they've upgraded the checkout stations to current technologies? The UK ones are mostly quite recent seem to work fine (though I have no experience outside London).

For what it's worth, some customers DO put unwanted items back where they found them. I'm guessing you wouldn't necessarily notice those cases so you might perceive a higher proportion of those bad actor customers than there are in reality.

I'm assuming that the position on the shelf has very little to do with adding the item to the cart. I'm guessing that taking the item out of range of a scanner and into range of your phone is what does it.

I've not seen anyone so far suggest that people will still work at the store. There will still be people who help customers do what they need, but instead of mindlessly scanning groceries, they'll be there for the things people actually need help with. Returns, price problems, finding things, lifting heavy items, etc.

If that's the case, that this is a solved problem, they should have showed that off!

I would argue the fact that you are automatically charged for an item when it is removed from shelf is incentive enough for people to ensure they put it back in its right place if they decide not to purchase it.

The emphasis on machine learning and computer version makes me think that they're recognizing the products based on what they look like, not just where they are.

In a regular store, customers are anonymous, so they can be jerks without fear of consequences.

If I understand well, in Amazon Go, every customer has an account linked to a credit card. I guess that this will work as a deterrent. I a customer does something wrong, they can get a warning. Next time, the account is canceled. Problem solved.

Are people forgetting the legal perspective?

You can't charge your customer for an item they didn't buy. Trying to make the computer "dumb" doesn't change this. You're going to get angry customers on twitter, chargebacks, and possibly sued.

you are forgetting by the time this becomes a reality we would have all become perfect human beings with perfect ethics and morals and social responsibility just the way silicon valley has always intended and designed every product for

They will detect it and have people in the store replacing it.

In the video they said they use deep learning algorithms. Perhaps each product has a unique number and even if it is not put in the right place the algorithm saves its new position.

If youre on EBT/WIC, what % of those folks have both a smartphone, Amazon Go installed AND an amazon account?

So Amazon Go is a defacto rich peoples store? Sadly, this will probably be a selling point for some.

You would be surprised how many EBT/WIC folks have smart phones since they are often part of educational plans and back-to-work initiatives. It was often easier to order product off Amazon than locally (gift cards, debit cards).

I wasn't meaning it as a class-based statement... just that I don't expect one of these to pop-up in either oakland or the tenderloin in the next few decades...

This made me think of a possible exploit. If these sensors are using weight sensors and cameras, it might be possible to pick up similar items and then put back the much cheaper version.

It turns out the most people aren't interested in complex strategies to steal a grocery items while under intense video surveillance with ID verification.

Or just go to literally any other store and you won't have to defeat sophisticated algorithms to shop lift.

Knowing Amazon they will charge the first few customers for the products, claim it was a glitch the do something panicky to either fix or prevent the problem.

I hate it when companies offer a "how this works" section that doesn't actually tell you a damned thing about how it works.

* How does my Amazon account get associated with the items I take?

* How are items detected when leaving the store? If my friend and I walk out side by side, how does it know (if it does) which items are mine and which are hers?

* What happens when someone picks up an item and leaves without first doing whatever check-in/registration/setup is necessary?

These are purely assumptions based on what I saw on the video alone:

>If my friend and I walk out side by side, how does it know (if it does) which items are mine and which are hers?

Looks like the entry/exit is the same type of set-up you find at most large office buildings with the tap in/out gates (see screenshot: http://i.imgur.com/e7fDglY.jpg). I assume it only lets one person out at a time which also suggests your friend could only enter the store if they had an amazon go account and tapped in themselves.

>What happens when someone picks up an item and leaves without first doing whatever check-in/registration/setup is necessary?

Again, based on those gates, I'd assume you can't actually access the store without going through the necessary set-up first.

Personally, I'd like to know what happens if/when my phone battery dies in-store.

It looks like you don't need to scan anything when exiting. So your phone could die once you're in the store and you'd still be able to check out.

I'm guessing the Amazon Go app generates a one-time-use QR code that gets scanned on entry and then security/tracking cameras follow you all over the store. If the cameras see you in front of an item when its RFID sensor detects it was picked up, they make an educated guess that you picked it up. Then they can re-scan all the RFIDs as you exit for extra confirmation.

They probably don't have RFID sensors on the shelving, and if they did they would only be able to detect when you walk several feet away from the shelf, and would have huge problems if two people were standing right next to each other at the shelf.

It looks like they have a row of cameras along the top of each shelf that will be used to detect when you pick up and place items back down.

The RFID tags are mostly useful at the gates to confirm the visual data and feed back into the machine learning algorithm.

i believe this is it too, mainly because of their mentions of machine learning and computer vision in the post.

Or if you lose your internet connection. or if your app crashes.

I think you meant http://i.imgur.com/e7fDglY.png

Its probably too creepy in terms of privacy to detail how it works which in turn will give themselves PR. I can only assume it uses arrays of cameras and sensors with deep learning. Glorious amounts of tracking data.

My current best guess is, basically what you've suggested. The Amazon Go page says "Computer Vision" which means "Cameras, Cameras Everywhere!". It also says "Sensor Fusion," I bet that means at the very least "Wifi and Bluetooth," with which they'll use to place your location on a virtual map of the store. That location could then be compared with the computer vision object tracking location to "double check" that you really are at that location. And so on.

yep. full on dystopian tracking was my first thought.

It's not "dystopian" if you're on private property and have accepted their terms of service. You can avoid surveillance by, you know, choosing not to shop there.

If there is no way for me to delete my personal data after i leave the store, i find it distopyan.

If one day all the stores are like this, I guess you may argue I can still grow my own food, and everything is fine.

> If there is no way for me to delete my personal data after i leave the store, i find it distopyan.

By "personal data", if you mean the video footage of your shopping, then it will probably be deleted after a few days/weeks. It will just be used to train (reinforce) the machine learning model, and will be discarded eventually. But your shopping history will always be there, as it is now in online shopping sites...

Nope, get back on narrative! If it involves tracking me it's dystopian and literally 1984, especially if I've voluntarily signed myself up to be tracked

Exactly. They say, it uses Deep Learning, Computer Vision and blah blah.. (Add 2 more buzzwords). But, for christ's sake, say how!

As a customer-facing thing, "how it works" means "what do I have to do to get stuff". Like, how does the process work on a UX level.

Which is already addressed earlier in the video. The "How it works" section of the video adds no practical details on how to use the system. It's just a slot to drop in a bunch of impressive-sounded jargon.

Oh. I have to admit I ignored the video and thought the comment was about the "How does Amazon Go work?" paragraph on the page.

* What happens when someone throws a product in my bag. Who pays it?

if you don't notice it and don't put it bag you pay for it, just like in a normal store. That's what I'd assume at least.

Sensor Fusion. Duh.

So will they have weighing scales to walk over? And also register eaten candy bars and toilet visits?

> Sensor Fusion

I wonder, when did the obvious way you use sensors become a buzzword?

How it works sections aren't designed for the technically inclined user they are designed for the layman.

They don't have to get intensely technical but they could say a little more than basically, "magic".

Except to a layman it is basically magic. Using X technology we did Y so you can do Z. I'm sure in coming months you'll get the in depth technical explanation you are looking for but this announcement isn't for you it's for the average person.

That's a rather defeatist attitude to have about explaining technology to laymen.

For example (not saying this is how it actually works): "Scanning your phone when you enter lets us know who you are. Then, our advanced vision tracking system follows you around the store and lets us see what you're choosing. Finally, scanners in our turnstiles verify your purchases as you're leaving the store."

This is an advertisement announcement not a technical breakdown. There is no need to even say that much.

I'd assume the sensors can do a couple things - sense your proximity to your goods (the closest) and sense the difference in time between when you leave and when your goods leave (the smallest). I'd also assume that they use logic along the lines of "your goods are those which have been closest to you (and not on the shelf) for the most time since you entered the store.

I'd assume trying to forcibly enter the store without registration will set off an alarm. I can't wait to show up with 15 of my friends and run amok in the store - it'll present an interesting legal experiment (unless they just get us for trespassing).

RFID, image recognition with cameras... what is it? Can someone go and check it out? ;)

Yeah that was pretty terrible. I half expected to hear the word synergy in there somewhere.

Sensor fusion? You mean multiple sensors?

I'm more interested to know how it handles one person taking something off the shelf and handing it to another person.

Companies have been discussing "checkout-less" stores since forever, but nobody has been brave enough to do it due to the perceived threat of shoplifting.

And while shoplifting is a legitimate threat, are non-shoplifters going to be turned into shoplifters without a checkout? Are normal shoplifters stopped by checkouts? These are the core questions, and until it is tested nobody will know for sure.

Target is getting awfully close to this. With their Cartwheel app you're meant to scan all your items as you shop (so it auto-applies coupons and discounts); but they haven't taken it to the next logical step and allowed you to provide your Cartwheel output at the checkout for checking out.

I will say that the way Target has implemented smartphone barcode scanning makes me think that there might be a future in all this. It is extremely painless, they just need to stop kicking you out of the scan screen when it finds a discount (i.e. it doesn't kick you out if no discount is found, but does when a discount IS found, that's problematic for efficiency reasons).

In the UK, Tesco have been running a 'Scan as you Shop'[0] thing for a couple of years now. Customers pick up a scanner as they enter, scan their items as they go into their cart, and they have special checkouts which read your scanner.

There's a random chance that your scanner will be audited by a human against the contents of your shopping cart. Usually the first time you use it, then it backs off.

[0] http://www.tesco.com/scan-as-you-shop/

Tesco started that back in the late 90s and Sainsburys also did it 5+ years ago, then Tesco seems to have had another go at it recently.

edit: Safeway was first, in 1997

The Camden store illustrates the progress Safeway has made in other directions too. As part of its customer friendliness, Safeway was the first of the major food multiples to introduce self-scanning, the system it calls Shop & Go.


Our local store installed them to great fanfare back in the 90s, kept them for about a year, then quietly removed them. I was always very curious as to why: perhaps they didn't get used very much, or perhaps customer satisfaction with them was low.

A Waitrose I lived near had this system, and I think it must have been fairly successful because they put in a second wall of scanning devices shortly before I moved.

Offhandedly asking why next time you're at a (self-) checkout can't hurt.

Huh...I had travelled around the UK/Ireland last year and I don't think I saw any of these. Maybe I was hitting the wrong stores, or people were using them and I didn't recognize them.

They're only in big stores, and you could easily miss them if you weren't looking. Also most people don't use them.

It is typically found in very large stores. I only have only seen this in two Tesco stores.

Sam's Club has been doing similar for the last year[0], as well. They already have a built-in "check your receipt at the door" flow that customers are used to, so it wasn't that hard to offer an electronic tweak to the existing system.

[0] http://www.samsclub.com/sams/pagedetails/content.jsp?pageNam...

I used this for the first time two days ago. I was pleasantly surprised, and it felt really awesome --almost to a point of guilt-- skipping the checkout lines.

However, once I reached the exit where the man was inspecting receipts, things screeched to a halt. I showed him the bar code on my phone, and he looked at it, exclaiming, "Oh no, Scan and Go. You used Scan and Go."

He then turned to find a powered-off scanning device that he couldn't get to start up, as he muttered, "I wish they'd never started that."

I replied, "As a customer, I love it."

He had to call over a manager, who used her scanner to scan my barcode, blindly scroll through my purchased items list to get to the green button, and declared me good to go.

Meanwhile, I had held up all the people trying to exit behind me. It was still faster than checking out, but hopefully they get proper training for their employees implemented.

Tesco also has pushed self scanning tills for years now.

However, the number of employees working at the cash register is still the same because those scanners sometimes do not work and most importantly their user experience is deplorable. So, you frequently have to ask someone for help (and I'm in my mid-thirties and very tech-savvy. I can only imagine how someone twice my age would feel when using these scanners).

> However, the number of employees working at the cash register is still the same

Actually the ratio of staff:active-self-checkout machines is planned to reduce in each new deployment as customers become accustomed to the machines. Starts around 1:2 and usually sits around 1:4 with a target of 1:8 or even 1:12 at quiet times.

That's definitely a reduction in cashiers since the machines displace existing check-out lines & registers.

Source: an acquaintance is a manager in a Tesco Superstore.

Similar at supermarkets in Australia, one staff member could be watching over about 12 self checkouts. They still offer regular registers for people with a lot of stuff and people that don't like self checkout.

The machines are good enough now that I almost never get stuck, although some people are a lot more prone to needing staff assistance to complete checkout.

Terrible UX is spot-on.

I was using one of these just this morning: got all my shopping onto the scale/shelf and was getting ready to pay. The machine asks how many 5p bags I used. So I start packing the stuff into my backpack to find out whether I need a bag. The machine pipes up: "Did you remove something from the scale?" The screen has a full screen modal warning that I have to put the shopping back.

I put the things back on the scale and guess that I won't need a bag and I pay using contactless but what if my shopping won't fit in my bag?

I understand why we have to "pack" things onto a scale (it makes it much harder to take things without scanning them) but it has to trust you at some point.

Just say you brought your own bag and put the backpack on the scale. Then put things into the backpack as you scan them. Sometimes it'll request an attendant to confirm that it's empty (probably because it's heavier than most bags).

Yes, but i always think that will take ages ;-)

You're supposed to put your bag on the scale at the start when it asks if you brought bag(s). Fill it, then fill 5p bags if necessary.

You can blame that bad UX on the government, not the store. Next time you can just lie. The store doesn't care. They'd be happy to give you as many bags as you need for free if they were allowed.

Don't ever use a Chip + Signature credit card at a Tesco self-checkout, they had to reboot the thing when I accidentally used it instead of my regular UK Debit card.

Other than that I really like self-checkouts, usually much quicker and they're excellent for coin disposal (dump all your coins in it and pay the rest by card).

>> Don't ever use a Chip + Signature credit card at a Tesco self-checkout

Are UK banks giving these out? I've never come across them before.

They're standard issue from American banks these days. I believe you can request them from UK ones too if you have trouble remembering a PIN, though.

I used them in Tesco many times it's was horrible expirience too. Sometime they didn't work properly or failed to scan something.

What worse shortly after some stores in London got those there was less cashiers working at eventide. So you had to choose between unpleasant expirience or long queue.

Self-checkout registers in Finland have about 1 employee per 6 machines. They work pretty well, I haven't had any issues after the first few months they were launched.

> There's a random chance that your scanner will be audited by a human against the contents of your shopping cart. Usually the first time you use it, then it backs off.

I must be really suspicious-looking because I've been subjected to a "random audit" all 5 times I used it. Gave up because with a weekly shop it's quicker to go to an actual checkout than having someone scan all of my items again.

The idea's good, but my experience of the execution has been bad.

Stop & Shop in the NE US released this in ~2006

I worked on that project (Shopping Buddy) and a few later iterations of it. This was before the iPad, we had these huge IBM super rugged tablets on the carriages. We used something like Zigbee to triangulate the precise location of the carriage in the aisle, so we could tell whether the cart was at the beginning, middle or end of aisle 5 or aisle 6 and target ads appropriate for where you were. There was a small scanner for you to scan your groceries as you went. It was ridiculously expensive to outfit one store w/the tech so we ended up with a simpler version using Motorola MC70 handheld scanner guns and ditched the exact triangulation and instead went to more generic stuff where we would just check which AP you were connected to so we would know which third of the store you were on (produce, frozen foods, or middle of the store). That system then got ported to iOS and Android so you could use your phone instead of the scanner. The original version despite its issues was pretty advanced for the time (mid 2000s).

They still have the mobile / MC70 version rolled out to the entire chain now, it's very successful from what I hear.

Awesome insight! This is why I love HN.

My mother makes uses of it every week and gloats that we don't have the tech available here in Boulder (a tech hub).

Where is that at? I am Omaha and google maps did not show a "Stop & Shop".

It's a grocer in the Northeast.

Good one :). NE = North East, not Nebraska in this instance though I can see the confusion.

Waitrose have had the exact same thing for many years now.

This exists in the US, too. I saw one at Giant in Pennsylvania when I was home for the holidays.

Sounds great if it's not as unbelievably user-hostile as the self checkout machines. Those things are just total crap, every one I've ever tried.

Once I got the hang of it, I've had almost zero problems with self-checkout machines

It's usually one of two things: a purchase that must be approved (either alcohol or some medicine) or scanning the single item barcode of a multi-item

They screw up for me about 60% of the time. Alcohol is annoying but understandable. More problematic is the weight sensor that misreads the product after I set it down and forces me to wait for an assistant to click a button that says it's okay. I'll occasionally get the register that isn't calibrated properly and then I need assistance for every item. That's fun.

I have given up on bagging as I check out. It goes crazy every time I open a bag and set it in the bagging area. So now I just pile up my groceries and bag them when I'm done, which wastes everyone's time.

It's actually an interesting problem to solve. You would think that a particular item weighs the same no matter which store you scan it in, but in reality item weight can vary. Sometimes it is due to the manufacturer changing the packaging or adding a free sample of some other product to the item, but more interesting it could be the region. One chain I worked in had issues w/self-checkout on paper items such as toilet paper, paper towels, etc. The items were all weighed in the home office in Massachusetts and input into the master item file. This worked fine in most of the stores. The stores in Florida however always had issues because, due to the humidity in the air, the paper would absorb some of that and be just heavier enough to trip the scales. Eventually it was fixed by adding a larger weight variance allowance to the southern stores' POS registers.

> I have given up on bagging as I check out. It goes crazy every time I open a bag and set it in the bagging area. So now I just pile up my groceries and bag them when I'm done, which wastes everyone's time.

This. Trying to do the following

a) Open one of the cellophane bags that are stuck closely together

b) Put the product I just scanned in it

c) Take my hand away so it reads the weight correctly

...are way too difficult for me to do without the machine locking up and saying "okay, assistance required, let's hold you up until the overworked 1 person for 12 checkout machines gets here". So I do the same thing you do, just throw them in the checkout area and bag later.

It wastes everyone's time, but sometimes allows me to bag more efficiently and use fewer plastic bags.

We don't have plastic bags anymore in Seattle. The act of opening a paper bag and setting it on the scale (or simply opening a reusable bag that I'd already set on the scale) is enough to make me "require assistance" 99% of the time.

The ones in supermarkets in Australia started off with a lot of weight related issues. Seems to work fine now, not sure if they have gotten smarter or just less picky with weights.

At the mall where I live, Woolworths insists on bag weighing but Coles actually has it turned off!

This is awesome, as it's incredibly time-saving. At Woolworths I generally have to look like a doofus holding my next item to scan in front of the scanner for at least 2 seconds while the machine is frozen as it slowly weight-checks the item I just bagged...

At Coles the bottleneck is the speed at which I can bag things. Usually I just cram as many things into one bag as possible (I always repack later) and checking out takes under a minute.

NB: The Coles where I live is toward the back of the mall and reasonably far away from the center's exits. I think this has had an impact on the number of items that go walkies, which is why they were able to disable it. (The Woolworths on the other hand is practically outside - leaving there is like going through airport security, they physically rummage through your bags!)

I think there are bigger problems, like how they implicitly assume everyone is a thief and are will to damage usability for everyone because that assumption.

That stupid scale isn't going to stop any shoplifter, but it does inconvenience me every time.

EDIt - Grammar.

It's not just to catch thieves, but to catch errors in scanning too. When I was a cashier (this was back in 2004 but these self check outs appear to be running the same UI), about 70% of the time I had override the machine, the person had thought they scanned something but didn't. And it rarely appeared to be an attempt to steal.

I even catch myself doing it. If you are smart you can notice the error, take it out of the bag and rescan and the machine will keep going.

Most people just freeze and wait for help.

That is not when the machine messes up for me. If I just got done scanning the 5th case of some beverage and it needs to be weighed what should I do. The screen provides little hint. Sometimes there is a "remove my bags" button but often its not there. If I try to move goods to my cart it complains and stops the process, I put the stuff I Just scanned directly in my cart I get an error.

If you say that the scale does find honest mistakes (which I doubt, I think you just have dumb shoplifters) and I say the scales are strictly harmful to user experience then it seems these things must have a bad UI until some advance comes along.

Right. They're user-hostile.

I always either get erroneously told to put item in the bag, or have to take an item out of the bag because for some reason it shouldn't be there according to the register.

The secret is to hate the environment and just use the plastic/paper that is already there.

For me: When I am doing my weekly food shopping, I just use plastic. When I am grabbing something I forgot or picking up beer or something during the week, I just leave everything in the area, pay, and then bag with my reusables after the fact.

These days I only need human interaction when I buy booze, and that consists of a cashier eyeballing me and approving it.

Also when something's on sale but the non-sale barcode scans. I think employees are trained to cover the barcodes but it doesn't always happen.

Tesco's last generation machines are particularly bad, although their newest ones have removed a huge amount of latency from the UI and weighing scales from what I've seen.

If you're shopping for a single item, scanning it and dropping it in the bagging area is sufficient to start the flow, then simply wait a second, click finish, then tap your contactless card is all required to complete a transaction, can be done in under 5 seconds

The best machines I've seen in the UK in terms of user flow are probably at Waitrose, they're just supremely faster to use than those at Tesco

You get what you pay for :-)

The Sainsbury's near me have started to go contactless as well, and it's great.

I had problems with early generations of self-checkout devices, but I have no such problems now. In fact, I will use them whenever I can.

I used to work in a grocery store many years ago though, so it's possible that I have some skills kicking around in my brain that make it easier for me to use them.

That's surprising to hear. I love them (at normal grocers, not Sam's Club) because I find I am faster at scanning and checking myself out than at least half of the cashiers I encounter.

Almost all cashiers in my country are under orders to scan things slowly enough that the customer doesn't feel rushed to put them away. There are discount stores (Aldi and Lidl) that don't do this, and the cashiers there scan things much faster.

I wish more stores employed bag boys/girls. They speed the process up so much. It's a good first job for teenagers and the mentally disabled.

I found it depends on the 'class' of the supermarket. The higher class the supermarket, the less calibrated they are to treat you like a shoplifter.

In Waitrose you can use your own rucksack without getting flagged!

At M&S, you must wait to get approval after indicating that you're using your own bags. Awful system.

I've had nothing but trouble with the self-checkouts I've used in the UK. I'm quite slow and methodical about it, but the machines always manage to inexplicably fuck up halfway through.

I find in Tesco and the Coop (don't think I've used M&S) they let you say you're using your own bags, but they assume that's a carrier bag and complain at my rucksack. I always leave packing until the end if I can now, although they'll still complain about unexpected items even after paying..

I own a house in Bentonville, Arkanasas. I now live in Denver, but owning a property in Walmart's company town means I always pay careful attention to retail developments that threaten Walmart.

I know a lot about the various retail companies and how effective they are with technology. Anyone that thinks Target is going to do anything effective in this space doesn't understand how god-awful their entire logistics/supply chain is. It's legendarily bad.

Amazon is another story. And this particular technology is something that can destroy Walmart, eventually. I'll be watching this very closely, because if I see enough headway being made I'll be selling my house before the disaster hits the market there.

Honestly though, groceries and CPG are HARD. I suspect that Amazon has simply thought about cost-savings from having minimal staffing in a store and used it to justify the insane capital costs of an RFID tag on every item and the scanners/camera/compute needed to operate the Go store.

What's always so funny to me is how the millennial generation in general has a hugely negative view of Walmart for paying poorly and destroying small businesses, while having a positive view of Amazon. In this case, Amazon will continue to do what it's been doing (destroying businesses) while paying nothing, because it's automating away a huge segment of work. I'm not saying this is a bad thing, but it's definitely going to happen, and like self-driving vehicles, this is going to disrupt society big time.

> What's always so funny to me is how the millennial generation in general has a hugely negative view of Walmart for paying poorly and destroying small businesses, while having a positive view of Amazon.

The difference is that while both may offer low prices while destroying small businesses, the customer experience on Amazon is great while the customer experience at Walmart sucks.

>>Amazon is another story. And this particular technology is something that can destroy Walmart, eventually.

It's a nice innovation, but it does almost nothing with regards to competing on price with Walmart. And with the self checkouts, checking out isn't even that big of a deal. So i don't see how it's going to affect Walmart.

I actually read this completely differently with regards to shoplifting. How can you shoplift when you have multitudes of cameras watching your every movement and automatically assigning to you any products you take off the shelf? This is the dream anti-shoplifting system; it's automated loss prevention personnel.

It will flag clients for "suspicious" behavior, compare your face with known shoplifters in the last 10 years, and check if you are on the FBI most wanted list.

You'll be arrested before having picked up a single item.

When it is installed and working it could be part a "The Culture" like utopia. But even if it has an error rate as low as 0.01% and ruins the lives and careers of people because of technical mishaps it can make their lives a dystopic nightmare.

So, you're saying your non-member friends have to wait outside if you decide to pop into Go for a few items.

> Target is getting awfully close to this. With their Cartwheel app you're meant to scan all your items as you shop (so it auto-applies coupons and discounts); but they haven't taken it to the next logical step and allowed you to provide your Cartwheel output at the checkout for checking out.

Tesco in the UK is already doing that. You just finish your shopping by scanning an "end of shopping" barcode, you pay and then you go with your trolley. You sometimes need someone to remove the security tags or check that you're 18+ for alcohol but that's it.

However, home delivery is even better in my opinion, it's very common in the UK nowadays. You see those vans everywhere, from nearly all big supermarkets.

I'll admit, my first thoughts were "How can I exploit this system? How would I place sensors? What kind of sensors? Pressure plates? Cameras? IR trackers?" etc etc.

And I'm no shoplifter, never done that, I just tinker with stuff all the time. It's a kind of occupational psychosis. Same with lockpicks.

Yeah, but ultimately it's pretty easy to steal from a lot of regular stores too, and most people just...don't. Because they're not thieves.

If the cost of sensors + additional theft is less than the cost of the removed salespersons. Then the system makes business sense.

I thought it might be vulnerable to a human DOS-attack, where a flash mob of shoppers wearing similar cloths and/or masks all walk in at the same time, grab items, pass them around, and then all try to leave at once.

My take is that a foil-lined bag will bypass the whole system.

I think that could easily be detected if they wanted to. It sounds like they may be tracking your store movements and what items you pick up, so it would be odd if a shopper carrying items walked out of the store without registering any of the bought items.

This reminds me of the story about unmanned stores in Sweden https://www.washingtonpost.com/lifestyle/kidspost/sweden-ope... where you just walk in with a smartphone with a app, buy and checkout all by yourself

> are non-shoplifters going to be turned into shoplifters without a checkout?

One could respond that you can never truly shop lift if you follow the rules.

First, you have to check in when you enter. Second, they tell you to grab and go.

As long as you don't try to be malicious about it, then it's on Amazon to figure out what to charge you for. If they fail to see something, that's on them.

Presumably there will be some terms and conditions that require you to examine the receipt for the visit and make good any omissions. I don't suppose it would be impossible for Amazon to protect themselves from this kind of loss.

It seems to be that as long as you're malicious (I.e. deliberately hiding items), then that's on them. At least that's what the implied contract is to me. Otherwise it's just a "self checkout" app.

In Apple stores, you can buy anything off a shelf using an app on your smartphone and walk out with it without even needing to talk to an employee.

Of course, all the really pricey stuff like watches, phones, and computers is kept in a back room.

Amazon has the money to run the studies to prove whether it increases revenue. I suspect it wouldn't be too hard to both detect shoplifting and differentiate it from legitimate purposes using a modern high-end multisensor fusion machine learning system.

Better make sure to carefully balance the samples in your training set to avoid confounding variables though, or else your AI is going to get quite racist and sexist very quickly.

Here in Chicago the Jewel/Osco chain of grocery stores tried handheld scanners mated with the self-checkout systems about 10 years ago. I lived near one of the test stores.

It worked pretty well, but from what I heard the shoplifting rate was a lot higher than the comfort level of the store's executives. There really is no way to check that you've scanned every single thing in a large cart full of goods. They experimented with things like random checks from cashiers, but that just added to the labor and confusion. The project was scrapped a year or two later.

RFID could fix that if we're really close to being able to scan a whole cart full of goods in one sweep (which the Amazon project seems to imply). But then you have an issue where every single vendor to your store needs to be inserting compatible tags into the packaging. That adds cost and logistics.

> RFID could fix that if we're really close to being able to scan a whole cart full of goods in one sweep (which the Amazon project seems to imply). But then you have an issue where every single vendor to your store needs to be inserting compatible tags into the packaging. That adds cost and logistics.

I remember seeing this working when I was in university [0]. They just couldn't focus on supermarkets because the tags were too expensive (10-15 cents per tag).

[0] http://www.vilant.com/

Apple does it. You can use the Apple Store app to scan and pay and walk out.

> perceived threat of shoplifting

If that's the problem, couldn't these companies just test it? I mean, even if all the items in the single pilot store were stolen, would it really be such a big cost for Wallmart R&D department to have exact empiric knowledge as opposed to "perceived threat"?

I'm surprised people read Amazon Go as shoplifting friendly.

My first thought was: wow, and they made it much harder to shoplift. In a regular store you can just tuck an item in your jacket. With this tech, they know when the item left the shelf, that you were standing there, and that it left the store with you.

Sure, it's hackable. Everything is hackable. But this actually seems like an anti-shoplifting measure to me.

Heck, I could imagine them installing this system in a normally staffed store just to detect shoplifting!

Even if they gain more shrink/ loss by having customers scan their own items, the reduction in employee labor is almost certainly worth it.

real (part of the Metro group of retailers) and IKEA do it in Germany (and have been out of the evaluation phase for some years).

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact