You bring up a good point. Perhaps the potential savings from shoplifting, especially in the targeted urban markets, will pay for the technology/camera systems.
There is no mechanism to stop someone from walking out though- guards are under orders not to touch anyone. The old school dry goods store where you order at the counter is probably the only solution.
The tracking system basically has to be able to identify people in the store who haven't provided a valid credit card - the ability to track a person and know what their credit card (or lack of card) is.
If the store was literally unstaffed, I guess those cameras couldn't stop people, but they're not, so I assume these stores will do the same thing regular ones do, and rely on their employee's presence to discourage theft.
You are completely wrong about Hampton Inns. I spend 220 nights a year in Hotels. I’m Hilton Diamond. Mostly Hampton Inns. But also HI Express and Resident Inns and the like.
What people don’t realize there’s all these hotels basically handle your infrastructure workers of America. They aren’t filled up with businessman, will they have them, but largely they have technicians, road workers, railway workers, people who work on water plants, engineers, all the sort of specialties that are required to keep stuff working in America, but is so specialized of people have to travel. Usually, many of these people will clear out on Friday to go back home for the weekend, and that’s when the families come in. But Monday through Friday in any average Holiday Inn or Hampton Inn… This is where you truly essential workers are.
Trust me when I say, people absolutely do care about clean rooms. Now a lot of these hotels have gone to a service on demand model. Where are the only key in your hotel room in there for two days or three days, or upon request. This is reduce the labor requirements a little bit, but they are still having issues.
I think Matt means cleaning your room during your stay, as opposed to cleaning your room before your stay (and after the last person’s stay in that room). Everyone cares about the latter, people care less about the former.
If I'm somewhere for a stay of only a few days, I just leave the Do Not Disturb sign up non-stop, because I _actively dislike_ having the staff come in and mess with the room.
(In particular, what the hell is up with folding the sheets under the mattress along the side? Does _anyone_ do that at home?)
That's not what he meant, he meant while you're staying there. All hotels I've stayed at recently have stopped cleaning rooms every day unless you request it.
You can be Gold with a $95/year credit card or Diamond with the $450/year version. I have the latter; although Diamond isn't worth it to me, there's been enough perks to justify keeping it for this year at least.
Amazon’s system has a number of limitations you didn’t mention. You can only put items in the 2 bags in your shopping cart. You can’t just load up your cart, because the cameras can’t recognize that. So if you need more than 2 bags of food, you’re stuck.
Also, critically, because the cart has some expensive technology, you *can’t remove the cart from the store* so you either transfer to a dumb cart or have to carry it to your car.
Overall, I don’t think these trade offs are worth it vs a regular checkout. I think this system more works for a convenience store, like an Amazon Go. Although based on the ones I’ve been to, I don’t see much foot traffic there either.
Why would there be a two bag limit? It seems like the sensible thing would be to monitor the products and figure out who is doing the grabbing when someone picks something up.
I’m not armchair quarterbacking here- legitimately curious about the strategy behind the imaging rig plans.
These both seem like issues that will be solved with the benefit of time and/or innovation, though, no? Like, these are *now* problems, but probably not forever problems.
Well, I’d be careful about making easy predictions there. The history of improvements in computer vision technologies has been one of sudden jumps followed by long plateaus rather than a smooth slope upward. In 2010 everyone thought we had just a tiny bit more work to do before autonomous vehicles were ready for prime time, but reality had other ideas, and there’s no reason to expect that this problem will be any different.
There is one pretty significant difference between vision technology applied to autonomous vehicles and vision technology applied to groceries. The consequences for infrequent errors are massively different. (Potential death and destruction of valuable property vs. a free watermelon or the need for a single employee to process a transaction instead of the eye in the sky.)
Vision technology is more amenable to incremental but significant improvements in areas where minor errors in the technology do not lead to massive destruction and death.
Sure, no one will die if the ceiling cameras make a mistake. But one too many errors on your bill will insure that you go to Krogers and not Amazon Fresh.
It's difficult to predict, but human nature is that you're going to remember the time Amazon tried to stiff you for something you didn't buy more than 50 times they didn't charge you for a watermelon, and depending how hard they prosecute shoplifters, you may decide it's not worth the risk compared to an old-fashioned supermarket where that didn't happen.
Autonomous vehicles are ready for prime time now. They are already far safer than human drivers. The problem isn't technology, it's regulation and prejudice.
“Amazing, every word of what you just said was wrong.”
Sorry, I have dear friends who work at Waymo and…no. AV technology is somewhere between “intriguing demo” and “can work under limited circumstances in places with no inclement weather.” Whoever can make an autonomous car that can make it through a Boston winter will earn a mountain of cash but it hasn’t happened yet and we’re minimally 25 years away from it.
(If I were a betting man I’d expect that we’ll see semi-autonomous long-haul trucks pretty soon, but general adoption of it on passenger cars? Don’t bet your paycheck.)
Good, because as a suburbanite I don't really want Amazon coming in and eliminating competition that utilizes union labor. I also only have time to make bigger trips and prefer to be able to shop at a single store for everything rather than making multiple stops. Sounds like I am safe from Amazon.....for now.
I've always preferred that. In general I hate grocery shopping, but for whatever reason I find shorter, more frequent trips (especially off peak) a lot easier. Even when I owned a car this was the case.
Go is definitely a mini-mart, and has the product variety to prove it. Fresh is more traditional but with the expectation that you walk there and don't buy very much in one trip
This is not my experience at the one I went to this morning - you just put stuff in your backpack or whatever - I don't even think the store has carts, I've certainly never seen anyone using one. That being said, I agree with you that the Amazon Go is definitely a better use of the tech - my best case has been walking in, grabbing a soda and being out before my friends even finish walking down the block.
Interesting. This is more than enough to have me avoid Amazon Fresh stores. Maybe this will change in the future, but as of now this sounds like it's aimed at a niche market.
How big are these bags? I can't imagine a typical grocery trip where my purchase fits into two bags, at least not two bags that a normal person can lift, and especially once you throw in cases of soda and gallons of water and watermelons that don't tend to get bagged.
Out of curiosity, how does it deal with weighed products? Does the cart somehow weigh the items?
FWIW, I didn't even know this was a constraint because I physically can't carry more then two bags when I walk home. I don't think I've ever seen anyone using a cart in the store.
It doesn't deal with weighed products - only a pre packaged pound of beef or bag of 6 avocados. This is the worst thing about the store, the produce and such are pretty meh.
I live in the suburbs, so I drive to the store. I typically go to the grocery store once a week and come back with 4-6 tote bag sized bags of groceries, plus jugs of water, cases of soda, and multi-packs of paper towels. If I was limited to two bags, my "trip" to the store would probably be 4-5 trips.
In urban areas where these are being rolled out most people walk or take public transit to grocery shop and as a result have to shop more often, buying less stuff per visit
Great thing that I live in Philadelphia where this is impossible to do as in 2019 local city council lawmakers and the mayor signed a bill that bans cashless stores. They say it will help ease the day-to-day lives of many lower-income residents. I think it is bullshit for reasons I won't get into with this post.
Which is sad, as Philly was one of the few cities that were getting a flagship Amazon fresh store. And we need this type of innovation... especially grocery store innovation. There are plenty of food desert or desert-lite jawns where some innovative small grocery-store or co-op could do a lot of good. Sad that this city is rejecting this type of 21st century innovation.
The Mann Center - in Philadelphia - has a sign at the entrance that says, This is a cash free facility, and then directs you to where you can convert your cash to a debit card for use at the Center if you don't have your own plastic money.
Another angle here is that Amazon has seen a huge explosion in their grocery deliveries. These stores can double as last mile distribution centers and cut down delivery times. They’re working hard to get sub one hour delivery on groceries into large urban markets.
A lot of back and forth on the utility of self-checkout at the grocery store and I think it’s a reasonable point that the utility of it depends a lot on how many items you’re buying and how many of them need to be weighed rather than just scanned— for the moment both self-checkout and auto-checkout are really most effective in urban areas where you’re likely to be making more and smaller shopping trips.
But I think this points out where the puck is going: grocery stores are a dress rehearsal. Whoever can first make a serious dent in the checkout expertise at a big box store is going to make an unholy amount of money. Imagine a Saturday trip to Costco where you don’t spend 20 minutes in line to check out. Whoever builds that better mousetrap first will win big.
Another area where bigness can be good is health insurance. Only when an insurer reaches a certain behemoth size does it have the leverage to hold the line on rent-seeking price increases by providers. Insurers that become vertically integrated and directly acquire providers have more control to reduce inefficiency out of the provider system, as well as the incentive to do so. That's why a theme of health care reform efforts since the HMO fad of the Clinton years, through ACOs under Obamacare, up to today, has been to try to encourage providers to become risk bearing, i.e., to become whole or partial insurers. But small mom and pop physician practices and local hospitals can't easily manage risk like that - it takes size, to spread risk, to support a good data management infrastructure, etc.
I think that one of the main advantages (perhaps THE main advantage?) of healthcare in the EU is exactly this. Instead of having an army of administrators trying to figure out how much something costs and whom to bill for it, you just provide health services. Reducing make-work jobs reduces costs.
I was shocked in Belgium when, at the conclusion of my doctor visit, he said “that’ll be €25.” He pulled the change out of his middle drawer. He even took the simple X-ray of my son’s injured ankle himself. What a sensible system!
I had the opposite experience. A friend of mine here broke his leg and got an Uber to the hospital to avoid paying for ambulance. A few years after that, I needed to call the fire department here, and I was trying to figure out how much the call to 911 would be.
The general thrust of the various efforts here to move away from the fee-for-service model is to go in that direction, to push payment to providers up to a higher level of generality. So instead of providers presenting an itemized bill that lists each individual procedure, syringe and cotton swab used on the patient, they just get one fee for the entire episode of care and to cover all the details themselves out of that fee. Or better, a provider gets paid one fixed fee to cover all of a patient's healthcare needs for the year, and then it's up to them to do what's necessary during the year, instead of separately billing for each encounter with the patient.
"Or better, a provider gets ... encounter with the patient."
This is really the solution (more than Medicare for all or anything else). Out of curiosity, is there any actual movement in this direction in the US? I don't really know.
HMOs like Kaiser are designed to follow this model. They face two major obstacles: 1) People (conditioned by providers fighting against the model) (over)value freedom of choice in doctors and treatment facilities. I say this is over-valued because it's like trying to choose your pilot and plane when you want to take a trip. Quality has less to do with the individual characteristics of a physician or facility than the efficiency and quality control of the systems that support them. 2) Price competition on premiums. HMOs are in competition with fee for service plans that charge copays and deductibles, so the upfront premium costs are less. To match them, HMOs also have to charge copays and deductibles, which tends to undercut the model.
I'm a fan of Kaiser! I really hope their model would expand more.
1) Totally agree with you.
2) That's interesting. I learnt about the concept of deductible in healthcare once I moved to the US, and it seems very confusing to me. I can't understand why people would accept such a thing to be legal.
There is a logic to deductibles that comes from the principle of insurance as a hedge against risk. In this context, insurance shouldn't pay for everything, just those things that would put you out financially. As a practical matter, you don't expect your insurance to pay for things like Band-Aids and Tylenol - a deductible expands this a bit. You accept a deductible on your car insurance: I take the maximum one and pay a lower premium because I can afford to pay $1000 or so if I have an accident.
I've had nothing but excellent care at Kaiser for over ten years. There's a lot to be said for the model of "you broke your arm, the urgent care is available 24x7 at three regional locations, and your appointment for the ortho is on Tuesday, no choices."
All of their care is geared towards giving you the minimal possible to save costs and doing only evidence supported care. Works great and saves big bucks on the premiums for both me and my employer.
There are efforts by private insurers and CMS (at a snails pace) to push providers more toward that model (value-based payments, etc). But a precondition for risk-based payments to providers to really take off and replace fee for service, I think, is a sufficient number of large integrated healthcare delivery systems that can handle that. There has been consolidation on the health system side and some have ventured into risk plans such as with managed Mediciaid or Medicare Advantage or HMO plans, so there's some movement toward providers becoming insurers.
Recently the more notable movement has been insurers becoming providers (vertical integration from the top down instead of bottom up, so to speak). For example, the granddaddy of them all, UnitedHealth, now controls one of largest physician networks in the country through its Optum division, as well as surgery centers and other providers. And the combination of Aetna and CVS brings together an insurer and the MinuteClinic chain and other CVS providers.
In Maryland, I have heard Johns Hopkins just gets a yearly lump sum from the state. To work around this, they've moved procedures to satellite campuses that get reimbursed at a different rate, lol.
The theoretical advantage of private insurers is that they will play the role of doing the cost benefit analysis of medical treatments and drugs that doctors can't do and patients individually do not want done.
Yes, that's a function insurers can perform - to the extent people are willing to let them do it. But providers and patients have generally not been very tolerant of insurers that try hard to keep costs down, so insurers generally just shrug and pass on increased costs in the form of higher premiums.
Their main function historically though, has been to assume and spread risk that patients and providers are unable or unwilling to carry.
Right. We have not figured out how to get insurers to play the role that patients individually do not want them to play; they see the refusal to pay but not the lower cost of coverage. Medicare does not do it either. :(
Hampton Inn slam out of nowhere. Like, they aren’t the most up-market brand, but they’re not a Travelodge or Motel 6. People going there are going to expect clean sheets on the bed and fresh toiletries in the bathroom at least.
Really, if we are getting into a place where the labor market it too tight for Hampton Inn-level places to hire enough cleaning crews, I’d expect them to raise prices enough to cover the difference, and for a larger gap to open up between the low-mid market they occupy and the truly down-market places (who might actually eschew cleaning crews, and quickly become even dirtier and less attractive to all but the most desperate customers).
I think Matt was talking about when they clean rooms during your stay. Obviously they will clean rooms between stays and can do so with fewer workers if they're not worried about cleaning evey room every day.
Part of the issue with that idea is that cleaning is an area where delay can often compound the amount of work that needs to be done. If a guest stains the floor or a piece of furniture you're going to have a much harder time cleaning it >3 days later than you would within 24 hours of it happening. This is one of the reasons a lot of hotels will not let you go more than 2-3 days without letting in the cleaning crews.
I’m not sure a lot less. But I do prefer Hamptons. There is definitely variability in any area. There is a Holiday Inn Express in McAllen Texas that is just awesome. They do free food every night like a residence inn. Are usually choose a hotel based on how close it is to Starbucks. And the job site. With Hilton always getting the night over a Holiday Inn Express. But I will take a residence Inn or equivalent over anything if the job is more than three or four days.
Also, the thing I don’t care about. Breakfast. I haven’t eaten a hotel breakfast in three years. It just makes me nauseous thinking about it. I just fast until lunch.
I just stayed at a Hilton Garden Inn, which is nominally slightly more upmarket than a Hampton, and they have eschewed daily cleaning as well. I could have requested it, but honestly I prefer someone not messing with my stuff.
In reading this, I think Matt is ignoring a very important feature of Amazon‘s business model. Amazon essentially operates four separate but vertically integrated businesses: a platform serving other retailers, a retail business, a logistics apparatus that serves both, and an IT shop that is now primarily engaged in outsource cloud computing.
There is no way that the antitrust authorities of a prior era would look at this and think it’s a good thing. It allows for all sorts of competition-limiting behavior on Amazon’s part. You can hobble competitors by denying them access to services, steal crucial business information off AWS, subtly overcharge for logistic services to drive competition out of select markets… The list is endless.
And the reality is, Amazon’s ability to innovate is not contingent on all four of these business models being linked, at all. Amazon, the retailer could still quite effectively pursue the fresh model without having preferential access to AWS or Amazon, the retail platform
I think it’s possible to make cogent arguments in both directions here. I will go to my grave thinking that slipping the noose of antitrust investment in 2000 was actually a terrible deal for Microsoft shareholders and that splitting the company up into 3 or 4 independent units would have been substantially better for everyone than letting them hobble through the 00s in the moribund way they did. And frankly right now Google is in a much more similar place than anyone in the leadership team there wants to admit: the entire engineering arm is stuck in a defensive crouch around YouTube’s as revenue because it’s their only firewall against Facebook and their ability to manage their other products (always dicey to begin with) has evaporated. Divesting YouTube and android from Search and Ads would be a boon to both company and country.
But contrariwise, Amazon and Apple are firing on all thrusters and making vertical integration and scale work very very well for them and the benefits of divesting them from some of their portfolios seem pretty theoretical at best right now.
Stock performance is not relevant to anti-trust decisions.
Google absolutely should be forced to divest at least YouTube and Waze. Facebook should be stripped of Instagram. But it has nothing to do with how they impair/assist profitability.
Meanwhile, I really don't see much of a case for Apple being gutted in the same way; it should be forced to reform its App Store practices and the openness of the iOS ecosystem, but not much more is needed.
As for Amazon...
"the benefits of divesting them from some of their portfolios seem pretty theoretical at best right now."
Tell that to the retailers who are seeing popular goods cloned under the AmazonBasics brand as fast as they can develop them.
Or the would-be e-commerce competitors who can't overcome Amazon's pricing advantages because they don't have gobs of money from AWS to subsidize loss-making operations.
Or the dev and IT folks trying to find employment after leaving Amazon in Seattle.
My point about Microsoft's stock is a little different than you think it is: I'm saying that Microsoft's reason for trying to _avoid_ divestment (their profits and stock performance) turned out to be chimerical. Divestment would have been better for their shareholders _as well as_ their customers.
Having worked for a non-amazon Seattle-based tech company for several years myself, I'm a little curious for you to expand on why you think Amazon is having a bad effect on tech hiring in the pacific northwest or why looking for a tech job with Amazon on your resume is problematic. Or am I misunderstanding you?
(I'm really, really, really incapable of caring about Amazon Basics. Store brands have been a fact of life in the retail space for decades now, Amazon just brought the experience to a few verticals who weren't previously used to it: they'll cope. Brands fall over themselves to sell through Costco even though they know it means there will be a Kirkland equivalent within six months. This is not a problem the state needs to fix.)
(Likewise, in a world where more dumb VC money is sloshing around than ever before, I'm a little dubious that "amazon gets to cross-subsidize initially money-losing gambles" is actually a critical issue. If you have a dumb might-be-profitable-in-a-decade plan, Softbank is very much here for you.)
I understand your point about Microsoft/Google, and I'm saying it's entirely orthogonal to the discussion of anti-trust enforcement.
I don't care, and the anti-trust authorities shouldn't care, whether breaking up a vertically-integrated trust improves or degrades the investment performance of that firm and its successors.
As for Amazon, I find it a bit hard to be as cavalier as you're being. Gutting business models that require product development is a bit different from making pasta on an off-label. And when it comes to Seattle, I was just pointing out the local labor market monopsony within a few fields. Consolidation in tech has repeatedly been found to exert a depressive effect on wages in employment hubs like Seattle and the Bay Area.
I also see exactly no downsides to forcibly separating Amazon into "Amazon, the e-commerce platform and logistics outfit", "LifeBasics, the online retailer", and "AWS, the cloud service provider".
LifeBasics might or might not die within a few years without the reams of inside information that makes it possible, but the other two will be just fine.
I think you will find that food and apparel makers are just as sensitive about their product development budgets as USB cable makers -- nonetheless they've all found a way to live with Costco, Target and Walmart's store brands. They'll live with Amazon Basics as well.
I'm seriously and non-snarkily gonna ask for a citation on the wage issues. I'm very aware that _collusion_ between major tech employers has sometimes depressed wages (and indeed happily cashed my settlement check from the google/apple case as a result), but claiming that amazon _on its own_ is depressing engineering wages in the PNW seems like a bit of a stretch to me: Google, Microsoft and Oracle are all happy to bid for your time as a seattle-based engineer, as are a raft of well-financed startups. I'm willing to learn differently (lord knows I've been surprised by counterintuitive facts before) but this is the tightest market for engineers that I've seen in my lifetime.
I was using softbank more as a (notorious) example than as the be-all and end-all but yes, they have their fingers in a ridiculous number of pies right now including apparel and groceries. https://visionfund.com/portfolio. (And the Vision Fund isn't their only fund!)
"You can hobble competitors by denying them access to services, steal crucial business information off AWS..."
Do you have evidence that they actually do this? Hypotheticals are not a good reason for anti-trust enforcement. E.g. AWS seems to be perfectly happy to host Netflix even tho Prime video is a direct competitor. You think Amazon is stupid enough to break contracts with other companies by stealing crucial business information off AWS? What a bunch of absurd fear mongering.
Amazon already skims sales data from vendors on their platform and uses it to develop competitor products; it’s a good and a skip to “take a peek at something stored on our servers.”
Hypotheticals are a perfectly valid reason for anti-trust enforcement. Why allow the possibility to exist?
There are plenty of ongoing anti-competitive behaviors going on as is, so it’s not really necessary to resort to hypotheticals, but whatever.
"Amazon already skims sales data from vendors on their platform and uses it to develop competitor products"
Why do people bring this up as something nefarious? Every major retailer does this! I was just at a Kroger supermarket. They have multiple house brands. How does Kroger decide which products to make house-brand knockoffs of? They look at their sales data! Heaven forbid a company look at its sales data!
"it’s a good and a skip to “take a peek at something stored on our servers.”"
I have no idea what "a good and a skip" is, but the two things are very different. Amazon has contracts with the companies they sell cloud services to that says Amazon won't access the companies' data, and Amazon would lose huge amounts business if it turned out they were spying on cloud customers. Amazon does *not* have a contract with vendors saying they won't look at their sales data. Again, it's standard practice to look at sales data! Heck, I wouldn't be surprised if it's stated in the contract that Amazon *will* look at its sales data.
"Hypotheticals are a perfectly valid reason for anti-trust enforcement. Why allow the possibility to exist?"
Doesn't the possibility exist for *any* cloud computing provider to steal data from their customers? Why is this Amazon specific? Is it ok for the second largest cloud computing provider (Microsoft) to steal customer data but not Amazon? Should we then take preemptive anti-trust action against all cloud computing providers because they could hypothetically steal data?
If you can't show harm, then you don't get to take anti-trust action. If you think amazon is causing harm, then provide evidence rather than just spouting off about hypotheticals. If spouting off about hypotheticals is sufficient cause for legal penalties, then I submit that you could hypothetically be a child molester and we should lock you up just to be on the safe side.
I don’t know if this is because I got lucky in the grocery stores I’ve frequented since self-checkout became a thing, but in my opinion self-checkout is much, MUCH preferable to full-service checkout. Partly because I don’t need a human being standing there judging the vegetable:dessert ratio of my grocery purchases, but back in LA Ralphs used to send me coupons regularly and to use those I’d need to flag down an attendant anyway. The real reason self-checkout is better is because I know how to pack groceries into my backpack and grocery bag much, much more efficiently than any stranger, no matter how skilled. Also, no risk of the next person’s groceries rolling down the conveyor belt into mine while I’m still packing up.
If you're filling one backpack and one grocery bag, then I agree with you on self-checkout. If you're trying to check out an entire trolley (cart) and fill up a dozen or more bags and then load that all into your car, then that's inconvenient.
I've found it depends on the place. Places where people tend to work at the grocery store for a long time will often have really experienced and efficient staff, that can process goods faster than I ever could.
Places with higher turnover can be more hit and miss.
In Sweden, if you have a loyalty card, you can get a scanner at the supermarket, self-scan everything you buy and self-checkout. No cameras needed. There is a gate you have to scan your receipt to get out and there are random checks. All very convenient. Around 80% of shoppers use this system where I shop. I think this system exist in most of Europe too. The Amazon Fresh solution seems convoluted and kludgy to me, and creepy too.
The Big thing is interesting, since you can make a case that Amazon is getting to a level where you know so much about the consumer that price movements is no longer the most efficient indicator for allocating resources, hence planning potentially superseding market allocation.
I think Dan Karreman means the hand held scanners you carry with you (not the semi-supervised self checkout scanners) which I think people like, no? We've had these some 2014 or so at least.
Yep, that’s correct. I like self-checkouts too, but can see why you find them annoying. The scanner system basically eliminates lines, and fresh produce is a breeze since you put it on a scale, point at the image of what you are weighting , get a bar code, scan it and move on. Sure, it means you have to do some work but arguably not more than what is involved in getting the stuff from somebody (instructing, inspecting, and receiving the stuff).
That's interesting. I feel like I'm not really slowed down by anti-shoplifting measures, just that it's not very efficient when I need to unload my own cart, scan everything myself, bag it myself, and put it all back. What really slows it down though is that if I have any beer or wine or anything, I have to wait around for someone to come check my ID, which is a real pain. I do like it for when I just need to check out a small number of items though.
Most of the big supermarket chains in the US (and at this point even a fair number of the smaller regional ones) are already most of the way into transitioning to a similar self-checkout system. It’s generally pretty convenient but with one big caveat: nobody has really cracked the fresh produce issue yet. If there’s a real possible upside to the Amazon approach it’s that maybe someone will figure out how to get a camera to reliably discern cucumbers from zucchini.
I think that most modern European grocery stores (because I've been to a numbers there) have scales and attached bar code printers right in the produce section where you weight your produce in the usual plastic bag, type in plu code right there, it prints a bar code sticker you put on bag, and you can hand-held scanner that, or later scan it at either a manned or unmanned checkout lane. I have always liked this system.
If you have a hand scanner, it means you don't really have to wait in line (I think my store even has something where you can just link your card and don't have to check out at all). It also saves you the time and effort of loading everything onto the belt and you can just bag as you go.
Those exist in Whole Foods! It's used by the Amazon shoppers (who were getting lazy and not weighing produce) but it seems like such an easy thing to allow customers to use as well. I think it's the store being a bit over paranoid about the shoplifting potential
This is kinda how it works in China, too, but there's normally a store person to help you. Gotta say I'm rather "meh" about this system, as there's often a queue (And Chinese people are ruthless about cutting in line). Which translates into delays.
I normally don't mind having to type in the PLU codes. What I find annoying is the times produce has a bar code that scans by accident, incorrectly, and then I have to wait for an employee to come over. Or the times the scale guesses wrong about how much something should weigh. I do wonder if there would actually be a significant increase in losses if they just turned off the bagging area scales at self-checkout. I'd move much faster if I could bag directly in my cart.
Cucumbers from zucchini is hard, but I think it'll be even longer before it can tell whether I bought gala, fuji, or honeycrisp apples, unless they're bagged with distinctive packaging, which I don't want.
Well, the Amazon Fresh model certainly understands what bins certain varieties of produce are supposed to be in. They don't necessarily need to be able to distinguish apple varieties visually if they can tell that you took one apple from the honeycrisp bin.
(Also, I'd be surprised if Amazon wasn't also planning to use your previous behavior to build a probabilistic model of what you purchased. If they're like, "Oh shit, our computer vision is having difficulty telling if you grabbed a fiji or a honeycrisp," they say, "well, the last five times he bought an apple, it was a honeycrisp.")
Yeah. For that I think they’re going to have to break down and put RFID tags in the stickers, and with the caveat that I have no idea what the market is currently like for such things I suspect that it’s going to be a prohibitively margin-eating cost for a while.
I just bought running shorts at Target using self-checkout. No problem! Except that I didn't notice they had RFID tags on them until I got home (didn't trip the door alarm, of course). So I would have to drive back to the store to ask them to remove the tags and prove that I wasn't trying to cheat them.
Or I could pound on the tags for five minutes with a hammer until they finally sprang open.
Yup, there are advantages to having human cashiers.
That has been the main argument against them for at least the past ~9 years since I first looked into it, not sure how much closer we've gotten to feasibility.
This whole piece boils down to "big companies have economies of scale and that's efficient, and sometimes consumers benefit from this!"
Everyone knows this already, including the economists that want to fight Amazon with anti-trust efforts. The problem is Matt misses half the picture, which is that there are trade-offs: fundamentally, between efficiency and resiliency, and more specifically, trusting Amazon to keep the consumer happy once they have achieved complete dominance in market. I don't see any engagement with this, just "big efficient, efficient good!"
The pro-Amazon take is it will never be possible for them to have complete dominance. Target, Wal-Mart and others are still out there and it doesn't seem plausible they can run them all out of business.
Every time I leave the US I'm mesmerized by, as I walk through a typical neighborhood, the menagerie of small grocers, bakeries, and butchers. It's deeply sad to me that as Americans we only get to drive a few miles and then choose between two Everything Stores, and because there's not literally only one option we consider this acceptable competition.
I agree with the feeling, but Matt has convinced me this is pastoralism. We want customers to get good prices and we want workers paid well. Those things usually come from chains, because they have better margins.
I feel that I can't oppose residential NIMBYism, but support commercial NIMBYism.
I'm always amazed how those small grocery stores in Manhattan survive but then I think about the density and what the volume of customers per hour must be for them, and then it starts to make sense.
I'm very much looking forward to the day when we can avoid all those interactions with "humans" most definitely including those friendly ones Trader Joe's hires.
I love Trader Joe's and it's the best check-out experience I have (except maybe Whole Foods, but I am rarely at Whole Foods.) It's delivery that would dramatically improve things -- which I guess is not a technology issue.
1. Matt Stoller is one of the absolutely dumbest people on the twitter.com. When he's not fellating Josh Hawley because he says "Amazon bad", he spends the rest of his time being extremely wrong about everything and it's mystifying why anyone pays any attention to him.
2. Years ago, Food Lion, a large chain of middle of the market grocery stores, decided to re-brand their stores based on the income levels of the communities in which they were located. higher income area stores became Bloom, middle income stores staid Food Lion, and lower income communities got Bottom Dollar Food.
The concepts failed after a few years, but Bloom tried to do what Amazon is doing now nearly 15 years ago. You could grab a self-scanner when you entered the store, then walk around and scan and bag your own stuff as you shopped, and at the end you simply walked up to a self-serve kiosk to return your scanner and pay. It was a great concept and I was disappointed when it disappeared along with the Bloom branding. That amazon has taken this a step further and can autobill you as you walk of the store is a no-brainer of a concept.
3. I find self-checkout awful. Unless I've got less than 5 items, I avoid self-checkout at all costs. Its slow, the area is too small to effectively organize your groceries, and it inevitably yells at you for not putting the items down quickly enough or too quickly into your bag and needs a manager to unlock the terminal.
4. I try to support unionized labor anytime I can, but here in DC I find the unionized grocery stores to be generally terrible, with surly, unhelpful employees and far more spoiled produce. I'm glad there are unionized stores in town because it likely inflates the wages at the non-unionized Harris Teeter I shop at, but every Safeway in DC is a disaster, and Giant is only marginally better.
Our local mega-chain, GIant, does have those gun scanners which my partner loves but I hate because I prefer a cashier at checkout rather than force me to remember what variety of tomatoes or rolls I picked and what their cost code was.
The history of grocery shopping is one of pushing more and more labor onto the customer. 19th century dry goods stores essentially had everything behind the counter. A&P's innovation was to allow customers to select their items from shelves themselves. Self-checkout enlists the customers to do their own scanning and bagging. This has always struck me as a poor form of specialization as the general public will never be as fast or efficient as checkout clerks who do it all day long. Some stores still have bag boys (who in many areas historically worked just for tips) but even this form of labor division is vanishing. Since the store pockets the saved labor from self-checkout, the inefficiency is a time tax on the customer.
I'm not sure how Fresh works in this paradigm as it is replacing check-out machines and scanners with just a different technology.
To put it mildly, my experience is diametrically opposed. Being a checkout clerk is boring work that requires staying on your feet for extended periods of time: it’s a job designed to produce boredom, disengagement and ennui, and as a customer unless I am buying a large amount of produce (where the clerk having memorized the product codes for the most popular items will mean that they can’t help but be faster than me) I will almost always aim for the self-checkout line and expect that I will save quite a bit of time by doing so.
Both this post and the replies to it are missing the mark.
The factor holding things back has always been theft losses.
General stores with a sole proprietor had that one person and a physical barrier between customers and products primarily to prevent theft.
A&P's innovation was realizing that with modern policing and low crime rates, they could set up a checkpoint between customers and the exit to permit customers to get on with it. Which everyone very clearly wanted; customers didn't view it as taking on labor, they viewed it as speeding up the process compared to going to the grocer, butcher, and produce market in turn and waiting to be served.
Self-checkout only became widespread once camera surveillance started to limit "self-discount", and it's become clear that people mostly want it as well, especially for small-to-medium purchases.
It might be faster to have clerks grab things from behind the counter when there's just one kind of everything and not that many choices in total. But when there are many varieties and brands and you need to browse to figure out what you want, not so much.
I think an Amazon warehouse/instacart model would probably be the most efficient for car centric grocery shopping. You go online, punch in what you want, drive to the store, and just load it into the back of your car and leave. The inside of the store/warehouse is a mess where only the computers know how things are laid out, but people and machines move things around according to whatever the computer tells them to do.
That said, this model works best for packaged goods. For produce, consumers want to be able to inspect the goods and decide on the best whatever. So maybe just have two stores, one for just produce and another for packaged goods.
I hear they are already doing this at Bay Area CVS stores. People walk in, take what they want, and leave.
CVS is having a lot of "accidentally forgot to charge you for that watermelon" problems with this pilot program.
Ouch.
You bring up a good point. Perhaps the potential savings from shoplifting, especially in the targeted urban markets, will pay for the technology/camera systems.
There is no mechanism to stop someone from walking out though- guards are under orders not to touch anyone. The old school dry goods store where you order at the counter is probably the only solution.
The mechanism is that you have the guy's credit card and charge him for whatever he takes.
You can't rely on customers to prevent door kiting, so I am not sure what good that does.
The tracking system basically has to be able to identify people in the store who haven't provided a valid credit card - the ability to track a person and know what their credit card (or lack of card) is.
If the store was literally unstaffed, I guess those cameras couldn't stop people, but they're not, so I assume these stores will do the same thing regular ones do, and rely on their employee's presence to discourage theft.
Which doesn't deter people who are there to get stuff in bulk for a fence, which is the topic of this thread.
reducing shoplifting, I meant to say
You are completely wrong about Hampton Inns. I spend 220 nights a year in Hotels. I’m Hilton Diamond. Mostly Hampton Inns. But also HI Express and Resident Inns and the like.
What people don’t realize there’s all these hotels basically handle your infrastructure workers of America. They aren’t filled up with businessman, will they have them, but largely they have technicians, road workers, railway workers, people who work on water plants, engineers, all the sort of specialties that are required to keep stuff working in America, but is so specialized of people have to travel. Usually, many of these people will clear out on Friday to go back home for the weekend, and that’s when the families come in. But Monday through Friday in any average Holiday Inn or Hampton Inn… This is where you truly essential workers are.
Trust me when I say, people absolutely do care about clean rooms. Now a lot of these hotels have gone to a service on demand model. Where are the only key in your hotel room in there for two days or three days, or upon request. This is reduce the labor requirements a little bit, but they are still having issues.
I think Matt means cleaning your room during your stay, as opposed to cleaning your room before your stay (and after the last person’s stay in that room). Everyone cares about the latter, people care less about the former.
If I'm somewhere for a stay of only a few days, I just leave the Do Not Disturb sign up non-stop, because I _actively dislike_ having the staff come in and mess with the room.
(In particular, what the hell is up with folding the sheets under the mattress along the side? Does _anyone_ do that at home?)
Yeah I was confused about MY’s comment about not cleaning the hotel room. I also agree that ppl really care about that!
Agreed. Sometimes he just tosses off comments which are, shall we say, not evidence-based.
"No Room Cleaning for Stay less than 7 Days" without paying an extra fee at Hilton Grand Vacations on Paradise (Convention Center) Las Vegas
It was a very strange comment. “They’re just not going to clean them.” Wut? Even the shadiest motel cleans the rooms, even if not terribly well.
That's not what he meant, he meant while you're staying there. All hotels I've stayed at recently have stopped cleaning rooms every day unless you request it.
That's a heckuva lot of time on the road!
220-250 days a year.
Damn, don't think I could hack that! Must be interesting to see so much of the country once you get used to it, though.
It sounds more glamorous than it is. It’s usually 12 hour shifts 7 days a week. In obscure places. But it’s not the worst job.
You can be Gold with a $95/year credit card or Diamond with the $450/year version. I have the latter; although Diamond isn't worth it to me, there's been enough perks to justify keeping it for this year at least.
Amazon’s system has a number of limitations you didn’t mention. You can only put items in the 2 bags in your shopping cart. You can’t just load up your cart, because the cameras can’t recognize that. So if you need more than 2 bags of food, you’re stuck.
Also, critically, because the cart has some expensive technology, you *can’t remove the cart from the store* so you either transfer to a dumb cart or have to carry it to your car.
Overall, I don’t think these trade offs are worth it vs a regular checkout. I think this system more works for a convenience store, like an Amazon Go. Although based on the ones I’ve been to, I don’t see much foot traffic there either.
Hey Zach. I'm an engineer at Amazon who works on this stuff.
You're mixing up Dash Carts (https://www.amazon.com/b?ie=UTF8&node=21289116011) with Just Walk Out (https://justwalkout.com/). Some Fresh stores use one experience, some use the other.
Why would there be a two bag limit? It seems like the sensible thing would be to monitor the products and figure out who is doing the grabbing when someone picks something up.
I’m not armchair quarterbacking here- legitimately curious about the strategy behind the imaging rig plans.
These both seem like issues that will be solved with the benefit of time and/or innovation, though, no? Like, these are *now* problems, but probably not forever problems.
Well, I’d be careful about making easy predictions there. The history of improvements in computer vision technologies has been one of sudden jumps followed by long plateaus rather than a smooth slope upward. In 2010 everyone thought we had just a tiny bit more work to do before autonomous vehicles were ready for prime time, but reality had other ideas, and there’s no reason to expect that this problem will be any different.
There is one pretty significant difference between vision technology applied to autonomous vehicles and vision technology applied to groceries. The consequences for infrequent errors are massively different. (Potential death and destruction of valuable property vs. a free watermelon or the need for a single employee to process a transaction instead of the eye in the sky.)
Vision technology is more amenable to incremental but significant improvements in areas where minor errors in the technology do not lead to massive destruction and death.
Sure, no one will die if the ceiling cameras make a mistake. But one too many errors on your bill will insure that you go to Krogers and not Amazon Fresh.
Not sure, but it seems more likely that errors would accrue to the benefit of the consumer (e.g. missing an item like the watermelon example).
It's difficult to predict, but human nature is that you're going to remember the time Amazon tried to stiff you for something you didn't buy more than 50 times they didn't charge you for a watermelon, and depending how hard they prosecute shoplifters, you may decide it's not worth the risk compared to an old-fashioned supermarket where that didn't happen.
For sure the risk profile is more friendly to experimentation!
Autonomous vehicles are ready for prime time now. They are already far safer than human drivers. The problem isn't technology, it's regulation and prejudice.
“Amazing, every word of what you just said was wrong.”
Sorry, I have dear friends who work at Waymo and…no. AV technology is somewhere between “intriguing demo” and “can work under limited circumstances in places with no inclement weather.” Whoever can make an autonomous car that can make it through a Boston winter will earn a mountain of cash but it hasn’t happened yet and we’re minimally 25 years away from it.
(If I were a betting man I’d expect that we’ll see semi-autonomous long-haul trucks pretty soon, but general adoption of it on passenger cars? Don’t bet your paycheck.)
Seems ideal then for urbanites. I only walk to get groceries so I can't carry more than two bags anyway
Good, because as a suburbanite I don't really want Amazon coming in and eliminating competition that utilizes union labor. I also only have time to make bigger trips and prefer to be able to shop at a single store for everything rather than making multiple stops. Sounds like I am safe from Amazon.....for now.
Now that we don't have minor children at home, its good for us as well. We make daily store trips instead of one big one.
I've always preferred that. In general I hate grocery shopping, but for whatever reason I find shorter, more frequent trips (especially off peak) a lot easier. Even when I owned a car this was the case.
Hey Zack. Thanks for that info. So basically it’s a mini-mart not a grocery store.
Go is definitely a mini-mart, and has the product variety to prove it. Fresh is more traditional but with the expectation that you walk there and don't buy very much in one trip
The one in San Francisco I just stuffed stuff in my pockets
This is not my experience at the one I went to this morning - you just put stuff in your backpack or whatever - I don't even think the store has carts, I've certainly never seen anyone using one. That being said, I agree with you that the Amazon Go is definitely a better use of the tech - my best case has been walking in, grabbing a soda and being out before my friends even finish walking down the block.
Interesting. This is more than enough to have me avoid Amazon Fresh stores. Maybe this will change in the future, but as of now this sounds like it's aimed at a niche market.
How big are these bags? I can't imagine a typical grocery trip where my purchase fits into two bags, at least not two bags that a normal person can lift, and especially once you throw in cases of soda and gallons of water and watermelons that don't tend to get bagged.
Out of curiosity, how does it deal with weighed products? Does the cart somehow weigh the items?
FWIW, I didn't even know this was a constraint because I physically can't carry more then two bags when I walk home. I don't think I've ever seen anyone using a cart in the store.
It doesn't deal with weighed products - only a pre packaged pound of beef or bag of 6 avocados. This is the worst thing about the store, the produce and such are pretty meh.
I live in the suburbs, so I drive to the store. I typically go to the grocery store once a week and come back with 4-6 tote bag sized bags of groceries, plus jugs of water, cases of soda, and multi-packs of paper towels. If I was limited to two bags, my "trip" to the store would probably be 4-5 trips.
I could see one doing well in the little downtowns suburbs have or even a Mall.
In urban areas where these are being rolled out most people walk or take public transit to grocery shop and as a result have to shop more often, buying less stuff per visit
Like you can bring your own bags it is fun I had a big old backpack
Great thing that I live in Philadelphia where this is impossible to do as in 2019 local city council lawmakers and the mayor signed a bill that bans cashless stores. They say it will help ease the day-to-day lives of many lower-income residents. I think it is bullshit for reasons I won't get into with this post.
Which is sad, as Philly was one of the few cities that were getting a flagship Amazon fresh store. And we need this type of innovation... especially grocery store innovation. There are plenty of food desert or desert-lite jawns where some innovative small grocery-store or co-op could do a lot of good. Sad that this city is rejecting this type of 21st century innovation.
The Mann Center - in Philadelphia - has a sign at the entrance that says, This is a cash free facility, and then directs you to where you can convert your cash to a debit card for use at the Center if you don't have your own plastic money.
I'll guess that the approximate number of Mann Center patrons without debit or credit card is 0.
I wonder if people would rather go back to gas attendants. Was that better? I hate driving through Jersey for this very reason.
Mandatory gas attendants as in NJ and OR are maddening. That said, when the weather is terrible it’s nice to have the option.
Another angle here is that Amazon has seen a huge explosion in their grocery deliveries. These stores can double as last mile distribution centers and cut down delivery times. They’re working hard to get sub one hour delivery on groceries into large urban markets.
A lot of back and forth on the utility of self-checkout at the grocery store and I think it’s a reasonable point that the utility of it depends a lot on how many items you’re buying and how many of them need to be weighed rather than just scanned— for the moment both self-checkout and auto-checkout are really most effective in urban areas where you’re likely to be making more and smaller shopping trips.
But I think this points out where the puck is going: grocery stores are a dress rehearsal. Whoever can first make a serious dent in the checkout expertise at a big box store is going to make an unholy amount of money. Imagine a Saturday trip to Costco where you don’t spend 20 minutes in line to check out. Whoever builds that better mousetrap first will win big.
Another area where bigness can be good is health insurance. Only when an insurer reaches a certain behemoth size does it have the leverage to hold the line on rent-seeking price increases by providers. Insurers that become vertically integrated and directly acquire providers have more control to reduce inefficiency out of the provider system, as well as the incentive to do so. That's why a theme of health care reform efforts since the HMO fad of the Clinton years, through ACOs under Obamacare, up to today, has been to try to encourage providers to become risk bearing, i.e., to become whole or partial insurers. But small mom and pop physician practices and local hospitals can't easily manage risk like that - it takes size, to spread risk, to support a good data management infrastructure, etc.
I think that one of the main advantages (perhaps THE main advantage?) of healthcare in the EU is exactly this. Instead of having an army of administrators trying to figure out how much something costs and whom to bill for it, you just provide health services. Reducing make-work jobs reduces costs.
I was shocked in Belgium when, at the conclusion of my doctor visit, he said “that’ll be €25.” He pulled the change out of his middle drawer. He even took the simple X-ray of my son’s injured ankle himself. What a sensible system!
I had the opposite experience. A friend of mine here broke his leg and got an Uber to the hospital to avoid paying for ambulance. A few years after that, I needed to call the fire department here, and I was trying to figure out how much the call to 911 would be.
The general thrust of the various efforts here to move away from the fee-for-service model is to go in that direction, to push payment to providers up to a higher level of generality. So instead of providers presenting an itemized bill that lists each individual procedure, syringe and cotton swab used on the patient, they just get one fee for the entire episode of care and to cover all the details themselves out of that fee. Or better, a provider gets paid one fixed fee to cover all of a patient's healthcare needs for the year, and then it's up to them to do what's necessary during the year, instead of separately billing for each encounter with the patient.
"Or better, a provider gets ... encounter with the patient."
This is really the solution (more than Medicare for all or anything else). Out of curiosity, is there any actual movement in this direction in the US? I don't really know.
HMOs like Kaiser are designed to follow this model. They face two major obstacles: 1) People (conditioned by providers fighting against the model) (over)value freedom of choice in doctors and treatment facilities. I say this is over-valued because it's like trying to choose your pilot and plane when you want to take a trip. Quality has less to do with the individual characteristics of a physician or facility than the efficiency and quality control of the systems that support them. 2) Price competition on premiums. HMOs are in competition with fee for service plans that charge copays and deductibles, so the upfront premium costs are less. To match them, HMOs also have to charge copays and deductibles, which tends to undercut the model.
I'm a fan of Kaiser! I really hope their model would expand more.
1) Totally agree with you.
2) That's interesting. I learnt about the concept of deductible in healthcare once I moved to the US, and it seems very confusing to me. I can't understand why people would accept such a thing to be legal.
There is a logic to deductibles that comes from the principle of insurance as a hedge against risk. In this context, insurance shouldn't pay for everything, just those things that would put you out financially. As a practical matter, you don't expect your insurance to pay for things like Band-Aids and Tylenol - a deductible expands this a bit. You accept a deductible on your car insurance: I take the maximum one and pay a lower premium because I can afford to pay $1000 or so if I have an accident.
I've had nothing but excellent care at Kaiser for over ten years. There's a lot to be said for the model of "you broke your arm, the urgent care is available 24x7 at three regional locations, and your appointment for the ortho is on Tuesday, no choices."
All of their care is geared towards giving you the minimal possible to save costs and doing only evidence supported care. Works great and saves big bucks on the premiums for both me and my employer.
There are efforts by private insurers and CMS (at a snails pace) to push providers more toward that model (value-based payments, etc). But a precondition for risk-based payments to providers to really take off and replace fee for service, I think, is a sufficient number of large integrated healthcare delivery systems that can handle that. There has been consolidation on the health system side and some have ventured into risk plans such as with managed Mediciaid or Medicare Advantage or HMO plans, so there's some movement toward providers becoming insurers.
Recently the more notable movement has been insurers becoming providers (vertical integration from the top down instead of bottom up, so to speak). For example, the granddaddy of them all, UnitedHealth, now controls one of largest physician networks in the country through its Optum division, as well as surgery centers and other providers. And the combination of Aetna and CVS brings together an insurer and the MinuteClinic chain and other CVS providers.
In Maryland, I have heard Johns Hopkins just gets a yearly lump sum from the state. To work around this, they've moved procedures to satellite campuses that get reimbursed at a different rate, lol.
https://en.wikipedia.org/wiki/Maryland_hospital_payment_system
The theoretical advantage of private insurers is that they will play the role of doing the cost benefit analysis of medical treatments and drugs that doctors can't do and patients individually do not want done.
Yes, that's a function insurers can perform - to the extent people are willing to let them do it. But providers and patients have generally not been very tolerant of insurers that try hard to keep costs down, so insurers generally just shrug and pass on increased costs in the form of higher premiums.
Their main function historically though, has been to assume and spread risk that patients and providers are unable or unwilling to carry.
Right. We have not figured out how to get insurers to play the role that patients individually do not want them to play; they see the refusal to pay but not the lower cost of coverage. Medicare does not do it either. :(
Hampton Inn slam out of nowhere. Like, they aren’t the most up-market brand, but they’re not a Travelodge or Motel 6. People going there are going to expect clean sheets on the bed and fresh toiletries in the bathroom at least.
Really, if we are getting into a place where the labor market it too tight for Hampton Inn-level places to hire enough cleaning crews, I’d expect them to raise prices enough to cover the difference, and for a larger gap to open up between the low-mid market they occupy and the truly down-market places (who might actually eschew cleaning crews, and quickly become even dirtier and less attractive to all but the most desperate customers).
I think Matt was talking about when they clean rooms during your stay. Obviously they will clean rooms between stays and can do so with fewer workers if they're not worried about cleaning evey room every day.
Part of the issue with that idea is that cleaning is an area where delay can often compound the amount of work that needs to be done. If a guest stains the floor or a piece of furniture you're going to have a much harder time cleaning it >3 days later than you would within 24 hours of it happening. This is one of the reasons a lot of hotels will not let you go more than 2-3 days without letting in the cleaning crews.
You can always count on the Hampton Stans to rally in defense!
Hey. I’m Diamond Hilton. Spire Holiday Inn. Gold Marriott. I do like Hampton Inns over HI Express.
I'm Diamond with Hilton too (by credit card, not nights). And every time I check in, I can't help but think of this video: https://www.youtube.com/watch?v=WCqm4H3m3Ew
Same. Express is a lot less consistent, in terms of cleanliness, service etc.
I’m not sure a lot less. But I do prefer Hamptons. There is definitely variability in any area. There is a Holiday Inn Express in McAllen Texas that is just awesome. They do free food every night like a residence inn. Are usually choose a hotel based on how close it is to Starbucks. And the job site. With Hilton always getting the night over a Holiday Inn Express. But I will take a residence Inn or equivalent over anything if the job is more than three or four days.
Also, the thing I don’t care about. Breakfast. I haven’t eaten a hotel breakfast in three years. It just makes me nauseous thinking about it. I just fast until lunch.
I just stayed at a Hilton Garden Inn, which is nominally slightly more upmarket than a Hampton, and they have eschewed daily cleaning as well. I could have requested it, but honestly I prefer someone not messing with my stuff.
In reading this, I think Matt is ignoring a very important feature of Amazon‘s business model. Amazon essentially operates four separate but vertically integrated businesses: a platform serving other retailers, a retail business, a logistics apparatus that serves both, and an IT shop that is now primarily engaged in outsource cloud computing.
There is no way that the antitrust authorities of a prior era would look at this and think it’s a good thing. It allows for all sorts of competition-limiting behavior on Amazon’s part. You can hobble competitors by denying them access to services, steal crucial business information off AWS, subtly overcharge for logistic services to drive competition out of select markets… The list is endless.
And the reality is, Amazon’s ability to innovate is not contingent on all four of these business models being linked, at all. Amazon, the retailer could still quite effectively pursue the fresh model without having preferential access to AWS or Amazon, the retail platform
I think it’s possible to make cogent arguments in both directions here. I will go to my grave thinking that slipping the noose of antitrust investment in 2000 was actually a terrible deal for Microsoft shareholders and that splitting the company up into 3 or 4 independent units would have been substantially better for everyone than letting them hobble through the 00s in the moribund way they did. And frankly right now Google is in a much more similar place than anyone in the leadership team there wants to admit: the entire engineering arm is stuck in a defensive crouch around YouTube’s as revenue because it’s their only firewall against Facebook and their ability to manage their other products (always dicey to begin with) has evaporated. Divesting YouTube and android from Search and Ads would be a boon to both company and country.
But contrariwise, Amazon and Apple are firing on all thrusters and making vertical integration and scale work very very well for them and the benefits of divesting them from some of their portfolios seem pretty theoretical at best right now.
Stock performance is not relevant to anti-trust decisions.
Google absolutely should be forced to divest at least YouTube and Waze. Facebook should be stripped of Instagram. But it has nothing to do with how they impair/assist profitability.
Meanwhile, I really don't see much of a case for Apple being gutted in the same way; it should be forced to reform its App Store practices and the openness of the iOS ecosystem, but not much more is needed.
As for Amazon...
"the benefits of divesting them from some of their portfolios seem pretty theoretical at best right now."
Tell that to the retailers who are seeing popular goods cloned under the AmazonBasics brand as fast as they can develop them.
Or the would-be e-commerce competitors who can't overcome Amazon's pricing advantages because they don't have gobs of money from AWS to subsidize loss-making operations.
Or the dev and IT folks trying to find employment after leaving Amazon in Seattle.
Kill it with fire.
My point about Microsoft's stock is a little different than you think it is: I'm saying that Microsoft's reason for trying to _avoid_ divestment (their profits and stock performance) turned out to be chimerical. Divestment would have been better for their shareholders _as well as_ their customers.
Having worked for a non-amazon Seattle-based tech company for several years myself, I'm a little curious for you to expand on why you think Amazon is having a bad effect on tech hiring in the pacific northwest or why looking for a tech job with Amazon on your resume is problematic. Or am I misunderstanding you?
(I'm really, really, really incapable of caring about Amazon Basics. Store brands have been a fact of life in the retail space for decades now, Amazon just brought the experience to a few verticals who weren't previously used to it: they'll cope. Brands fall over themselves to sell through Costco even though they know it means there will be a Kirkland equivalent within six months. This is not a problem the state needs to fix.)
(Likewise, in a world where more dumb VC money is sloshing around than ever before, I'm a little dubious that "amazon gets to cross-subsidize initially money-losing gambles" is actually a critical issue. If you have a dumb might-be-profitable-in-a-decade plan, Softbank is very much here for you.)
I understand your point about Microsoft/Google, and I'm saying it's entirely orthogonal to the discussion of anti-trust enforcement.
I don't care, and the anti-trust authorities shouldn't care, whether breaking up a vertically-integrated trust improves or degrades the investment performance of that firm and its successors.
As for Amazon, I find it a bit hard to be as cavalier as you're being. Gutting business models that require product development is a bit different from making pasta on an off-label. And when it comes to Seattle, I was just pointing out the local labor market monopsony within a few fields. Consolidation in tech has repeatedly been found to exert a depressive effect on wages in employment hubs like Seattle and the Bay Area.
I also see exactly no downsides to forcibly separating Amazon into "Amazon, the e-commerce platform and logistics outfit", "LifeBasics, the online retailer", and "AWS, the cloud service provider".
LifeBasics might or might not die within a few years without the reams of inside information that makes it possible, but the other two will be just fine.
I think you will find that food and apparel makers are just as sensitive about their product development budgets as USB cable makers -- nonetheless they've all found a way to live with Costco, Target and Walmart's store brands. They'll live with Amazon Basics as well.
I'm seriously and non-snarkily gonna ask for a citation on the wage issues. I'm very aware that _collusion_ between major tech employers has sometimes depressed wages (and indeed happily cashed my settlement check from the google/apple case as a result), but claiming that amazon _on its own_ is depressing engineering wages in the PNW seems like a bit of a stretch to me: Google, Microsoft and Oracle are all happy to bid for your time as a seattle-based engineer, as are a raft of well-financed startups. I'm willing to learn differently (lord knows I've been surprised by counterintuitive facts before) but this is the tightest market for engineers that I've seen in my lifetime.
I’ll have to see if I can find the paper. Been a while, I’ll get back to you.
Is the SoftBank dumb money really going to new CPG ideas? (As opposed to wild moonshot other ideas)
I was using softbank more as a (notorious) example than as the be-all and end-all but yes, they have their fingers in a ridiculous number of pies right now including apparel and groceries. https://visionfund.com/portfolio. (And the Vision Fund isn't their only fund!)
"You can hobble competitors by denying them access to services, steal crucial business information off AWS..."
Do you have evidence that they actually do this? Hypotheticals are not a good reason for anti-trust enforcement. E.g. AWS seems to be perfectly happy to host Netflix even tho Prime video is a direct competitor. You think Amazon is stupid enough to break contracts with other companies by stealing crucial business information off AWS? What a bunch of absurd fear mongering.
One assumes you’ll happily use Alicloud, then.
Amazon already skims sales data from vendors on their platform and uses it to develop competitor products; it’s a good and a skip to “take a peek at something stored on our servers.”
Hypotheticals are a perfectly valid reason for anti-trust enforcement. Why allow the possibility to exist?
There are plenty of ongoing anti-competitive behaviors going on as is, so it’s not really necessary to resort to hypotheticals, but whatever.
"Amazon already skims sales data from vendors on their platform and uses it to develop competitor products"
Why do people bring this up as something nefarious? Every major retailer does this! I was just at a Kroger supermarket. They have multiple house brands. How does Kroger decide which products to make house-brand knockoffs of? They look at their sales data! Heaven forbid a company look at its sales data!
"it’s a good and a skip to “take a peek at something stored on our servers.”"
I have no idea what "a good and a skip" is, but the two things are very different. Amazon has contracts with the companies they sell cloud services to that says Amazon won't access the companies' data, and Amazon would lose huge amounts business if it turned out they were spying on cloud customers. Amazon does *not* have a contract with vendors saying they won't look at their sales data. Again, it's standard practice to look at sales data! Heck, I wouldn't be surprised if it's stated in the contract that Amazon *will* look at its sales data.
"Hypotheticals are a perfectly valid reason for anti-trust enforcement. Why allow the possibility to exist?"
Doesn't the possibility exist for *any* cloud computing provider to steal data from their customers? Why is this Amazon specific? Is it ok for the second largest cloud computing provider (Microsoft) to steal customer data but not Amazon? Should we then take preemptive anti-trust action against all cloud computing providers because they could hypothetically steal data?
If you can't show harm, then you don't get to take anti-trust action. If you think amazon is causing harm, then provide evidence rather than just spouting off about hypotheticals. If spouting off about hypotheticals is sufficient cause for legal penalties, then I submit that you could hypothetically be a child molester and we should lock you up just to be on the safe side.
I don’t know if this is because I got lucky in the grocery stores I’ve frequented since self-checkout became a thing, but in my opinion self-checkout is much, MUCH preferable to full-service checkout. Partly because I don’t need a human being standing there judging the vegetable:dessert ratio of my grocery purchases, but back in LA Ralphs used to send me coupons regularly and to use those I’d need to flag down an attendant anyway. The real reason self-checkout is better is because I know how to pack groceries into my backpack and grocery bag much, much more efficiently than any stranger, no matter how skilled. Also, no risk of the next person’s groceries rolling down the conveyor belt into mine while I’m still packing up.
If you're filling one backpack and one grocery bag, then I agree with you on self-checkout. If you're trying to check out an entire trolley (cart) and fill up a dozen or more bags and then load that all into your car, then that's inconvenient.
I've found it depends on the place. Places where people tend to work at the grocery store for a long time will often have really experienced and efficient staff, that can process goods faster than I ever could.
Places with higher turnover can be more hit and miss.
In Sweden, if you have a loyalty card, you can get a scanner at the supermarket, self-scan everything you buy and self-checkout. No cameras needed. There is a gate you have to scan your receipt to get out and there are random checks. All very convenient. Around 80% of shoppers use this system where I shop. I think this system exist in most of Europe too. The Amazon Fresh solution seems convoluted and kludgy to me, and creepy too.
The Big thing is interesting, since you can make a case that Amazon is getting to a level where you know so much about the consumer that price movements is no longer the most efficient indicator for allocating resources, hence planning potentially superseding market allocation.
Self-checkout is ubiquitous in the US but infamously despised
I think Dan Karreman means the hand held scanners you carry with you (not the semi-supervised self checkout scanners) which I think people like, no? We've had these some 2014 or so at least.
Yep, that’s correct. I like self-checkouts too, but can see why you find them annoying. The scanner system basically eliminates lines, and fresh produce is a breeze since you put it on a scale, point at the image of what you are weighting , get a bar code, scan it and move on. Sure, it means you have to do some work but arguably not more than what is involved in getting the stuff from somebody (instructing, inspecting, and receiving the stuff).
Oh, it could be. In that case yeah, I'm not familiar with that approach, but I can see it being convenient.
No worry, here's a link to what think he meant. (Giants in DC area have had forever now)
"Hand-held scanners cut check-out time at Giant | Business | dailyitem.com" https://www.dailyitem.com/business/hand-held-scanners-cut-check-out-time-at-giant/article_2bece2db-9862-5567-b024-4fbfd8e23e9d.amp.html
I think it’s despised because the stores put so much energy into preventing shoplifting that the system is simply not very fast
That's interesting. I feel like I'm not really slowed down by anti-shoplifting measures, just that it's not very efficient when I need to unload my own cart, scan everything myself, bag it myself, and put it all back. What really slows it down though is that if I have any beer or wine or anything, I have to wait around for someone to come check my ID, which is a real pain. I do like it for when I just need to check out a small number of items though.
Most of the big supermarket chains in the US (and at this point even a fair number of the smaller regional ones) are already most of the way into transitioning to a similar self-checkout system. It’s generally pretty convenient but with one big caveat: nobody has really cracked the fresh produce issue yet. If there’s a real possible upside to the Amazon approach it’s that maybe someone will figure out how to get a camera to reliably discern cucumbers from zucchini.
I think that most modern European grocery stores (because I've been to a numbers there) have scales and attached bar code printers right in the produce section where you weight your produce in the usual plastic bag, type in plu code right there, it prints a bar code sticker you put on bag, and you can hand-held scanner that, or later scan it at either a manned or unmanned checkout lane. I have always liked this system.
This does not seem like a major advance over having a cashier do all this work for you.
If you have a hand scanner, it means you don't really have to wait in line (I think my store even has something where you can just link your card and don't have to check out at all). It also saves you the time and effort of loading everything onto the belt and you can just bag as you go.
It seems to me that a lot of cool new technologies require more work, not less, from the consumer.
Those exist in Whole Foods! It's used by the Amazon shoppers (who were getting lazy and not weighing produce) but it seems like such an easy thing to allow customers to use as well. I think it's the store being a bit over paranoid about the shoplifting potential
This is kinda how it works in China, too, but there's normally a store person to help you. Gotta say I'm rather "meh" about this system, as there's often a queue (And Chinese people are ruthless about cutting in line). Which translates into delays.
This describes my local Stop & Shops. The hand scanners are great, esp. combined with reusable bags, since you can scan and bag as you shop.
I normally don't mind having to type in the PLU codes. What I find annoying is the times produce has a bar code that scans by accident, incorrectly, and then I have to wait for an employee to come over. Or the times the scale guesses wrong about how much something should weigh. I do wonder if there would actually be a significant increase in losses if they just turned off the bagging area scales at self-checkout. I'd move much faster if I could bag directly in my cart.
Cucumbers from zucchini is hard, but I think it'll be even longer before it can tell whether I bought gala, fuji, or honeycrisp apples, unless they're bagged with distinctive packaging, which I don't want.
Well, the Amazon Fresh model certainly understands what bins certain varieties of produce are supposed to be in. They don't necessarily need to be able to distinguish apple varieties visually if they can tell that you took one apple from the honeycrisp bin.
That said, I'm sure the devil is in the details.
(Also, I'd be surprised if Amazon wasn't also planning to use your previous behavior to build a probabilistic model of what you purchased. If they're like, "Oh shit, our computer vision is having difficulty telling if you grabbed a fiji or a honeycrisp," they say, "well, the last five times he bought an apple, it was a honeycrisp.")
Yeah. For that I think they’re going to have to break down and put RFID tags in the stickers, and with the caveat that I have no idea what the market is currently like for such things I suspect that it’s going to be a prohibitively margin-eating cost for a while.
I just bought running shorts at Target using self-checkout. No problem! Except that I didn't notice they had RFID tags on them until I got home (didn't trip the door alarm, of course). So I would have to drive back to the store to ask them to remove the tags and prove that I wasn't trying to cheat them.
Or I could pound on the tags for five minutes with a hammer until they finally sprang open.
Yup, there are advantages to having human cashiers.
That has been the main argument against them for at least the past ~9 years since I first looked into it, not sure how much closer we've gotten to feasibility.
This whole piece boils down to "big companies have economies of scale and that's efficient, and sometimes consumers benefit from this!"
Everyone knows this already, including the economists that want to fight Amazon with anti-trust efforts. The problem is Matt misses half the picture, which is that there are trade-offs: fundamentally, between efficiency and resiliency, and more specifically, trusting Amazon to keep the consumer happy once they have achieved complete dominance in market. I don't see any engagement with this, just "big efficient, efficient good!"
The pro-Amazon take is it will never be possible for them to have complete dominance. Target, Wal-Mart and others are still out there and it doesn't seem plausible they can run them all out of business.
Every time I leave the US I'm mesmerized by, as I walk through a typical neighborhood, the menagerie of small grocers, bakeries, and butchers. It's deeply sad to me that as Americans we only get to drive a few miles and then choose between two Everything Stores, and because there's not literally only one option we consider this acceptable competition.
I agree with the feeling, but Matt has convinced me this is pastoralism. We want customers to get good prices and we want workers paid well. Those things usually come from chains, because they have better margins.
I feel that I can't oppose residential NIMBYism, but support commercial NIMBYism.
I'm always amazed how those small grocery stores in Manhattan survive but then I think about the density and what the volume of customers per hour must be for them, and then it starts to make sense.
Maybe I'm being a bit backwards- looking but a typical neighborhood in even a small town had a dozen food retailers 75 years ago.
“…once they have achieved complete dominance in market.”
Yes, that’s the fairy tale that drives left-wing policy.
*painfully Yuppie brain* can we get this instant checkout technology at Trader Joe’s?
I'm very much looking forward to the day when we can avoid all those interactions with "humans" most definitely including those friendly ones Trader Joe's hires.
Who’s gonna ring the bell? You?
I love Trader Joe's and it's the best check-out experience I have (except maybe Whole Foods, but I am rarely at Whole Foods.) It's delivery that would dramatically improve things -- which I guess is not a technology issue.
WF has indifferent checkout people, not the charmers that TJ's has at least in Washington DC
A couple of random thoughts -
1. Matt Stoller is one of the absolutely dumbest people on the twitter.com. When he's not fellating Josh Hawley because he says "Amazon bad", he spends the rest of his time being extremely wrong about everything and it's mystifying why anyone pays any attention to him.
2. Years ago, Food Lion, a large chain of middle of the market grocery stores, decided to re-brand their stores based on the income levels of the communities in which they were located. higher income area stores became Bloom, middle income stores staid Food Lion, and lower income communities got Bottom Dollar Food.
The concepts failed after a few years, but Bloom tried to do what Amazon is doing now nearly 15 years ago. You could grab a self-scanner when you entered the store, then walk around and scan and bag your own stuff as you shopped, and at the end you simply walked up to a self-serve kiosk to return your scanner and pay. It was a great concept and I was disappointed when it disappeared along with the Bloom branding. That amazon has taken this a step further and can autobill you as you walk of the store is a no-brainer of a concept.
3. I find self-checkout awful. Unless I've got less than 5 items, I avoid self-checkout at all costs. Its slow, the area is too small to effectively organize your groceries, and it inevitably yells at you for not putting the items down quickly enough or too quickly into your bag and needs a manager to unlock the terminal.
4. I try to support unionized labor anytime I can, but here in DC I find the unionized grocery stores to be generally terrible, with surly, unhelpful employees and far more spoiled produce. I'm glad there are unionized stores in town because it likely inflates the wages at the non-unionized Harris Teeter I shop at, but every Safeway in DC is a disaster, and Giant is only marginally better.
Bloom at the top and Bottom Dollar at the bottom! That’s some George Saunders shit.
The Giant in my (DC) area offers scanners and an app that scans but there always seems to be a network problem that keeps them from working.
Just got back from the Giant. They just finished a renovation and the scanners and app work now
Any idea if the DC Costcos are unionized? I recall one having the worst and slowest service of any Costco I've been to.
Our local mega-chain, GIant, does have those gun scanners which my partner loves but I hate because I prefer a cashier at checkout rather than force me to remember what variety of tomatoes or rolls I picked and what their cost code was.
The history of grocery shopping is one of pushing more and more labor onto the customer. 19th century dry goods stores essentially had everything behind the counter. A&P's innovation was to allow customers to select their items from shelves themselves. Self-checkout enlists the customers to do their own scanning and bagging. This has always struck me as a poor form of specialization as the general public will never be as fast or efficient as checkout clerks who do it all day long. Some stores still have bag boys (who in many areas historically worked just for tips) but even this form of labor division is vanishing. Since the store pockets the saved labor from self-checkout, the inefficiency is a time tax on the customer.
I'm not sure how Fresh works in this paradigm as it is replacing check-out machines and scanners with just a different technology.
To put it mildly, my experience is diametrically opposed. Being a checkout clerk is boring work that requires staying on your feet for extended periods of time: it’s a job designed to produce boredom, disengagement and ennui, and as a customer unless I am buying a large amount of produce (where the clerk having memorized the product codes for the most popular items will mean that they can’t help but be faster than me) I will almost always aim for the self-checkout line and expect that I will save quite a bit of time by doing so.
Both this post and the replies to it are missing the mark.
The factor holding things back has always been theft losses.
General stores with a sole proprietor had that one person and a physical barrier between customers and products primarily to prevent theft.
A&P's innovation was realizing that with modern policing and low crime rates, they could set up a checkpoint between customers and the exit to permit customers to get on with it. Which everyone very clearly wanted; customers didn't view it as taking on labor, they viewed it as speeding up the process compared to going to the grocer, butcher, and produce market in turn and waiting to be served.
Self-checkout only became widespread once camera surveillance started to limit "self-discount", and it's become clear that people mostly want it as well, especially for small-to-medium purchases.
It might be faster to have clerks grab things from behind the counter when there's just one kind of everything and not that many choices in total. But when there are many varieties and brands and you need to browse to figure out what you want, not so much.
I think an Amazon warehouse/instacart model would probably be the most efficient for car centric grocery shopping. You go online, punch in what you want, drive to the store, and just load it into the back of your car and leave. The inside of the store/warehouse is a mess where only the computers know how things are laid out, but people and machines move things around according to whatever the computer tells them to do.
That said, this model works best for packaged goods. For produce, consumers want to be able to inspect the goods and decide on the best whatever. So maybe just have two stores, one for just produce and another for packaged goods.