Hacker News Reader

AWS is far more vulnerable to online competition right now than Amazon the retailer is. There's several companies that have the available infrastructure to compete with Amazon in the cloud... Google and Microsoft are already out there, and Amazon is finding themselves cutting prices to compete.

Nobody really has Amazon's physical goods delivery infrastructure. You can look at Wal-Mart and Target, but they have infrastructure designed to solve a very different problem, getting large amounts of the same goods to certain physical stores. They rely on semi-truck drivers to deliver palettes of goods to stores. What they don't have, for example, is Amazon's very close relationships to UPS and USPS. Could they develop it? Maybe. At great expense. Would they be able to peel off more existing Amazon customers than they'd be cannibalizing their own current sales? Who knows.

I was not suggesting that Walmart compete in home delivery, I was thinking that they should ship things to their store (which as you mentioned they already have that system in place) where it would be boxed up and ready for pickup by customers (because everyone has a walmart fairly close to them, and that's a huge advantage). The cost to them would be minimal and the customers they would gain that have been using online shopping with shipping to their homes could be significant. Even for a small surcharge it would be worth it to me.

>The cost to them would be minimal and the customers they would gain that have been using online shopping with shipping to their homes could be significant.

This does not compute. They already have ship-to-store, and have had for quite a while. It hasn't made a significant impact on their online power. Customers don't want to drive to a Walmart to pickup their online order. That's an extra step for no reason.

Site-to-store does indeed solve a problem. If there is something I want at Wal-Mart, then I have to find it. Not only that, after finding it I need to stand on a long check-out line. With site-to-store, I get it on the web site, then just go to the back of the store and pick it up. The problem is that at Wal-Mart it can be a crap shoot as to whether the site-to-store place is properly staffed.

I actually use the Home Depot site-to-store equivalent much more than the Wal-Mart one [probably several times a month, vs. once or twice a year]. Home Depot gets it right. There are still occasional problems, and it would be better if they could fulfill the orders faster, but it still avoids having to search for stuff, which I consider to be of significant value.

See this comment about the site-to-store please: https://news.ycombinator.com/item?id=8028352

For driving it really depends on the city and the persons situation. Many people drive by a walmart on their commute, the ability to order something in the morning then pick it up on the way home would be convenient. For businesses someone could order something the night before and pick it up in the morning and use it at work that day, that would be very valuable.

Everyone has a Wal-Mart fairly close to them? I think more people have the ability to reliably receive UPS deliveries than have a Wal-Mart close to them. Who wants to drive into the suburbs to buy stuff?

Plus, Amazon lockers.

They already offer that.

http://www.walmart.com/cp/Site-to-Store/538452

This makes no sense.

Walmart Online to dominate Amazon? Walmart has been trying hard in the online space for ages! What indicates they aren't trying hard? I'm sure they're trying as hard as possible. A major problem for Walmart is engineering talent. Top engineers don't want to work for Walmart. The brand has negative value for an engineer. Innovation and the ability to leap frog will be hindered by this brand problem.

And as far as "pick out items online and have them boxed at the customer service desk". Yes, it does sound reasonable - in fact they've been doing this almost since day one (Site-to-store). It's obvious and it doesn't solve any real problem. I don't want to have to get in my car, deal with traffic, park in a huge parking lot, wade through crowds and Walmart, wait for a disinterested customer service rep to find my online order (let's hope they didn't forget anything), ring me up, then haul everything back to my car, drive home, etc.

I want to click a button for what I want, then open my door that same day and see my items sitting there waiting for me. That is human progress and innovation that Walmart isn't going to be able to match.

Site-to-Store actually isn't available for most items.

http://www.walmart.com/search/search-ng.do?search_query=ensu...

http://www.walmart.com/search/search-ng.do?search_query=cent...

If it is then I can't figure out how to use it. Those were just the two items I would like to purchase and typed in first.

Regarding the "click a button for what I want, then open my door that same day and see my items sitting there waiting for me", that's not true for many places. I live in a tech-focused capital city and shipping is at least 2 days for Prime but I've had it take much longer (yes, for items sold by Amazon). I regularly see Amazon do the "our shipping carrier handed your item to USPS and we are no longer responsible for how long it takes" routine and it is completely unacceptable.

Once they get the site-to-store thing right (in it's current state you are correct, it has flaws) then they should offer some kind of large promotion to encourage people to register and use their service one time. Right now I'm hesitant to register and start shopping with Walmart, I don't have a logical reason for this though.

Real innovation will be when I can order and pay online, then send out my autonomous vehicle to walmart, then one of their employees will put the items in my car and the car will drive home and text or email me when it arrives. That's probably 10 years away but it will happen.

As for the point of items there same day/next-day, that is the problem Amazon Fresh is trying to solve. I have it here locally and it is pretty much amazing.

For the last point - why send out your autonomous vehicle and deal with employees at all? That's so 2024 (as you said). I want an autonomous Amazon Fresh truck, or maybe even an Amazon Prime drone to deliver to me. No employees needed, no human inefficiency.

"For the last point - why send out your autonomous vehicle and deal with employees at all? That's so 2024 (as you said). I want an autonomous Amazon Fresh truck, or maybe even an Amazon Prime drone to deliver to me. No employees needed, no human inefficiency."

Several reasons. It would work with any store or even fast-food/restaurants, sending my car to pick up a burger would be fantastic. Human employees walking the packages to your car (which would have a special lane in front of the store) could be easily replaced by a robot, at the moment it's just cheaper to have a human do the task. For many people who need a low paying and low skill job, that would be a great opportunity for them, there wouldn't be any dealing with customers either.

How would the autonomous delivery trucks you are talking about actually deliver the packages? The very difficult part would be getting the packages from the vehicle to the person's door. It would not work if it required the customers to come outside to pick up the packages, people wouldn't be home, they might be busy at the moment, it would be difficult to ensure they don't take the wrong or someone elses packages. I've considered that before and I just can't figure out a way for that to work.

The drones are also impractical and could only be used for small items. You would basically need a landing zone in your yard, I just don't see that happening nor being part of the free shipping. In my opinion the drone announcement was just a research project (that was full of flaws that can't be overcome) or a marketing stunt.

> How would the autonomous delivery trucks you are talking about actually deliver the packages? The very difficult part would be getting the packages from the vehicle to the person's door. It would not work if it required the customers to come outside to pick up the packages, people wouldn't be home, they might be busy at the moment, it would be difficult to ensure they don't take the wrong or someone elses packages. I've considered that before and I just can't figure out a way for that to work.

Tightly specified mailboxes.

Imagine that if you installed a special "UP-EX" mail box, you could receive packages securely when not at home. The truck has an arm that places the item directly into the receptacle.

> How would the autonomous delivery trucks you are talking about actually deliver the packages?

For the first iteration, I'd have a human take the boxes from the truck to the door. The truck would have one of those laser designators like Kiva uses for amazon warehouses to help the human pick the box. The difference is that the human is a minimum wage drone instead of a unionized teamster, so the savings could be quite substantial.

Plus, since the truck is automated, it doesn't have to be static while the picker is out running with the package.

Good points, good discussion!

Most groceries are bough weekly , so why is 1-day important for you? is it only for the times around the week that you miss some stuff ?

I have experienced the problem, but I have no idea as to the reasoning or lack thereof behind it.

I love to order my groceries to be delivered, it's very convenient. But I generally fail to order them significantly far enough (24 hours!) in advance, and end up going to the supermarket anyway. I would love same-day delivery. Don't ask me the reason, because I don't know it!

I disagree with most of your points, but agree somewhat with the non-sold by Amazon nonsense, and the generally poor quality of 3rd party sellers. You have to be very watchful with this. Not using Prime isolates you from this pretty well, as you can easily filter by items available under Prime shipping terms, and 3rd party sellers that put their stuff in the Amazon warehouses to make them Prime eligible avoid the slow shipping problems and are generally not an issue. However, the third party sellers on there are giving them a bad reputation, because people are not distinguishing them from Amazon proper.

I seriously doubt there is any personal retaliation going on.

I don't think they should split the business, as one of the things that makes AWS good is that it has to keep Amazon running. The initial concept of AWS as something that Amazon can use to meet demand peaks, while selling excess capacity makes a lot of sense.

There's also products such as this: http://www.amazon.com/Schlage-FE599NX-CAM-619-ACC/dp/B0083GJ...

This has the option for a single product from Amazon, or a "2 pack" for a $100 surcharge from some random supplier. This is on MOST products from Amazon and a real pain in the ass to avoid. I hate that Amazon allows this crap which is basically spam. I would be willing to bet that the supplier simply orders 2 of these products from Amazon after the purchase, then ships them out to you and makes $100. It offers zero benefit to Amazon's customers, but it's a rampant issue.

There's also the companies with names such as "Prime inc." which are just trying to trick people.

They probably make the most money off of assistants or employees who are asked to purchase a product and don't really look at the prices or shipping.

Imagine this scenario: I send the link above to one of my family members saying this is the product they need for their doors. They see 2-pack and immediately click it and continue through the checkout process, because I had told them this was the best value product and I had done the research for them.

A query parameter such as (?only_show_amazon=true) would help the issue, or better yet would be a permanent setting, but Amazon doesn't see the problem with allowing these sellers or products. Maybe utilizing the Amazon API and creating a website to avoid this would be a solution.

AWS was never really something to sell excess capacity, thats a myth. And anyway now, it is vastly bigger than Amazon's compute needs. The stockmarket would like it to be sold off, but not sure it will be.

It would be terrifying if AWS was just Amazon's excess capacity, as then one might want to ask what would happen when Amazon suddenly needed the capacity... Apart from for spot instances, a lot of people have never seemed to get that Amazon needs to over-provision massively because of how the product works, and that their customers still need to cover the cost of all that over capacity.

No one said it was "just" Amazon's excess capacity. However, they do use it...so it is way of not paying for their excess capacity, just like everyone else's use is.

Amazon needs extra capacity at peak times, are you saying they don't use AWS for that?

I have had excellent experience with Amazon. In reading the reviews of a couple of high end monitors that I am considering, I see that Amazon will give a refund even when the seller will not. I imagine Amazon hammers them back into line.

My only complaint is when they ship by FedEx. In my town FedEx hands packages off to the Post Office, where two day delivery becomes three or four.

In my (limited) experience, Amazon boots third-party sellers off pretty fast if they get complaints.

If anything, AWS is overrated. It already has some decent competitors and more will grow up soon, though they're massively underutilized. The idea that Amazon itself is critical to internet infrastructure is very disturbing, yet in many cases, it's true these days. We really need to get other cloud providers in the mix.

Amazon as a retail store has been consistently great for me in all major aspects. Any issues I've had with orders have been promptly resolved and Amazon has gone to bat for me several times. I agree that Wal-mart or another major department store could give Amazon a run for its money if they played their cards right, but like you, I don't think they're in that headspace.

SES is really the only AWS service that has not been done by anyone else as far as I know. Some of the SES features (spam reporting for instance) requires cooperation between email service providers and Amazon, something that many other companies would have difficulty arranging.

Yeah, the issue isn't really that there are no similar services available, it's just that people don't really use those services very much. There are several search engines competing against Google too, but if Google blackballs your site and you're dependent on organic search traffic, you're still going to lose 90% of your visitors overnight.

Internet infrastructure is pretty important, so it's a little more disconcerting to me if some huge percentage of sites are dependent on the same single ISP/datacenter/etc.

SES is really the only AWS service that has not been done by anyone else as far as I know.

sendgrid is like 5 years old: http://sendgrid.com/

mailchimp's version: http://mandrill.com/

rackspace: http://www.mailgun.com/

also: https://postmarkapp.com/

I had never explored Mandrill, thanks.

SES is still the cheapest (especially when scaling up) and extremely fast from my experience with it.

Here is a good article I found about SES and competitors such as Mandrill: http://www.ventureharbour.com/transactional-email-service-be...

The difference betwen SES and Mandrill? With Mandrill you'll be patronizing a great company with some fantastic folks.

(First used MailChimp in 2003 and they've always been the bomb.)

Before 2012 (when they launched Mandrill) MailChimp used Amazon SES. http://us1.campaign-archive1.com/?u=f7b9ee22124ff6454424dc10...

The Mandrill service is nice, I agree.

Also, MailChimp has a very cool 404 page: http://kb.mailchimp.com/article/transactional-reports-overvi...

The right way to implement "ship to store" is using something like "amazon locker". Assuming that's the correct route, amazon will win, because it can choose better locations than a walmart store . And in retail location is everything.

But the fact that amazon has tried some lockers but it doesn't seem the common way to buy stuff, and there's no huge scaling effort from amazon, seems to indicate that people prefer to recieve packages at home.

And bezos understands that in this game, if you want to make profits, you need for customers to love using your services.Having that ability at low enough costs could earn you a monopoly. And than might be the time to play with "amazon lockers" and other cost cutting measures.

The Amazon Lockers seem to only benefit Amazon because the shipping is much cheaper for them. If they offered a discount or credit for using the lockers then it could potentially be a useful service.

What is the benefit of using an Amazon Locker rather than shipping to your home? Maybe in neighborhoods that have a high stealing packages percentage.

The term lockers also imply (incorrectly) that they are supposed to be purchased by people for packages to be shipped to (like a PO box), the name might be hurting their success.

I can't find anywhere on any of the Amazon Locker pages where it lists a single benefit of using the lockers, http://www.amazon.com/b/?node=6442600011&ref=locker_fd_brws

I think that shipping fees for amazon are much lower than prices offered for people at say ups. Offering half of that, wouldn't be attractive and will probably hurt the amazon brand.

They already offer (or use to) a credit if you choose standard shipping while being a prime member. The Subscribe & Save does something similar.

How much credit do they give ? And what's the difference between services ?

It was $1 when I saw it, the link below says it's now $3. It's a credit for Amazon movies/music though, so it doesn't actually cost Amazon $3.

http://www.businessinsider.com/amazon-no-rush-delivery-2012-...

From my experience the difference is that they delay shipping your item. They claim to have one of the best shipping and handling systems in the world but for some reason they save money by delaying boxing an order for a week, it doesn't really make sense to me. I think they're really just trying to get people to purchase Prime.

I'm sure that there's an aspect of what you say at work, but I can also imagine that they have a certain amount of "pick and pack" capacity at their fulfillment centers.

By choosing 2-day, you're forcing Amazon to pick and pack that order today probably, tomorrow at the very latest.

By choosing "no rush", you allow them flexibility to fulfill from a wide variety of centers, on a wide variety of days. Maybe there's a weekly cycle in order volume and no rush orders may allow them to even that out with a more constant labor force.

In the UK they also offers "Collect Plus" - a delivery option where they ship to a participating nearby corner store and you pick it up from the tills. In my case it's a 3 minute walk to the nearest Collect Plus location, and they're "everywhere" near us, or we have 15 minute bus ride to the nearest locker..

Collect plus sounds really good. How much does it cost ?

I'm with you with the splitting, I don't think these to fields belong together and they can work much more independently with a split. But they probably will only split if one half goes down :\

But I like the Amazon store quite much, don't know any other company that offers a similar service (in Europe/Germany).

Should there ever appear a seemingly real fault line in Amazon's long game, I think at some point they'd be forced to sell the infrastructure / hosting platform to keep investors happy.