Hacker News new | past | comments | ask | show | jobs | submit login
Amazon will pull the plug on dash buttons (marketwatch.com)
95 points by blumomo 78 days ago | hide | past | web | favorite | 107 comments

I never really got the point of them since they were usually for things that stored really well and get used up at a relatively constant rate so it's usually not a surprise when you run out, so like why would I need to rush and blindly push the button to buy more detergent/toothpaste of a certain brand without knowing the price when I'd just eventually buy it in my normal shopping run/order.

The goal was to reduce friction in the purchase. You had to walk to a computer, bring up Amazon, look up the product, click the version of it you want, and push "one click order" (or maybe two or three more clicks) -- that's a lot of opportunity for you to just add it to your shopping list and just buy it at the grocery store.

Or from your point of view, that's a lot of effort for a product that you already know you want. If you're busy doing laundry, you want to keep doing that, rather than stop and order detergent, or risk forgetting it in the next five minutes (and then not having it the next time). The price is somewhat variable, but the variation is likely below your threshold of caring.

These aren't huge costs, but it's very much web-UX thinking: people have very short memories (and attention spans) and even a single extra click is an opportunity for people to change their minds or get distracted. If your washing machine came with an infinite supply of detergent, or your cat food canister were infinitely deep, your life would be a tiny bit better. If they knew the cost were comparable to other sources, many people would take that deal without worrying too much about whether the exact price.

That's Amazon's schtick. They make your life a tiny bit better and take a tiny profit on that. It works only if they can get that tiny profit on basically everything you buy. Everything everyone buys.

They've since discovered other ways. As you say, you know the rate at which these things get used up, so it's even better for them to just replace the button with a recurring order.

One problem with their approach IMO is there was no direct feedback on when you would actually receive the product both before AND after you press the button.

You don't know if the toilet paper is going to arrive the same day, next day, 2 days, or 2 business days. Many times if you know beforehand it's going to take 2 days to get toilet paper you'd just run to the grocery store instead.

Not just toilet paper -- even with shampoo, body wash, kitchen soap, laundry detergent, kitchen sponges, I'm usually unaware I'm running out until it actually runs out, and with most of these things I can't wait 2 days. I also live 2 blocks from a 24-hour Safeway so it's not actually an issue.

These things all keep well, so I just keep extras and order when I have to start using my last one, so it doesn’t really matter how long shipping takes. For me, the worst part of dashes was that Amazon didn’t keep the exact same thing in stock for the exact same price, and sometimes it’d be a very different price if it had to fall back to a 3rd party seller. So I usually ended up checking and then ordering on the computer anyway.

Spot on. It just needed a cheap LCD display with the basic info. They are remote terminals and need to provide standalone feedback for the experience.

Not only that, but you also had a limited selection of brands.

How much friction is there really in adding “toilet paper” to your shopping list and grabbing it during the next grocery run? It really feels like a solution in search of a problem.

Not much, but a little. You don't find yourself thinking, "I'm running out of shampoo, surely I will remember it for the five minutes it will take to finish my shower, then put it on my list." I fail at that all the time.

It's the firstest of first-world problems, and I sure as heck don't see putting a "shampoo" button in the shower. But it kinda doesn't surprise me that at least some people wanted them.

a lot, if you're e.g. depressed

though to be fair, I am (was?) too depressed to even set up the dash buttons

I suffer from depression as well, but I don’t think dash buttons would do nearly enough to help me (not being able to do basic tasks is a symptom, and no amount of automation will make the depression go away). And at least for me, getting out of the house helps. It forces me to shower and put on pants, otherwise I run out of toilet paper and detergent.

> You had to walk to a computer,

Wouldn't a smartphone app that has a repeat-purchase item dashboard allowing 1-click orders for them cover most of those "issues"?

> Wouldn't a smartphone app that has a repeat-purchase item dashboard allowing 1-click orders for them cover most of those "issues"?

That's pretty much (not the exact feature set, but close enough) what Amazon cited in cancelling it, so...yes?

Amazon doesn't seem to have that feature in their app, though. The closest thing is a buy again section, but it doesn't hit the mark (to me):

(1) Many of the items aren't consumable items. On my page, there is a wall clock, a power strip, and an instant pot.

(2) The ordering of the items appears to be based mostly on what I bought recently rather than what I'm likely to buy over and over again consistently. (Or likely to need right now.) It could allow me to choose, but it doesn't. It could choose for me intelligently, but it isn't doing that because #2 on the list for me is a year's supply of allergy pills that I bought 3 months ago.

(3) There isn't a streamlined purchase process. The buttons just add them to my cart. (Though this could have something to do with my not having enabled one click ordering on my account.)

My Amazon app has a section of "virtual dash buttons" down the bottom with only consumable items and they all purchase directly with one tap

Huh, so it does exist! I had to scroll through an almost endless list of tiles, but I was able to find it. (I tried a second time to count how many, and it wasn't there that time.)

Turns out it's also possible to intentionally navigate to it: app menu (top left "hamburger" icon) -> "See All Programs" (under "Programs and Features" group in menu) -> "Your Dash Buttons". And the UI looks just like dash buttons. With buttons that say "Buy now" rather than adding to cart.

It'd be great to have the ability to somehow pin it in a more prominent / convenient location if I were seriously planning to use it.

I’m half surprised Amazon doesn’t have a widget or something that displays the virtual dash buttons

I say half surprised. They’re probably avoiding Apple’s 30% cut or something

FWIW, Apple's 30% cut doesn't apply. You can't use Apple's IAP system for physical goods.

Same story on Android. Google Play's in-app billing is for digital goods only. (See the over here: https://developer.android.com/google/play/billing/billing_ov...)

You can use Google Pay (note no "l" in second word) for physical goods, although I think you can collect your own payments if you prefer.

Unrelated: Your speech-to-text result threw me for a loop, "moose Leon" (mostly)

Related: A hardware "Dash" button made more sense in a time when mobile smart devices were not ubiquitous. The recommended items I see are really strange, often items that are "one-offs" which I might only buy once every ten years are recommended. This to me is a serious problem with their algorithm.

> Related: A hardware "Dash" button made more sense in a time when mobile smart devices were not ubiquitous

Dash buttons were introduced in 2015, which was hardly the pre-ubiquitous-smart-device dark ages.

Yeah, I just noticed that at the same time. Fixed "moose Leon" to "mostly on".

They were fun. I bought a few back when they were $1.

It’s more fun to press the red bull button than to say “computer, order more Red Bull.”

I didn’t really use them that much because amazon’s prices got a lot worse so I ordered my Red Bull and batteries from other places.

I hope I can repurpose them for other iot signals.

> I hope I can repurpose them for other iot signals.

I bought one for that - it's authenticated on my wifi but doesn't know what to order on Amazon. Pushing the button will make it visible on the network but do nothing else. So I wrote some scripts to run when it associates on the network, works pretty well.

I think most buttons were actually used for other purposes (so for the hardware) instead of for Amazon purchases.

In some countries selling something without showing the price at the time of purchase is not entirely legal.

It didn’t actually buy. It sent me an email saying “you bought this, are you sure?”

So it was a funny situation where I had to get my phone out anyway.

Where was that? In the US at least you didn’t have to confirm.

In the US you had the option to cancel

in germany it was like this after a while.

could you share where you purchase them now?

It's not for your benefit, it's for Amazon's. It's for them to train people to lower both the cognitive and physical barriers to buying. And specifically, buying from them by default without exercising the thought process of evaluating other vendors, noticing price increases, etc.

I've started only using Amazon as a last option, not a first option, because I've noticed that their process aren't really any better and buying through them can make some things more difficult (warranties, refunds on reduced price items, etc).

That being said, it's very easy to just default to Amazon and not bother looking elsewhere.

Yeah, what I really want is a service that buys detergent/toothpaste in bulk when the price is below a certain given threshold.

Could you script something with CamelCamelCamel?

I used them and liked them. It's nice never having to remember to pick up detergent. I just push the button when it gets low.

Amazon was no longer allowed to sell them in Germany earlier this year. Consumer protection agency said the button isn't a clear enough buying agreement (here payment forms need to make clear it costs money to click, exact price, what you buy). Amazon allowed itself to change the product, e.g. amount of washing liquid.

In the US, I looked at them with a high degree of skepticism. How do I know that I'm getting the same product, at the best price?

In the US, quite frequently the same product is sold in different packaging, and the best price often varies between the packaging. Sometimes the 20-pack of toilet paper is cheapest. Sometimes the 30-pack is cheapest.

Besides, on the mobile app, it doesn't take that long to figure out the best deal.

I had a few, but stopped using them because the pricing was erratic.

Push a button for pistachios, and a pound bag shows up for $8.99.

Push the same button two weeks later, and get charged $17.99.

I liked the idea, though. It would have been good if it had been incorporated into the whole Whole Foods thing.

I've advised clients to be cautious about Amazon "subscriptions" for toner and such because I've seen crazy looking pricing that seems designed to take advantage of such automatic orders, such as a 3-pack of toner for 4-5x the price of individual ones. I'm sure that at some point it was less expensive....

A simple solution would have been an e-ink screen that listed the price per unit/press, but that probably would have increased button costs by an order of magnitude.

Also more users probably would become aware of price variance (as you have) and changed their shopping habits, which is not good for business.

The simpler solution, is to just not price-gouge customers and hope they don't notice, while trying to build a subscription-based business.

If Amazon isn't able to offer any level of price stability, then don't offer the subscription. Toilet paper is convenient to get on a subscription basis, but that convenience isn't worth paying $5/roll.

Definitely. It would be amazing if they offered subscription with price lock-in and a preset range lockout (i.e. if current price reaches 4x subscription lock price the subscription is suspended until price goes down).

They will lose some money but they'd also gain money from people afraid to drop subscriptions (e.g. gym subscription model).

So almost like a store or using your phone anyway.

Same problem here. We used to use it at my company to order paper towels when we were running low. It was super convenient but the pricing varied so much - the same paper towel bundle could range from $22 to $37 depending on the day.

Those CueCat scanners were a great way to scan used textbooks when I was selling them in college. They just magically spit out a text string followed by a <CR> when scanning a barcode. They were going for dirt cheap on eBay for a number of years.

In college, on finals week I'd get $100-200 cash and go to the bookstore. I'd hear they were buying back certain popular textbooks for next to nothing. I'd offer the seller $10 cash right there on the spot. "Sell to them or sell to me, it's the same price. Here's a $10 bill." After getting a stack of those (or when profs were cleaning out their bookshelves), I'd lug them home, scan with my CueCat and sell them for a substantial markup on half.com (RIP). One semester I found a stack of books outside faculty offices labeled "free" - sold those for $500 that weekend.

For anyone interested in weird low cost embedded devices, I suggest googling for Amazon dash button teardowns. There's some interesting choices made in their components and spec.



There is also this thing, the Amazon wand:


I've no real love for the intended use of the dash buttons (re-orders of items) but do hope that some other company comes up with a similarly easily hackable wireless button for interactions.

As of now, Amazon appears to still be selling their "IOT button" (which is basically a programmable dash button)


The $15 "1st Generation AWS IoT Button" is out of stock, but the $20 "AWS IoT Enterprise Button" is still in stock. Is there a difference besides the label?

AWS IoT Enterprise Button creates a JIRA ticket and sends approval requests as well. /s

This is the first thing I’ve ever upvoted on Hacker News

The AWS IoT Button is for the AWS IoT Button service ("Cloud Programmable Dash Button"): https://aws.amazon.com/iotbutton/

The AWS IoT Enterprise Button is for the AWS IoT 1-Click service ("Trigger AWS Lambda functions from simple devices"): https://aws.amazon.com/iot-1-click/

Notice how the Enterprise button listing (https://www.amazon.com/dp/B075FPHHGG) says "This is not the AWS IoT Developer Button" with a link to the other button.

1) Double the lifetime clicks (2000 vs 1000) 2) More tightly integrated with the AWS IoT 1-Click service. (so arguably more restricted, but actually easier to manage and scale if you're an actual enterprise managing thousands of these in the field, rather than one developer building a toy app)

The first gen ones are able to speak MQTT as well as integrate with AWS IoT core, so they could also fire off lambdas, hit SQS, etc.

The new ones are strictly worse, only able to be bound to a Lambda

Usually the Enterprise version is twice as big, 100x the price and 1/4 as reliable.

Better support though.

Yes, but at $19.99 it's pretty expensive. I don't know what the right price point for something like this but, given the fact they are easily lost, definitely < $5, more like $1.

HW people on HN, is this impossible at scale?

$5: probably doable. An esp8266 dev board goes for $2 on AliExpress. That leaves $1 for plastics and $2 for a lipo battery.

If you wanted to make it for yourself. A general rule of thumb I've heard several times for makers is to retail at 4x the cost of goods sold - else your business is unsustainable. So for a 5$ retail price you'd have to be somewhere near $1.25 COGS. Going the otherway - 19.99 for a 5$ COGS is perfect.

That's zero margin for Amazon though and no budget for assembly. If the materials cost is $5 then $10 is the absolute lowest you could probably hope for. $15 is a more sustainable (profit generating) product.

If you're talking price point $5 would be $0 profit by that bill of items.

Was the original/intended purpose to always sell them at cost and make the profits up through increased sales?

Being tangentially involved in this project it was my understanding that the devices were financed in large part by the marketing departments of the brands that had buttons, which is why they were tied to brands.

"We lose a little money on every customer, but we make it up on volume." :-)

I think in Amazon's case it was just experimentation with a sunk cost. Based on the discussion above it looks like it's impossible to form a company to sell these to turn a profit, at least in the dash button form.

The original purpose of the “hackable” button was to give a fun piece of experimentation kit that worked with AWS and encouraged people to try it out. I set up my AWS account just to use one.

> If you're talking price point $5 would be $0 profit by that bill of items.

negative profit when you count assembly.

The actual dash buttons (including the IoT ones) don't use lipo batteries, they use a primary cell. So you can probably shave a bit off there.

That's interesting... But does Amazon provide a public API for making purchases?

like Puck-JS https://www.puck-js.com/ ?

That's an interesting project but not quite the same - the Dash buttons use wifi to connect (instead of bluetooth) which lets them operate more independently.

They're also roughly 2x the cost of the dash enterprise buttons. Granted they have more functionality

Not as cheap, but Fibaro has a ZWave button like this. Registers double clicks and some other behaviors too.

I guess you could make and program your own with an ESP8266.

I think I've had 3 physical buttons since they were introduced. Only used one of them one time. It was a novel idea but never really caught on.

I'm sure Amazon would rather you just ask Alexa to order things for you now anyway.

These were the ":CueCat" of 2015.


I wonder what today's equivalent is?

Rather than pressing a button when you run out, I think (totally anecdotally of course) people just use subscriptions.

Send me 12 rolls of kitchen roll, every 3 months, a bag of cat food every month, etc..

Today's equivalent is scanning a QR code with your phone.

(following this tangent) My impression of QR's problem was:

1) High ergonomics barrier; the user needed to install a QR app from the appstore or find the icon for it and start it.

2) People who are not on HN don't know what it is anyway.

Now that it's built in to most Android/iPhone cameras, #1 is solved. Now that #1 is solved, #2 is self-solving.

QR gets a lot of jank but when I use it, it's very handy.

Probably more accurate an analogy than you think:


Not of cuecat literally; the failure of cuecat in 2019.

It was always baffling to make non-trivial custom hardware in an age of universal smartphone ownership.

Does Amazon even have any good apps? Seems like easy subscription management and re-ordering is something they could do there.

Amazon already has their Subscribe and Save program.

I actually really liked the dash buttons for those items that I did not use as regularly, but the number of items like that are pretty low to warrant having more than 2 or 3 of these things.

Then when I rebuilt my wireless network I just did not care to move them over.

Really liking them and abandoning them anyway is a pretty solid indictment of the idea.

I will say this is an issue with iot devices in general right now.

When I rebuilt my wireless network, every smarthome device has to be modified and that was the worst part about it.

At least these were basically free anyways (you bought them for $5 and then got a $5 credit on your first purchase with them) so I did not feel a need to switch them over.

The IoT market is a testing bed for custom hardware ideas. For every 20 dash buttons you eventually get a Nest. It just doesn't seem that way because of survivor bias.

I’m probably alone but I’m pretty bummed about this. I regularly used them to order detergent and Clorox wipes when we were getting low. Haven’t been fully out of either in a couple years.

How much work would it honestly be for them just to keep the existing ones going.

Not alone. We have half a dozen of these scattered around the house and use them often. We'll eventually get used to asking Alexa to re-order items but the buttons will be missed. Much less friction than having to ask Alexa and going through the verbal confirmation back and forth. No clue as to why they would discontinue existing buttons. Very odd.

"Alexa, add X to my shopping list"

Well you can directly buy the same products with Alexa, which is what I'll probably do.

I don't understand why you'd need it. It's not like detergent is something you need to order every week. It takes me over a year to use a large jug of laundry detergent.

All you have to do is get two containers, one open and one backup. When you use up the first buy another one the next time you go grocery shopping.

> It's not like detergent is something you need to order every week

That's exactly why this is useful. It's something I forget to buy at the store because I run out so rarely. Right now while I'm doing my laundry if I notice I'm below about a third I'll hit the button.

Before I'd almost always realize I was out of detergent mid-washing, undoubtedly the night before some important meeting, and have to run to the store in the middle of the night.

Maybe, but the advice to have two containers of stuff like that seems like a good idea. The only issue I have with that is that I misplace the backup and then I buy another and accidentally have 3.

Does that really justify eventual electronic waste, an inefficient delivery and packaging material?

Leave the empty bottle by the doorway or in your car or something until you remember.

I wonder if anyone has gotten the Wi-Fi pairing process to work outside of the Amazon app. Otherwise, you can't change your wireless password without bricking any buttons you might be misusing for other DIY uses after Amazon kills support.

Relegated to the parts bin. Usefulness perhaps in a second life... https://hackaday.com/tag/amazon-dash-button/

Biggest problem I had with them is the products you could order were hard coded to the device itself. If the product you want to order doesn't have its own dash button, the whole system was useless.

Surely they will announce a Dash button recycling program. Right?

Edit: Read rest of article and have my answer. Sorry! Don’t spank me.

it was announced in the article!

They are ahead of their time.

If I pressed one and <1h delivery would happen it would be great.

But we aren't there yet.

Does anyone know of any programmable alternatives that are not tight to AWS?

Makes sense with the ubiquity of Alexa/Echos in homes. Why quantize consumer selection to a finite number of items when one can be able to order anything with the same seamlessness.

I don't understand the appeal behind ordering something without knowing the price.

The only thing I ever buy without knowing the cost is a drink in a pub, and that's some sort of strange anachronism.

And in that situation, if the bartender turns around and tells me 25 quid or whatever after pouring it, I'd just walk off.

The economy doesn't work properly if you just blindly pay whatever.

A lot of the products seem to be small household items, like razor blades, trash bags, water filters, and the like, that when you run out you are (1) probably going to want to replace with the same brand, (2) probably don't want to put off replacing to wait for a bargain, and (3) probably will still buy even if the price is a little more than the last time, but within the normal Amazon price variation.

Also, according to their documentation, pressing the button queues the order and sends you an email, which includes price information. You can cancel the order before it ships. I'd expect that the normal flow for most people is to press the button when they notice they are running out of something, and then check the order details when they next deal with email.

This seems incredibly non-optimal on a few levels.

I rarely buy small household items at full price, just wait until my brands go on sale or I have a coupon to buy several. They go on sale often enough that I almost always have items in the house. I probably have a five year supply of razors at my house, they take up basically no room. If I need something right away I'll put it on the grocery list and pick it up next time I go to the grocery store.

On top of that I don't really believe in shipping my household products to me one by one, seems very wasteful and inefficient. Plus you'd either need to pay extra for shipping every item you need or buy a prime subscription on top of that.

Amazon pricing is all over the place, not like the normal variation of other stores. You could easily be paying double or even triple since the last time you bought.

During setup of the dash button you tell amazon what you want the button to purchase. When setting up the button you get a few choices to choose and you pick one to link, so whenever the button is pressed it will order that exact product. When you pick the product you’re also told the price.

The price changes.

>The price changes.

And in some cases you won't like the change.

I have noticed on other "platforms" (I don't use Amazon, so it may be different on it) that vendors must be using some (evidently crappy) software to "re-align" their prices automatically to market/competitors/whatever and this makes from time to time some egregious mistake.

Recent anecdata, a couple weeks ago I bookmarked a plastic bench that I might have been interested in buying but had no "urgent" need for, I found a vendor selling it for 148.50 Euro (there were a number of other ones selling the same article ranging from 150 to 170 Euro).

Last week I accessed the bookmark and found that the price was now 1,148.50 Euro (and I just checked, it is still at that).

But I have seen it happen several times.

Not that anyone would have a dash button for that, but I suspect that the same can happen for common, ordinary house supplies.

They seemed to have little value.

Cool project though.

I've added up all the shits that are given, and the stunning total is... zero.

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