This is exactly the problem, and it's a difficult one to solve. I've seen research suggesting that the cognitive cost of filling out payment forms and giving out credit card information makes it easier to pay $10 than $1, since it doesn't feel "worth" all that effort to pay only one dollar. Unfortunately this is part of what reinforces the market dominance of mega-retailers like Amazon, because if you already have an account with them, the cognitive cost of signing up has already been paid.

> cognitive cost of filling out payment forms, etc.

That's exactly the case for me. I spend physical cash with much less hesitation than for online purchases. I'd probably spend way more freely online if I didn't have to keep track of the purchase just in case something goes wrong, like getting mis-charged, double-billed, a continuing subscription I wasn't expecting, or just to remember why that $1 debit appears on my statement.

As an aside:

It's curious to see that this made it to the HN front page and already has more than 25 comments, yet nobody has commented on the original article's own page which does allow commenting. Well, there is one comment that says, "Great analysis", but that was me just to check if commenting actually works. I suppose that ties in with the same idea -- it's not worth the "cognitive cost" to make a comment no one will read.

> it's not worth the "cognitive cost" to make a comment

Their comment form requires login or signup. I'm already signed up and logged in here. That's reason enough for me to post here and not there. I would never create an account just to leave a comment unless it was someplace where I intended to visit and comment often.

Not only that, but by now I've figured out that the comment sections of most sites are toxic with little, if any, moderation. HN has a good moderation set-up, respectful members, etc. I try to avoid even reading the comments of other websites since they just end up being depressing.

BBC is especially depressing.

You should try Fox News.

So if true federated non-statist, non-corp entity login actually worked we would probably comment freely and purchase stuff in 0.10-0.50 increments? Mozilla should be a payments platform.

The browser vendors are the only actors who might be able to make this work.

Seeing how Google makes basically all its money from internet ads, I don't think they have any interest to enable other monetization schemes. (Unless, of course, something makes them believe that ads are going to become unattractive and they need to pivot.)

I like the idea of no registration by having the browser locally generate a public/private key pair for use on the site.

The mobile stores come closer to that less hesitation scenario I think. Your card info is already there, you only need to input your pass to spend the money.

Though PayPal doesn't feel like that, I guess being in the browser is off-putting in terms of spending because of added risk.

Apple Pay and Android Pay are one touch, authenticated by thumb print and IIRC uses a temporary credit card number.

That's not going to work given that it is still CC on the back end -- per-tx fees will eat you alive at the $1/tx level.

I'd guess it'd eat $0.50 of a $1.00 transaction, maybe a bit less if you're lucky. That's bad, but it's workable if you have low operating costs - it certainly isn't a loss. Most people seem to think $5.00 is when credit transactions start making more sense.

Amusingly, Amazon charges me $0.14 every single month for my usage of Amazon S3. I wonder if they've got some sort of sick deal where they don't pay per-tx.

>Amusingly, Amazon charges me $0.14 every single month for my usage of Amazon S3. I wonder if they've got some sort of sick deal where they don't pay per-tx.

Off-topic but I had this (except I have a non-US card, so the foreign transaction fee was larger than the S3 bill). I contacted Amazon and asked if I could prepay $50 and they added a small balance to my account for free that'll cover my tiny S3 usage for the next few years

Aren't foreign exchange fees calculated by %?

My bank charges a flat fee of £0.75 per transaction for foreign currency transactions - not a big deal for a normal transaction but pretty significant if you're paying a $0.03 bill every month!

Some credit card companies charge $3 or $5 for transactions made in other currencies, with exchange rate being separate.

Are those fees > $1? Depending on the rest of your model - huge volume of transactions for a free-to-reproduce product? - even a massive cut might still be workable.

I disagree, I think the keyword is $1, because at least for me as strictly a user (not a stakeholder or business owner) generating a meme is simply not worth that cost no matter how easy it is to pay it.

> Has anyone made any real progress in solving this problem?

Wouldn't be too hard to implement in an iOS app with a IAP for credits. All the other places I can think where this has been done and it is relatively streamlined are other walled gardens like Steam.

Don't know about refunds, but for the rest Apple Pay (available via web) is literally click button, scan finger (on your nearby iPhone if you're on desktop and not using a tbMBP), done. I guess you can use credit card chargebacks, but beyond that refunds are hard to solve without making abuse possible.

I use privacy.com to generate merchant or burner cards. Funds come out your bank a/c, and Chrome integration makes it seconds to generate a card and fill out the details in the web page. And the card is locked to the merchant if it's not a burner.

Would this work for you?

In Portugal ATM systems are like tiny Internet, so some websites generate a one time ID number that you can use for an ATM transfer.

Basically you get the ID, then either make use of online banking or go to the next ATM machine around the corner and make the transfer.

So we've gone from filling out a form to either switching to a new tab and logging in to online banking to make the transfer, or even worse, leaving the building to go to an ATM?

Depends how much paranoid one is about securiy.

Online banking and mobile phone app are still an option as well.

Also we don't charge for ATM operations, so they are available on almost every street, at least on major cities.

Pretty sure Stripe does this - webmaster has to enable refunds on the Stripe side though.

Bitcoin

um... bitcoin? heard like 3 years ago that bitcoin will handle all sorts of micro-transactions...

anyone know what happened?

I haven't used BTC in a long time but last I read transactions take a very long time to process because the blocksize is too small to contain all the transactions that took place. I read (not sure if this is correct) that transaction fees needed to be upwards of $20 to ensure the transaction was included in the next block.

Bitcoin transactions have something like a $5 average transaction fee. You can offer less but your transaction won't get processed as quickly. Bitcoin is already slow at processing transactions.

Disclaimer: I don't use bitcoin. The $5 number is outdated, but from the graphs I could find in a quick search it seemed to be rising consistently.

I never spend more than $0.10 on transaction fees for bitcoin. I mean, you could give $5 if you wanted to, but why would you? If I was sending a dollar I wouldn't pay more than $0.01 or less for the transaction fee. And for something so cheap the merchant shouldn't bother waiting on confirmations. Exploiting 0 confirmations isn't easy, and it's not worth the work to anyone for a dollar. Even with sub cent transaction fees they'll still settle faster than 3-5 business days.

Stop spreading FUD. I spent 10c sending a transaction for $20 a month ago.

Smallest transaction I've seen is 191 bytes; [1] says right now to get a transaction in the next block the minimum fee is 240 satoshis/byte; and that today the 251-260 satoshis/byte has been the most common fee.

That's $2.82 according to [2].

And that's with current levels of demand - if there was a transaction every time someone visited a website, demand would be much, much higher.

[1] https://bitcoinfees.earn.com/ [2] https://www.google.co.uk/search?q=(191*240%2F100000000)+BTC+...

Yes, you're right. I miscalculated the fee. I actually paid $1 and was not trying to get it in the next block.

That's why you would use an off-chain solution like lightning payment channels with sub-cent transaction fees. Transactions are faster than credit cards too.

> Has anyone made any real progress in solving this problem?

Bitcoin micro-transactions via lightning:

http://lightning.network

> real progress

Let's not pretend that consumer-facing cryptocurrency-backed anything is little more than something for speculators, libertarians, and nerds.

Don't bother, the HN crowd will downvote this hard, bitter truth. I had this conversation N times, my argument is roughly this: "There's no legit use of bitcoin. Main usages are a) pyramid-like hold-and-hype b) illegal money transfers out of China. No, ordinary people don't want to use it for international transfer because you need to go through hops to get Bitcoin on the sender end, even worse the receiver needs to do something to get money and at the end of the day you have no idea how much money the receiver actually gets. And let's not even discuss outsourcing the security from banks to customers."

The community here is much more reasonable than, say, reddit. Notice the comment hasn't been killed as you expected it would. People here understand pretty well the barrier between cryptocurrency and "regular people."

Want me to collect the links of my downvoted comments which said this exact same thing?

> Has anyone made any real progress in solving this problem?

The reason they haven't is that the law puts the burden of fraud on the payment processors. The dispute resolution overhead is the same for a $.10 transaction as a $100 transaction, but if that average cost is $.30 then you can't have $.10 transactions and the fee on $1 transactions is onerous.

Using some cheaper less effective dispute resolution for micro-transactions doesn't work because if it's possible to consistently scam it but only for a dime then people will just scam it for a dime in a for loop.

This is why the off-chain Bitcoin processors have lower transaction fees -- they don't have this problem. But then the problem is that the customer has to get some Bitcoin and if they have to buy it with a credit card we're back to square one.

BTC processors don't have that problem because they do absolutely no fraud detection. That's a pretty big no-go for most consumers.

If it was actually the consumers who were averse to the risk then we could get rid of the law and nothing would change.

The problem is the law is inefficient because the consumer is the lowest cost avoider. They can do things like using a good pass phrase and not reusing passwords and not holding large sums of money in uninsured accounts for unnecessarily long periods of time.

Payment processors can't stop customers with no skin in the game from making poor decisions, so they have to sum up the cost of the resulting fraud and make everyone pay it as transaction fees, and then we can't have micro-payments.