What something is worth and what it costs are two different things. The big correlation is that if something costs more to produce than it's worth to the customer, nobody is going to make it. But if it costs less to produce than it's worth, who gets the surplus? In a competitive market, it's mostly the consumer rather than the supplier, because customers pick the supplier with the best price.

What ad-supported services did is zero out the price of anything that costs less to provide than the amount of ad revenue it generates. But the amount of ad revenue companies get per-user is already pretty small and companies are demonstrably willing to provide the existing services for that amount of money, so we know the upper bound and it's not that high.

▲

pdonis 17 hours ago | parent [-]

> What something is worth and what it costs are two different things.

Yes, surplus is a thing, I agree. But that doesn't materially change what I said. The thing still has to be worth at least as much as it costs for users to be willing to pay for it, so what users will pay at least sets a lower bound on what the thing is worth to users. (Note that it can be worth different amounts to different users; the more precise way of stating it would be that in a competitive market, price equals marginal cost equals marginal value, i.e., value to the marginal user, the user who just breaks even paying that price for it.)

> What ad-supported services did is zero out the price of anything that costs less to provide than the amount of ad revenue it generates.

Which also uncouples the price from any measure of value to the user. The price is now measuring the marginal value to the ad purchasers. The value to the users can be anything greater than zero--the fact that they're using it at all means (or should mean, if the users are rational) that the value to them is positive. But it could still be less than the cost to produce. And the worse the user experience gets, the more likely it is that the value to users is less than the cost to produce, even if that cost is small.

Plus, there's a whole other piece of this that the analysis we've just done doesn't even capture: externalities. One simple way of stating what many people think is wrong with the ad-supported business model is that it creates large negative externalities that, on net, mean that the value to users is negative--but the users don't see the externalities so they don't realize this, and the tech companies have offloaded the costs of the externalities onto others, so they don't see them either.

▲

AnthonyMouse 16 hours ago | parent [-]

> Note that it can be worth different amounts to different users; the more precise way of stating it would be that in a competitive market, price equals marginal cost equals marginal value, i.e., value to the marginal user, the user who just breaks even paying that price for it.

In a competitive market the price only depends on the value to users in the sense that it's required to be lower than that to make any sales. If something costs $1 to produce in a highly competitive market then the price is either going to be ~$1 completely regardless of how much more value people get from it than that, or no one will value it at even $1 and then no one will produce it. This is why farmers are always on the edge of bankruptcy even though their product is "without this you will die". Actual competitive market.

> Which also uncouples the price from any measure of value to the user.

It uncouples the lower bound. If the production cost is $1 but the user only values it at $0.50, now it still gets produced as long as the advertiser is willing to pay $1 to show the user the ads. But depending on how much you value not having ads, that could still be good. You got $0.50 worth of value without paying anything.

> And the worse the user experience gets, the more likely it is that the value to users is less than the cost to produce, even if that cost is small.

The real problem here is, they can do something you value at negative $10, but the advertiser will pay them an extra $0.05 to do it, and then they do it because "you're not the customer, you're the product".

In an idealized market this doesn't happen because then you just pay them the incremental $0.05 instead of the advertiser, but we've been screwed by the regulatory environment on both ends. For the seller it's hard to accept small amounts of money from arbitrary users without doing business with a fickle payment intermediary that wants to take a huge cut for small transactions and can shut down your business on a whim with no recourse, and for the customer it's hard to make a tiny digital payment to a service without linking it to your identity, which is often the exact thing you were trying to pay something to prevent.

But that seems like more of a problem caused by not having a good anonymous digital payments system than one caused by advertising. The advertising is just the infelicitous workaround.

> but the users don't see the externalities so they don't realize this

Externalities are when the costs are imposed on someone who isn't a party to the transaction. What you're describing is an information asymmetry.

In theory those can be solved just by providing the information to the users so they can make a better decision, but that's assuming the market is actually competitive. If e.g. you like Android and have one but don't like Google spying on you, are there viable alternatives to Google Play and the other Google services? Judging by how many people actually use them instead of the Google ones, big no. But then we're back to this really being a different problem again, this time antitrust.

▲

pdonis 15 hours ago | parent [-]

> In a competitive market the price only depends on the value to users in the sense that it's required to be lower than that to make any sales.

That's a restatement of what I said: price equals marginal cost equals marginal value--the value to the marginal user, who just breaks even by making the trade.

> It uncouples the lower bound.

Which you've already agreed is the only connection. So again you're just restating what I said.

> You got $0.50 worth of value without paying anything.

First, this is irrelevant now because the advertiser got at least $1 of value (or perceived value) in exchange for $1. The user wasn't a party to that exchange at all.

Second, the user didn't pay any money, but they did pay with their data. But they don't see that cost; it's a negative externality. And it's turned out to be a pretty large one.

> The real problem here is, they can do something you value at negative $10, but the advertiser will pay them an extra $0.05 to do it, and then they do it because "you're not the customer, you're the product".

If I actually assign negative value to using the service, I won't use it at all. They can't mine my data if I don't give it to them. I personally am in this exact position with respect to, for example, Facebook.

If the negative $10 is the net of all externalities, then yes, I can end up using a service that actually makes me worse off, because I don't see the negative externalities.

> Externalities are when the costs are imposed on someone who isn't a party to the transaction.

Yes; in this case the costs of having their data monetized using ads are imposed on users, who aren't a party to the ad transaction.

> What you're describing is an information asymmetry.

I suppose it could be viewed this way, in the sense that the services don't share with users all the relevant information about how their data is used. In many cases they share practically none of it.

> that's assuming the market is actually competitive

In the sense that there are massive thumbs on the scale, yes, I agree the markets in this area are not competitive.

▲

AnthonyMouse 13 hours ago | parent [-]

> That's a restatement of what I said: price equals marginal cost equals marginal value--the value to the marginal user, who just breaks even by making the trade.

The thing you're saying is backwards; the consequence rather than the cause. If the price is $1 and it can't be lower than that because that's the production cost then the user who is only willing to pay $0.90 doesn't buy it, but that has no effect on the price one way or the other.

If that user doesn't exist and the user who values it the least values it at $100, there is no user who is only breaking even. Every user has a surplus of at least $99 vs. not having it and the price is still $1 because there are a dozen companies willing to sell it for $1 who don't want to lose business by charging more than the others.

> Which you've already agreed is the only connection. So again you're just restating what I said.

It's the only connection in a competitive market. Not all of them are competitive.

> First, this is irrelevant now because the advertiser got at least $1 of value (or perceived value) in exchange for $1. The user wasn't a party to that exchange at all.

They are party to it though. They didn't pay money, but they paid in attention and have the option to patronize a different service if the market is competitive. And if it isn't then it's that rather than the ads which is the problem.

> Second, the user didn't pay any money, but they did pay with their data. But they don't see that cost; it's a negative externality. And it's turned out to be a pretty large one.

There are forms of advertising where this isn't required, e.g. you can be pretty effective with search advertising by basing the ads entirely on the search query while knowing nothing whatsoever about the user.

Obviously they then collect the data too because it makes the advertising marginally more effective, but that's the thing where you'd like to pay a nickel to have them not.

> If I actually assign negative value to using the service, I won't use it at all.

The service doesn't have to end up underwater for something stupid to be happening. You could have valued it at $15 originally and then the advertiser pays an extra $0.05 to reduce your $15 value to $5. It's still a positive number but you would much prefer to pay the extra $0.05 yourself than to lose $10 in value, except that's currently unreasonably hard to do.

	▲	pdonis 5 hours ago \| parent [-]
		> If that user doesn't exist and the user who values it the least values it at $100 In a competitive market, which is the case being talked about in what you responded to, that's impossible. If the production cost is $1, more users will keep buying the product until the marginal user values it at $1. If there are no users who value it between $100 and $1, then you don't have a competitive market; either the product or service is too specialized to permit real competition (which is highly implausible for the kinds of services we're talking about), or someone has their thumb on the scale. > Not all of them are competitive. The market not being competitive doesn't create any connection that isn't there in a competitive market. So it doesn't change what I said. > They are party to it though. They didn't pay money, but they paid in attention The transaction in which the users pay with attention is not the same as the transaction where the advertisers pay the tech company to get their ads shown. The latter transaction is the one I was talking about in what you responded to. The user is not a party to that transaction. > you can be pretty effective with search advertising by basing the ads entirely on the search query while knowing nothing whatsoever about the user. Yes, you do know something about the user: what search query they entered. True, it's not as much information about the user as they collect in other ways, but it's still information about the user. If I enter a search query for Depends, Google knows something about me that's of value to advertisers. > The service doesn't have to end up underwater for something stupid to be happening True, but irrelevant to what I was saying, or what you were saying that I responded to.