To illustrate, first I will lay out the economic template using the concrete case of the 2010s Facebook. Then, I will generalize to more interesting applications, including matters even so fundamental as friendship, and elections, and show you that it is obvious waste/economic mistake to provide or offer non-obvious choices, especially choices involving significant risks, in the provision of all aesthetic and many functional goods and services.
There are some averaging issues, but let us say Facebook generates $20 of revenue per user, per year. Consider (counterfactually) that Facebook was up-front about many of the changes it made to how user data was used, including integration with online shopping and auto-recommendations, as well as third-partying uses e.g. Cambridge Analytica, to include per-use authorization. Facebook applies data in wide variety of ways, including ad targeting, so it is hard, precisely to determine how many pop-ups/notifications users would have had to review and approve. However, if we were to use application permissions, or their settings pages, as rough indications of the policy decision space for users, we easily would be talking about dozens of decisions users would had to make, over the decade. To simplify even further: if Facebook has an existing user set, e.g. of ad targeting, then in order to start using Facebook, a user would have to sit down and make those dozens of decisions upon sign-up.
At $20 or $30 dollars of revenue per year for Facebook to provide all services, we can see that even an hour’s work for Facebook per user per year, is not justified. Hence any truly personalized service does not make economic sense to provide; “personalization” is more general customization, applied broadly to a specific data set, that may not actually yield a usefully improved set of recommendations. Hence to the average user, the actual value of Facebook, is capped above by the reduced value of customized service, vs. personalized service.
The user value from Facebook does not have to match the economic cost to provide services – indeed that is a common economic relationship of goods and services provided. However, there are common considerations within a market economy that allow us coarsely to bound said user value:
- The availability of paid alternatives, with differing service levels, etc.
- Whether those paid alternatives actually get used (implying that the additional value add may not be worth it)
- Non-direct substitutes to the good/service, and how much users pay for those
In this specific case, there is one particular feature that allows us to be very concrete about user valuation in ways explicitly acknowledged even by the social networking industry/ad networks/etc.: the effect of driving users away by intrusive/performance hogging ads. By turning this dial up, services lose a great deal of users, and by turning it down, they retain them. The time impact of advertisements on page loading, searching for information, etc. can be quantified not only as a percentage of how many users stop using the service, but also linked to how much revenue on a fixed user base, notionally would be generated by those additional advertisements, per user minute. Although this second consideration is not directly the user value, it relates to the consideration of users’ alternatives to the good/service because of their marginal opportunity cost, hence it effectively provides a rational stimulus to search for and use a different service, including subscription costs.
Hence from the preceding, we note that the service probably can’t impose time demands, say no more than an hour or two of ads per month, before people abandon it for other services. Just to be concrete, practically this means Facebook can’t be worth more than $1-2K dollars per year per user, based on this understanding and examination of alternatives e.g. cable TV subscriptions.
Now consider the original musing: that Facebook would modify its service, explicitly to prompt users for approvals, and that we consider such approvals over a decade, cumulatively to number in the dozens; furthermore, new users have to take the trailing cost up front. The time cost of reading, researching/understanding the implications (which most Facebook users clearly didn’t), and the minor cost of clicking through and checking the boxes, for each of these dozens of agreements, is going to be a task of at least a day or two if there are dozens of agreements. Hence whether you consider the matter based on time to time, or you refer to the concrete number above of $1-2K per user per year, any new user is going to be hard pressed to justify this cost. The more Facebook sells the data, the less useful the service gets, even for existing users. Note this does not consider that the users make the wrong decision (or even the right one and they just get unlucky) and their data gets sold to someone who outs their sexuality, blackmails them, or otherwise causes some adverse action; the point here is the configuration overhead.
Thus, given how marginal it is, Facebook feasibly can’t offer users data control choices. What actually happens is 99% of users either are forced to click through (businesses that want to advertise), or just choose to ignore the consequences. The 1% of voluntary users that might even consider the controls, clearly are not making a good use of their time.
Facebook is useless garbage, so why do we even care for this discussion? It is because it is easy for me to provide you with numbers and meaningful context without losing important detail; hence when I continue and tell you something more abstract, and controversial – that all aesthetic and most functional aspects of the economy must not require users to make choices, and certainly not hazardous choices – you know the outlines of the lines of reasoning I will present.
Let us review the rough economic situation of a typical developed economy worker:
- Sleep 49-56 hours a week
- 40-50 hours a week of work, to make a middle class income
- 5-15 hours a week of transport and errands
- A self-care routine of ~7 hours a week (shower, toothbrushing, possibly exercise, etc.)
- Some sort of child care/elder care responsibilities
- Of course we also consider there are going to be civic duties, elections, etc. (not double counted below because they are very predictable, but you could account for them with a larger budget also)
So specific occupational training, leisure, etc. has to fit into the free time available, or else you sacrifice something else, and the only one that makes long-term economic sense is moving closer to work or selling things that require maintenance. Practically, we are talking about 1-2 weekend days that even are available for allocation to any purpose. We know that additional schooling or time spent at work has net expected benefits (say amortized $10-20 additional per hour), so this forms a lower bound on the expectation associated with consumer product research, being thrifty with expenses, etc. as well as with free or low-cost activities; as long as the worker still has say half a weekend day to enjoy whatever they bought with their extra $100 per week, the free and low cost options can’t be justified over the more expensive ones unless they are about as good.
The specific acts of configuration, detailed consultation, and non-obvious monitoring (e.g. checking on whether your funds really are invested or whether Bernie Madoff has stolen them, making sure your spouse isn’t cheating) are losses. Of course we recognize that the reporting/facts and advice you receive from them may be extraordinarily valuable, but the acts themselves cost: per the above analysis of the individual worker, they at least are going to cost $10-20 an hour, AND, the average worker has both limited time and financial means. So, practically speaking, the average worker cannot configure, consult, and investigate a large number of items per year, within the general confines of this economic situation. Amongst other reasons, in order for that to be practical, the economic benefit of such would have to exceed that of job training in order to receive significantly higher salary, which on a lifetime or even 10 year basis at least will be on the order of tens of thousands of dollars, and usually hundreds of thousands of dollars.
So in a vague sense, we have established that there is a concrete choice budget, with a sort of harsh cut line. Before we continue, we also should lay out relatively frequent high-value choices that already consume some of this budget:
- Whether to continue working in the existing line of work/location, or to make a change
- Similar considerations e.g. for when parents get old/sick
- Non-routine investments and financial management – life insurance, property, etc.
- Romantic relationships – not the maintenance (that is routine) but the stay together or leave
- What to do about the kids – schools, colleges, not launching
- The nastiest, but also one of the most important: what do you want to do with your life, what do you enjoy, and what are you going to spend time on now (e.g. vacations with sick loved ones) vs. what are you going to try and defer (e.g. the RV retirement)
- The need to look for “gotchas” with the above, like lead paint, maintenance and use restrictions
Proper education and functional economy simplification helps you out here, for example if Warren Buffet tells you to buy the index funds. Conversely, a disorganized economy like most, means these consume more individual time.
The title word “marginal” is the final input. “Marginal” refers to incremental improvement, vs. some other phraseology like revolutionary or new line of business. Almost everything in the aesthetics economy (the obvious exception being physical women) is marginal. Consider as anchor points of large swaths of the aesthetics economy:
- That you can download a huge pile of art on your computer, and display it on a screen, or print it out
- Similar considerations for music, talkies, novels, etc.
- You can have your car painted, vinyl wrapped, and otherwise customized by a wide variety of vendors
- Similar considerations for your house, place of business, etc.
- If you go to a fancy restaurant, a host/hostess bar, or other aesthetic economy leisure establishment, they will dote on you regardless of your looks, as long as you have the money
- At Wrigley Field, a male gets the privilege of a trough
- Now there are even virtual idols, and their underlying AI is only going to get better
For all of these items, not only is the basic capability readily available, but (with the exception of the yakuza-run host/hostess bars) they do not require any sort of EULA, long-term contract, or ongoing engagement. You pays your moneys and you get your goods. If you want to upgrade, e.g. your computer art into a wall scroll, these progression follow the similar rubric; they are not qualitatively different except as to individual aesthetic judgment.
Moreover, when we talk about upgrading, due to the material costs and relatively expensive human labor involved, we clearly went well beyond a $10-$20 threshold of inconvenience. Normally we talk about hundreds or thousands of dollars for these upgrades. We are not even in the realm of trading off single hours.
Let us consider then, the legitimate purposes of configuration, consultation, and monitoring in aesthetics:
- To make the display most aesthetically pleasing
- To train the user on the presentation options
- To avoid the user destroying the work through poor use, neglect, or other fault (though this sort of invokes a functional)
- Proper display e.g. it is not crooked
- For safety purposes (again invoking a functional)
For the first two, clearly the goodness of the configuration is obvious, or at least the user recognizes the ambiguity of choice. To the latter three, since the functional is being invoked, you aren’t giving the user a new/configurable choice, rather you are trying to avoid them making an obviously incorrect decision.
So recapping that:
- All the average person’s choices in aesthetics (and functionals too) are expensive, and they can’t make that many of them
- If they do have important choices, usually either they will spend a lot of time with the subject of their aesthetics (to the exclusion of other important things), and/or they will spend a relatively large amount of money
- The physical reality of aesthetics (with the exception of personal relationships) does not lend itself to an obvious constructive use for any non-transactional mechanism
the concept that strings should be attached is limited by the low maximum number of strings you can attach, and also overcome by spending due to opportunity costs (i.e. you are going to work harder to spend more quality time with your aesthetics), in addition to the mechanical analysis.
As for the previously excluded item of personal relationships: here I merely will invoke cheating, divorce/alimony/child support, that they are in your house/can destroy things, and in general can set you up for problems. So to accept major flaws is accepting a large amount of risk that isn’t worth it.
You ask the question, well why not do only the mechanical analysis, isn’t it sufficient? For a functional economy, yes, it would work to look at time/money investment. However, the proponents of the fraudulent/crooked/complicated aesthetics economy invoke several subtle economic factors, such as value for money when people don’t have money, the opt-out argument, and the “voluntary” choice not to spend money on services. Value for money when people don’t have money, that usually is a false dilemma, because you can work harder to get money, and as shown above it is easy to do so, until you start upgrading, at which point the initial argument no longer applies. The review of limited choice resources above is one way of addressing the opt-out argument. The “voluntary” choice not to spend money on services is demonstrated as obvious human suboptimal behavior when additional work is available, and in particular, like the concrete example of Facebook above, the costs imposed by the alternative economic model are of the order of magnitude that just working or paying outright is easier.