According to legend—which means I don’t want to go look up the year and stuff—teabags were invented when a tea company sent out samples in pretty little scrap-silk bags, and the Americans were too stupid to know that was just packaging. They dropped the whole thing into the pot. Didn’t sell a lot of the tea blends, but they wrote back and absolutely raved about the wonderful little packets, could they get more of those? Not being utter fools, the company’s response was a solid “Oh? You want to buy? Sounds great!” and now my kitchen has a few dozen different flavors of tea ready to go for individual cups. But: stupid customers! How dare they misuse my brilliant system?!?
Part of keeping kids alive is figuring out how to deal with someone who hasn’t been trained to all the systems that we don’t even notice– all the assumptions we can’t see, because of course that’s just what you do. They do not inherently know what the important variables are when you say “do not touch that”– is it the effect on the thing being touched? The effect on the thing doing the touching? That if you touch it wrong, it will hurt you? That would be the icing on the cake I’m working on getting mussed, harassing the cat is a bad habit, and electricity or boiling water is dangerous, respectively.
Kids aren’t, generally, stupid. Nor are they failing to think– that theory was actually tested with brain-scans, and the researchers expected to find that adults were thinking more effectively before they made the obviously correct choice. Nope. Turns out that the adults weren’t thinking at all, because they already knew the correct answer, that’s why they were so uniform in giving it– it was the kids who were thinking, and considering vectors that the adults already knew had nothing to do with the problem being solved. The kids hadn’t found those things out, yet, so they tried different things. The more intelligent the kid, the more absolutely off the wall “what on earth were you thinking?” variables they could come up with, and test. (Why, yes, I did remember that study because it fit so well with the “something so stupid only someone very intelligent could have created it” situation, and yes I have very little tolerance for the supposedly Einstein quote about ‘doing the same thing over and over,’ mostly because it gets applied when the person quoting it hasn’t bothered to check if it’s actually the same thing, or just variables that didn’t change anything. The latter is called science.)
Moving up a bit in the “what were they thinking?” olympics, Redneck Engineering. Yeah, a lot of them are in the Hold My Beer category, but when it works, it’s stop-and-cuss brilliant, even when the folks who designed the system(s) and tool(s) used had never considered them being used in such a manner.
Sometimes, it’s simply: why would you want to do that? Video Game designers may be recognizing what was known at one point as the “dancing naked on mailboxes” effect, where what players want to do sometimes has absolutely nothing to do with what the designers were thinking people would do in the game. Yes, someone made a video game where you control individual characters, put in an athletic and busty group of females that could 1) dance in a style based off of a very attractive young lady noted for her sexy dancing, 2) strip down to underwear and 3) jump well enough to reach an elevated surface that let them be seen over a crowd, and they did not realize people would be doing exactly that. It took several re-designs of the system to make it so that mailbox dancers didn’t make it so you couldn’t access the mailbox. One of the early fixes was making it so game objects to occupy the same space, thus dancers just fell through the mailbox. This resulted in people deliberately causing distress in others by standing on the mail box, so it could not be accessed—both the want-to-be-seen dancers, and people who were griefing. (Keep this dual effect in the back of your mind, it’ll come up later.)
The mis-match of designer expectations and user expectations can result in major confusion. Users tend to have this crazy idea that systems should serve them, not the theoretical goal of the system designers.
Sometimes, this is something you really need to fight, because the goal is to get a specific result for a third party. Pencil-whipping documentation in the Navy is not going to make the measurements off of the system accurate, even if it does make things easier for the users.
When the users are the ones who decide if the results are desired, then demanding that they use it the way you want them to is just silly. My favorite example is the OK Cupid rating system, which was designed to help you find a date that was attractive to you. So it sorted the results on a scale of one to five, with “one” being “not if she were the last woman on earth”– which, if she also rated you a 1, would send you both a note to that effect– and five automatically sending her a “hi, you’re cute” break-the-ice note. The guys who designed the system then, with a straight face, a “https://archive.fo/BtTP2“reported that amazingly few women were rating men as a 5… (OK, mostly straight face, the profiles they used as examples are folks who work there so there’s teasing involved.)
Now, the data in the post makes perfect sense if you look at the system from the perspective of the results that an individual using it gets; roughly one in three men were rated as “I do not want to see this profile ever again, and tell him yeah, you too if he thinks I’m ugly” and low single digit percent were rated “automatically send them a note saying I think he’s cute.” Meanwhile, women were rated in a rather nice rainbow-arch, slightly heavier to the least attractive but high single digit percent being among the most attractive. Gals tended to contact guys who were below average on the scale, while guys overwhelmingly went for the top 25%. If you’re aware that guys are very visual, while gals are tea-leaf reading off of everything from ‘the profile picture is sports gear, I don’t do sportsball’ to ‘this guy is way out of my league, I don’t even want to see this, probably fake anyways’, the pattern makes perfect sense. If you’re expecting to get accurate data on visual attractiveness, you are going to draw the wrong conclusions.
So, is this an example of a foolproof system finding that the universe has made better fools? Or that the ‘idiot’ user is just using the system in a way the designer didn’t expect, because it gets them the results they want?
This is an important consideration when you’re making any system– say, social safety nets. It doesn’t matter that the idea is to help people. You have to think about what people will actually do, both in innocence and with deliberate malice. Like blocking off the mailbox in a game, because you can. At the same time, if you design with an eye on nothing but the possible abuses, you block off the ability of people to do anything good, much less anything fun, and if you can’t force them to use your system…they’ll go somewhere else.