How Much of the Internet Is Fake?
nymag.com
nymag.com
The rampant fraud we're talking about involves making a fake website (or taking a real one and hot-swapping its ad code) pointing a farm of devices at it and tricking the analytics services AND the buyers into thinking they were put in front of the audience of 500k people, give or take, when 0 of them are human.
The only way a scam like this works is because the ads are sold by robots, bought by robots, and measured by robots in lots of small batches, so a person can't just look at the site on their laptop and confirm that their ads are running even if they wanted to.
TV audiences were measured through surveys.
Also, the goal of advertising is often not to convince the user they want something, it's to get them familiar with something through repeated exposure so when they do have a need, they choose the thing they have been made familiar with through the ad campaign.
familiarity breeds contempt. And in the end we hate the ads and the products.
With all the recent news [0] coming out around FB and their practices, I've only one thing to say now: I knew they were evil, but I didn't think they were stupid too.
[0] 'fake', who knows after this article?
If Facebook wants to say I watched 1 minute of videos and its because they saw me stop scrolling and continuously digest a couple seconds of some buzzfeed video that is fine with me and I would think is honest.
But if their calculations includes videos that I "watched" for 1 second or less than 1 second, and didn't even pause my scrolling.... that is pretty wretched.
Worse, they count videos that auto-played on the page after you scrolled, then walked away from the computer to get a beverage.
[1] Yes, this actually is the correct figure: https://en.wikipedia.org/wiki/Television_consumption. 3 hr 58 min per day; 35.5 hr per week; 77 days per year. And if you are thinking to yourself "Well I don't watch any!", remember that this just means someone else is watching twice the average to make up the difference.
Given that the videos auto-play, I wouldn't be surprised if a significant portion of the count was from videos autoplaying as people scrolled theough.
It's a captive pair of eyes that can be subliminally programmed, even if no deliberately watching.
The FB numbers likely include accidental plays and plays by people who watch a few seconds trying to identify what a video is, then move on. Nobody would say they watched football last night if they flipped past ESPN and left the game on for a few seconds to see who was playing, but if Facebook were showing the game they would count these people as viewers.
A lot of people who don't watch a lot of TV hear the TV numbers and picture someone sitting on the couch watching the screen for four hours a day, when it's more likely people who leave the TV on while doing other things. If you called these people on the phone and said what are you doing, they'd probably say "making dinner" or something similar, not "watching TV."
While it only ever measures a sample size of participants and thus is still misleading, that's not how the Nielsen metrics work. You have to opt into the program and self-report your usage, specifically what shows you actually watch.
I'd note that television viewing statistics are also likely inflated, at least for a normal person's definition of "watching TV". How often have you been in front of a TV in a bar or airport, but no one is actually watching?
TV viewing in bars and other public locations is specifically excluded from TV viewing numbers. The same is true for hospitals, asylums, prisons, nursing homes, and even college dormitories.
Also, the way it works is that the broadcast only gets counted as viewed if someone watches for 300 continuous seconds. And then credit is only given for the 15-minute block (:00-:15, :15-:30, :30-:45, :45-:00) in which that 300 second span happened. And if the 300-seconds spans a 15-minute boundary, then it doesn't count at all.
For example, if someone watches a TV program from 6:01pm until 6:07pm, the broadcaster gets credit for 6:00pm-6:15pm.
If someone watches from 6:11pm-6:19pm, it doesn't count at all.
I'm not sure why this is defined as "300 seconds" and not "five minutes." When I was in broadcasting, I used the phrase "five minutes" twice and was harshly corrected by my bosses both times, so there must be a reason.
After a year the CEO stood up with the marketing VP to talk about their amazing marketing accomplishments. I was confused as still nobody really knew our company other than existing customers, our marketing was awkward, social media accounts even more awkward and dumb, and so on.
So they rolled out their big marketing accomplishment. Old youtube advertisements (also way awkward / looked like a video of a bad power point presentation) had been skipped by users at a high rate. Now, ZERO of our youtube ads had been skipped in the past 6 months. This of course made little sense as someone, somewhere, had to have skipped an advertisement right?
Turns out they shortened the ads so that they were below the time that allowed a user to click skip.... and this was their big accomplishment.
It was a non real accomplishment as far as I was concerned ... but they got all the high fives and bonuses, so I guess it was real.
Glad to hear you've moved on. I hope the current 'thing you do all day' isn't quite as easily fooled.
If I say I drink 2l of water a day, it's not a reasonable interpretation that I'm quaffing 2l in one sitting once every 24 hours.
Here's the HN discussion about that:
I wonder if attitudes towards trusting content on the internet are partially a generational thing. Frankly, the idea that "trust" ever really existed on the internet (post very early days, anyway) strikes me as slightly absurd. Having gotten online in the late 90s/early 2000s, my default position, augmented by countless stranger-danger-type warnings from parents and teachers and slightly tongue-in-cheek disclaimers on websites[1], was that everything was bullshit until proven otherwise. I don't want to sound too sanctimonious, but that sort of skeptical attitude is a good defense against being taken in.
In contrast, I've (totally anecdotally) seen older age cohorts and/or people who just ended up online later for whatever reasons develop much more relaxed, trusting attitudes towards internet content. The stereotypical boomer aunt for whom Facebook is the internet tends to be much less suspicious of content they see, as, after all, it's being posted by their real life friends! Why would one's real life friends go on the internet and tell lies? This is as good a place as any to plug my pet theory that real-name policies are actively harmful, as they incorrectly imply trust where none should exist. This dynamic, coupled with attitudes towards media outlets tuned by eras in which there were only a handful of (generally) credible outlets, seems to be a recipe for complete credulousness when transplanted into a low-trust medium like the internet.
So, rather than a redefining of the essential qualities of the internet, what seems to be in the offing is a general recalibration away from overly-trusting models of thought among segments of the population who never really developed the sort of awareness that I'd argue is necessary for not ending up as a total dupe.
Of course, the article's points on _why_ a lot of fake content exists totally stand - profit motive is a massive driver of that sort of behavior, and I genuinely hope that the ad model crumbles under the weight of its own depravity.
---
[1] "The stories and information posted here are artistic works of fiction and falsehood. Only a fool would take anything posted here as fact."
I particularly enjoyed this sentence coming as it did after the following:
> If a Russian troll using a Brazilian man’s photograph to masquerade as an American Trump supporter watches a video on Facebook, is that view “real”?
> Next time Russians want to puppeteer a group of invented Americans on Facebook, they won’t even need to steal photos of real people
Alternatively, virtue signalling refers to showing off our sincerely held beliefs to gain social credit, but I don't see why that would be bad either; we do many things (like make friends and show our accomplishments) to gain social credit, so why should voicing our opinions (with the effect of having like-minded people befriend you) be such an action to deserve the derision it gets?
It seems that "virtue signalling" is a very flimsy accusation; it is to say "you're lying, since nobody could possibly hold such a kind position sincerely". Maybe it's even an effect of an age in which we're so cynical that we expect anyone who holds a kind opinion to be acting insincerely.
The answer to accusation of virtue signalling should always be this simple: show evidence that I don't sincerely hold the opinion, then you might be onto something and we can start talking about social credit and instrumental rationality. Until then, an accusation of virtue signalling is as substantive as saying unicorns are real.
If there was a murder in the news and you immediately went on Facebook to post "Murder is bad" so people would Like your post, that's virtue signalling. Of course murder is bad. Only a crazy person would think murder wasn't bad. So there's no point posting that opinion on your feed, unless your goal is to get Likes.
A less silly example might be if an acquaintance posts a picture of themselves at a restaurant and you post an unsolicited reply saying "I would never eat that, I'm a vegan." Ok, maybe you really are a vegan, but you're just posting that to look morally superior.
While that's a fairly extreme example, you see some of this playing out with more recent murders, like those in anti-abortion extremism, where there is a constituency that sees them.as justified whether or not legally murder.
Btw when you say "recent murders, like those in anti-abortion extremism", that makes it sound like that's much more recent/common than it really is. The last three such cases were in 2015, 2009, and 1998, per wikipedia (granted the 2015 case left three dead). I wonder why those cases get disproportionately so much more attention than, say, the ongoing genocide of Christians by ISIL.
Whether or not it is virtue signalling about a "sincerely held belief" is kind of irrelevant. Sincerity is also not, as they say, computeable.
You see this debate regarding the value of, say, a Harvard education: is it the knowledge gained, or the "signal" of having been accepted to it?
Similarly, when I hear accusations of "virtue signaling", I understand the person to mean that the action in question was used to signal something, not to inherently accomplish something.
As an example, consider beliefs about global warming. Whether or not you "sincerely" believe in global warming is immaterial to most accusations of virtue signaling. What you do about it is the question. The question to ask is, what did that action accomplish? So, posting a link to some article on facebook is likely to not really accomplish anything, and so the point of posting it was likely to signal. On the other hand, posting about how every year your family takes a trip to Europe, but you've realized that uses too much fuel, and instead your family is going to have a nearby "staycation" in the woods out back: that actually accomplished something, and I wouldn't expect to see that called "virtue signaling".
Edit: That probably sounds more confrontational than I mean it to be. Rephrased: I think 'claudiawerner nicely exposes some of the flawed reasoning behind characterizations of "virtue signalling." What advantage in understanding do we gain by divorcing "signal" and "accomplishing something" that is not covered by either (1) (unsubstantiated) insincerity, or (2) sincerity that we don't find objectionable (the "Alternatively" clause)?
It could also be said, using your point of view, that the 'virtue' part is important, because the virtue isn't held sincerely if you're just posting a link about it. You tie the virtue signalling to the authenticity of the signal, as do I but in a different way. The sibling comment to this one makes the same point.
This issue also comes into play when one is accused of virtue signalling when in an Internet argument, which I think is mostly a bad faith accusation lacking evidence (which I mentioned earlier). Arguing for a particular position isn't necessarily virtue signalling.
Well, sure, I wouldn't say "virtue signaling" is "bad". Just that on a scale of 0-10 of virtuosity, it's like a 2 (but not negative, i.e. unvirtuous). I usually see it in the context of "this is just virtue signaling", meaning it's only a 2, while the accused is maybe acting like they're doing something more of a 5. Like, I see it more as calling out someone for being a little pompous, not insincere.
I've never accused anyone of virtue signaling or been accused, just seen it thrown around from time to time, so maybe my understanding of its meaning is different from yours. I see now in the article that the definition they're using is more aligned with yours, so maybe I'm the odd one out here. I'll have to pay attention the next time I see it "in the wild".
1) Take some small cut of commerce from would-be advertisers and/or subscribers. Use a not-for-profit model.
2) Provide curated searchable catalogs with aggregated subscriber feedback, 3rd party and internal reviews. Subscribers drive the discovery, and aren't indiscriminately sold to when they're not interested in buying.
3) Maintain some secondary and unobtrusive method(s) of random discovery, i.e. trending products, latest additions, or true-random exposure in some limited format to help underdogs and potential new markets get traction.
4) Help subscribers discriminate by disclosing which products/services are themselves advertising-subsidized and what their overall funding model looks like whenever possible. (Generally, help subscribers know who they're dealing with.)
It almost sounds like Amazon or a similar web store with important differences being they don't engage in advertising products themselves and they don't have a profit motive so much as solvency. If such an organization were to be democratically controlled by its subscribership that might also prove to be a good thing.
As for the Why of supplanting advertising as we know it, if that's not something clear:
1) The coupling of content/speech to advertising interests has bad effects on the trustworthiness of the content. It's a corrupting force.
2) Advertisements are wasteful of time (and money), and low on useful information. The targeting is awful and not voluntary or transparent. I like to think of ads as wasting mental space, though that's not a real quantifiable thing by any means I know of.
3) They incentivize some of the worst microeconomic behaviors and general human qualities (which have macro scale effects in aggregate) and do it through exploitative means, using psychology against people.
The proposed solution, when it originates outside of Silicon Valley, is always a Shakedown, and “Only” a Shakedown.