As most people in my life are tired of hearing I was once a historical researcher looking into the likely occurrence of infanticide in an early modern French village. If you buy this book, and read chapter twelve, you’ll see the fruits of my labor.
Full disclosure, I receive zero dollars if you buy the book.
Anyways, the point of bringing this up is to note that I relied of several thousand records to come up with my conclusions. While a discerning person could find these records themselves to see if I am full of shit or not – the sources are listed in the end notes – that would take time and effort. Instead, anyone who reads my contribution can assess the methodology and reasonableness of my argument and decide whether or not to trust me. Ultimately, it is this trust that decides whether my claims are legitimate or not.
The thing about trust is that it’s not an objective entity sitting in a field somewhere. One’s biases and preferences weigh heavily upon the matter. In the case of my work on peasants yeeting their kids, one can imagine a researcher who believes a priori that infanticide is a legitimate historical phenomenon and therefore trusts my conclusions and work because they align with theirs. If the work makes them look good, then they are more likely to accept it. This is quite natural. I don’t fault anyone for it.
A problem arises though when the underlying basis of a conclusion is far removed from the one deciding on the matter’s validity. If a manager tells a CEO that all the employees are happy and the CEO never has contact with the average employee, the CEO can be easily fooled. If the manager seems like a nice enough chap why question the report? What CEO wouldn’t vibe with the idea that his leadership brings joy to the masses?
My contention is that much of the digital economy is a series of self-aggrandizing lies, mistruths, and falsehoods. I believe this to be the case due to the all-too-easy ability to dress up reports about performance and prospects due to self-interest and distance between business leaders and the underlying real world sources. Few managers and executives can understand the piping that facilitates their businesses and are incentivized to assume the best. Allow me to explain.
One of the weird things about my job is that I get a decent amount of face time with senior people at other companies. Universally, these folks work in the realm of bits and bytes. These folks are decent people who come highly recommended and, most importantly, trusted. But often they will say things that simply cannot be true.
Let me be clear. They are not lying. They are just wrong. And so wrong that a simpleton like me can see the obvious falsity of the matter. Yet, they are not lying. They are merely mistaken. Somebody somewhere lied to them, or rather, lied to somebody forty-seven links in the chain of communication before them.
For example, I was on the call with two gentlemen who work for a large(ish) company and are in charge of a new app-based product. The app has a relatively niche addressable market. I won’t give specific information out here that could identify it. Let’s say that the app is aimed at people who swim at least once a month. It’s not. But this is for discussion purposes.
These dudes claim that their app has been downloaded two and half million times since launch two years ago. Additionally, they claim that in the last three months it’s been downloaded about three hundred thousand times. You have never heard of this app. I had never heard of this app. The app has zero marketing dollars allocated to it.
The team claims they engage in search-engine optimization but searching very specific terms associated with it leads to no mentions. I should also add that Google already offers an established free product that does the same thing as this app.
So, you tell me, do these guys actually have two and half million legitimate downloads?
Let’s add in a bit of colour here.
Independent of these folks I know a company whose legitimacy I fully believe in. Sticking with the swimming placeholder I would describe this company as having a top tier breaststroke app. Although most breaststrokers [ain’t that a hell of a phrase] don’t use the app on a day-to-day basis, they at least have it downloaded onto their phones. The app is focused on the US-market and has twenty thousand downloads. It is established, recognized, and has a significant marketing and sales team associated with it. So how is it that this app could have two orders of magnitude less downloads than the other one?
This is where wishful thinking comes in.
The first app – the one with allegedly two and a half million downloads – is not aimed at the US market. Instead, its target audience resides in sub-Saharan Africa, India, and southeastern Asia minus China. I’m told that most downloads come from India.
Let’s pretend it’s just India for discussion purposes. The country has a population around 1.4 billion people. 2.5-million downloads would represent an adoption rate of 0.17%. Hence why the folks on the call can believe the number they told me. The reality of large numbers being what it is, a penetration rate of less than a fifth of a percent seems possible. Especially if somebody somewhere is telling them so.
The story starts to fall apart though when you realize that the target audience of the app is not the entire Indian population but rather a subset. This subset – I estimate – is probably 2-million people. If I go and look up the app on the Indian version of Apple’s AppStore I see 65 reviews (Apple does not disclose download numbers). iPhone adoption in India is very low but this app is aimed at higher socio-economic status persons.
Over on Google Play the app is reported as having more than a million downloads and 6,600 reviews. If you jump into the reviews, you realize that the bulk of them are repetitive, incoherent, or exuberantly positive while echoing the same talking points. Searching the internet for mentions of the app leads to very few mentions. The subsidiary behind the app has 2,000 followers on Twitter and their pinned tweet discussing the app has 64 views. So, what happening here?
Now, I can’t prove anything, but I’m convinced that at most the app has legitimate users numbering in the thousands. My theory is that the app’s team – mostly based in the eastern hemisphere – needed to hit KPIs in terms of downloads. Getting real users is hard. Once upon a time I was on a call with a billionaire VC guy who said ‘the hardest thing in the world is to get someone to download an app.’
So, what did they probably do? I suspect they went and paid for downloads via click farms or bots to juice ‘user’ acquisition. But you can’t get a few thousand ‘users’ and then have it stall out. If you want the app to continue you must keep adding users. Week over week, month over month, more and more ‘users’ must be found. The snowball grows. You need accoutrement to go along with the alleged downloads. So, you pay for ‘reviews’ both in terms of starred reviews and comments in the android landscape because you can do things in Android land that you can’t with iOS because of the cost of hardware. In the end, it’s all a façade.
But if it’s a façade, shouldn’t someone have noticed this before me? The answer to this is no. Who benefits from exposing the issue? The app has not been monetized yet so there’s no customer related fraud. Additionally, everyone up the corporate ladder looks good if this thing grows. If the monetization doesn’t happen – which I doubt it ever will – then the app is just a fun thing the company did one time. In the meantime, everyone gets to point to user growth and say ‘see, my efforts are incredibly effective.
Yet there’s the question of where did the falsehood start?
My guess is that orders went down the corporate ladder to some poor underling who was tasked with ‘community management.’ The underling realized it’s really fucking hard to get someone to download an app. But their job depended on getting app downloads. So, they ended up buying users. They told their boss about the growth. The boss said ‘great I’ll tell the higher ups’ and the good news made its way to the two gentlemen on the call with me. At no point did anyone have an incentive to question the performance. No one would benefit from the truth. Instead, everyone would look like a bunch of incompetent wankers – especially after the first instance of false performance. Again, let me remind you that the guys on the call about the app were respectable dudes at an established company making six figure salaries and not consciously lying to my face.
This is far from a one-time issue. I was on a different call with a recently laid off ad-tech professional with twenty odd years in the ‘performance marketing’ space. As someone no longer working for his prior firm, he doesn’t have to toe a party line which led to some interesting disclosures. The biggest claim made on the call was when this guy claimed he could get app downloads – from real people in North America – at a cost of one to five dollars each.
Dude said this nonchalantly with a straight face.
You know that scene in The Big Short when Steve Carell’s character asks why the mortgage brokers are confessing and he’s informed that in fact they are bragging? That was my feeling in the moment. Except, that this man is a known person to me. He is respected and widely held to be professional and ethical. What he was saying was industry standard.
For context, performance marketing firms are ones that are judged and paid based on specific outcomes such as clicks/downloads/sales etc.… Think of it as simply ‘if I get someone to click your banner ad you owe me X dollars.’ This guy was paid by large corporations for the app downloads his team generated. Note that he was not let go due to having committed a crime. He was let go along with a significant number of other employees because, well, tech these days.
Why I thought he was confessing was because I can tell you right now that nobody will download an app for a dollar’s worth of marketing. At least not in the developed world. Instead, somebody somewhere told this guy that they were generating legitimate downloads of apps at the contracted price. As with the ‘swimming’ app mentioned above, it was a case of nobody benefiting from the truth. The KPIs were met and looked good.
I know what you’re thinking. ‘Evan, these two examples are from small-fry companies and are not indicative of a broader phenomenon.’ To the contrary my friend, the only difference between these folks and the bigger companies is that the bigger fish are believed due to their size and possession of some products that are legitimate.
Take Facebook for example. Hell, take Facebook, Twitter, YouTube, Instagram, and any or all pre-TikTok social media platforms. In the beginning these platforms were all about users pulling content from one another. I would make a post and then my boy Johnny Bag-o-Donuts would see it and be engaged.
Nowadays you’re looking at a 90% decline in the number of people you know who are active on Facebook relative to its peak. Allow me to put it this way. When was the last time you posted original content on Facebook? Or differently, all else being equal, if you had two possible love interests with the only difference being that one of them was regularly posting things to Facebook, would that enhance or detract from them in your eyes? Regardless of the social perceptions of social media content creation it is clear to see that the pull model is being replaced by a push model.
In a push model engagement is driven by the platforms sending its users content they did not explicitly seek out. Think of how noted hair-transplant-recipient Elon swapped the default feed on Twitter to ‘for you’ for a little while.
Why’d he do this?
Because if Twitter populates your feed so that every time you open the app there’s a smattering of new content to engage with then you will keep your eyeballs on screen longer. The folks you follow might only post, in total, 78 times a day. You can read 78 tweets in a handful of minutes. To show real activity on the platform they need more than that. So you send in the content your users are too lazy to find.
How does this fit into the whole ‘somebody somewhere is lying to you’ narrative. Well, let’s put it this way. Facebook et al., are in the business of controlling their own performance data. They have KPIs such as monthly active users and impression levels on branded content. They are, in effect, writing a test while knowing the answer. They need to show relatively consistent – if no longer growing – user numbers and engagement levels. Moving to a push model of content distribution means that they create the illusion of activity for those still on the platform. This is fine because the remaining users are engaged by the content. However, you can also just game the system.
Anyone who has ‘seen’ the backend of a website can tell you that there are an immense amount of benign and malicious bots roving the internet. These can be indexing bots from the likes of Amazon, Google, etc.…. Or less savoury bots looking to do dirty deeds.
Not to mention the whole ‘bot users’ that you see on Twitter, Instagram, and Facebook. Think of the amount of identical FX or crypto-shilling comments on posts.
The platforms combat these but the fact of the matter is that the folks running the bots move faster than the platforms. No Nigerian prince scam ever required a team to utilize agile development methods to reduce adverse modalities of yada yada junior schmucks making 300k a year to not contribute meaningful code.
Big companies move slowly.
The result is that these platforms create a false sense of aggregate engagement through pushing content, but then, the quality of the engagement falls off a cliff because real users are less common on these platforms while bots proliferate. But who’s to say what a real user is? If it’s a registered user who has been active on the platform sometime in the last month then a bot checks the box. Similarly, if it is a registered user who has engaged with content – click, liked, shared – then a bot checks the box.
These platforms have such a high need for active user numbers that there is little incentive for someone to say ‘eh, the numbers and/or quality are dropping like bricks.’
My hunch is that if Meta concocted a way to identify only active human traffic and publicly reported that number their share price movement would look like a bobsleigh track. Instead, somebody somewhere is saying that the KPIs are okay or fucking fantastic and no one is incentivized to say otherwise. Willful delusions abound.
One last example here. Let’s think about Google search results. If Google is as fantastic as they claim they are when it comes to indexing the internet, and website designers are competent, then there is no reason for Search Engine Optimization (SEO) to be a real thing that people are paid to do. Yet there’s an entire industry based around improving companies’ SEO performance. For the price of the SEO you could just buy ads on Google and tell them what parameters to use.
For example, if I type in ‘Proper Cloth’ (i.e. the online shirt maker) into Google right now I get three sponsored links at the top. The first is for a French pseuod-competitor called Pini Parma, the second is for Proper Cloth themselves, and the third is for another competitor called Duer. If I get less specific and instead search ‘custom dress shirts’ I get four sponsored links. In order of appearance they are for Proper Cloth, Pini Parma, Indochino, and Lanieri. The first non-sponsored result I see is Indochino followed by Proper Cloth. If I am Proper Cloth, what is the value of being the first or second organic link if there are multiple competitors above me? If a consumer is looking for an answer – in this case ‘where can I get a made-to-measure shirt?’ – then the sponsored links have already answered the question multiple times. The organic results are – for I suspect 95% of searchers – just there for window dressing.
You should note that Google only makes money from paid links. It has no incentive to show off SEO minded organic ones. Furthermore, Google only gets paid if someone clicks on the link. They are not in the impressions game.
Despite Google owning Google Analytics which seeks to sort out the bots from real users, you can see how being the arbiter of your own performance could lead to temptations to skim off the top in terms of claiming bot clicks as real ones. But who is going to argue with Google about performance? Certainly not the underling marketing major whose job it is to manage the search engine campaign performance.
It comes down to whether or not you want to actually trust Google. If you don’t trust ‘em you can’t make a claim to the good looking KPIs. If you do trust them then you have to take their word for it. I largely trust Google, but I find it hard to believe that somebody somewhere isn’t lying along the way whether it be the big G or the entire SEO industry.
I should point out that there are parts of the internet economy where the rubber hits the road and it’s very hard – if not impossible – to lie.
Typically, this involves money.
You can’t say you’re making a fuck ton of money when, in fact, you are not. Or at least not for long. At the very least money shows up in the real world through dividends, wages, the paying of bills. Shit like that. To that end, if SalesForce says they made X amount in revenue last quarter, I can tell you that they almost certainly did. It’s in these other metrics, like users and activities, where shit gets dicey. But, you have to realize that these metrics are integrally tied up with dollar considerations in terms of perceived market value.
I know I said that the Google thing was the last example, but here’s an inside baseball one. I know of a company that is both a medical publisher and a provider of telehealth solutions. This firm gets pharma companies to pay to advertise on its platforms because doctor-marketing is very lucrative and in theory doctors are on their platform.
One of the places they sell ads on is a secure-fax feature whereby a doctor sends in a prescription from their computer. Several physicians I’ve talked to have said that they use the system in their workplace. However, doctors are busy, and kind of smug. This means that they themselves are not typically the ones submitting the prescriptions.
Instead, it is a physician’s assistant, nurse, or medical administrator.
These folks ain’t the target audience that pharma is looking to reach with the ads. The platform company though turns around and says ‘Dr X saw your ad because his account saw it while sending a prescription.’
It’s a perfectly defensible position. But is it? Or is it a wilfull decision to not ask the obvious question and fudge the truth? Either way no one benefits from the truth. The company wants ‘doctor-eyeballs’ money and the pharma company person who paid to put the ad there wants to tell their boss that they engaged Dr X. Otherwise they’d have to tell the boss that they done goofed.
Anyways, that company’s revenue mostly comes from these ad operations. They’re a publicly traded firm. The revenue and user figures factor into their market capitalization.
In short, somebody somewhere profits off the fudging of reality. Which, if that ain’t the definition of the digital economy I don’t know what is.
Don’t get me wrong, all facets of the economy include made up bullshit, but, in the digital space the non-physical aspect of the landscape leads to people being removed from the reality of what’s happening through abstraction. This distance reduces the tendency to ask questions and increases our acceptance of positive news.
In short, the internet is a fantasy land that none of us know much of anything about. Perhaps Bo Burnham best encapsulated our limitless ability to be wooed into believing the impossible when it comes to the land of bytes:
Could I interest you in everything?
All of the time?
A little bit of everything
All of the time
Apathy’s a tragedy
And boredom is a crime
Anything and everything
All of the time.