The End of Internet History and the Last Ad

People make bad predictions all the time. Getting them wrong is not the problem, some are bound to be wrong. A source of bad predictions is often inherent to the enterprise; things are unpredictable. In almost any field, chaos reigns over order. In some sense, our economic system in the Western world, capitalism brought on by free enterprise, is designed to maximize this chaos and to reap its benefits. You might even call this innovation.

Of course, there’s no other industry that fetishizes innovation like the technology industry, and Silicon Valley in particular. Failing, especially fast, is a sign of success. Companies get founded, and funded, at a moment’s notice. Within a blink of an eye, new industries are born, old incumbents are disrupted. It is a breathless, some might call too much, desire to rethink everything. Nothing is sacred.

Except advertising.

I recently read this interview Antonio García Martínez, who did a 2-year stint at Facebook’s ad team around 7 years ago. When asked about whether he could imagine a different business model for internet companies, he quips:

Look, I mean, advertising sucks, sure. But as the ad tech guys say, “We’re the people who pay for the internet.” It’s hard to imagine a different business model other than advertising for any consumer internet app that depends on network effects. I just can’t think of many examples of viable businesses of that nature that weren’t based on advertising. What else are you going to do? How else do you pay for this?

Looking at some name brand technology companies like Alphabet (fka. Google) and Facebook that primarily make money via advertising, Martinez deserves credit. Just those two combined have a market capitalization over 1 trillion US dollars. Maybe more impressive than that, they employ just over 100 thousand employees all over the world. They are great success stories, especially if you are an investor or an employee. Good job, advertising. Let’s do more of that!

A more subtle source of bad predictions is the motivated reasoning, which ironically both causes and fed by bad predictions. When you have a stake in certain outcome, you seek information that supports that outcome. This is bad enough as is, but the real problem is when you make further predictions based on your now tainted information. From that point on, the machine feeds on itself, until it crashes into the reality. For example, 2016 US elections come to mind.

It is important that we acknowledge out our potential biases, and especially so when those biases are intertwined with our work, and personality so much.  Martínez fashions himself many things, but he’s an ad man, who worked at Facebook, after selling another ad company (under suspicious circumstances) to Twitter. As many other professionals do, ad people tend to see the world tinted with and distorted by their lens of advertising mechanics. Martinez also tries to lend an aura of self-effacing authenticity to his prediction (“advertising sucks, sure”) but people in every industry wax poetic about their fields.

But, here’s a different idea. Let’s imagine an internet where ads exist, but they aren’t how most people use it. In fact, let me show it to you.

I start my day checking m emails on my iPhone, which I paid Apple for, via internet that I Charter Communications for. Then I go for a run, recorded on Strava, which I pay for, listening to Spotify, which I also pay for. During the day, I answer emails on my phone and MacBook, both of which connect to internet via Verizon, which I pay for, on a domain hosted by Hover, using services from FastMail and Google, both of which I also pay for. I use CityMapper to look for transit directions, and Uber to catch a ride. I talk to my girlfriend via iMessage, some other friends via WhatsApp, neither of which have ads, though WhatsApp is subsidized by it. I use Slack to talk to clients and other friends. I read news from sources that both rely on as and subscription revenue. When I’m home, I read stuff on my Kindle, reading books I paid for and maybe watch some stuff on Netflix. 

Of course, I am not representative of all internet users; I have discretionary income to pay my way for some of these services. You could certainly skip some of them, or prefer ad based versions for things like Spotify and news. But I am also not a fringe user, on a mission. I am privacy conscious, but I am not a zealot. And these companies aren’t fringe companies, operating out of a Swiss datacenter in Alps. Most of them are giant, billion dollar businesses.

This is why “ads pay for the internet” is not just misleading; it’s plain wrong. But more importantly, such proclamations taken as maxims, blunt human imagination. It fits with our existing thoughts -advertising sucks-, provides some cherry picked data that fits that narrative -Facebook is great-, and the conclusion seems to be self-evident. “What else are you going to do?” says ad man.

But who says, that we should do any of this? Is progress of human civilization in general, solely marked by more ad based technology companies? 

I argued before that “tech company” isn’t a useful category. Instead of clarifying, it muddies the discussion and conflates unnecessary concepts together. Netflix executives seem to agree. Listen to Reed Hastings, founder CEO of Netflix and a Facebook board member, on his last investor call

I’m very glad that we built the business not to be ad-supported but to be subscription… Just objectively, we’re much more of a media company in that way than pure tech. Now of course, we want to be great at both, but again, we’re really pretty different from the pure tech companies.

I am always wary of people calling anything “pure”, as if there’s an ultimate purity you could reach. (What would a pure tech company look like? Would we able to interact with it in meatspace, or would it just be in the cloud? Or maybe it’d replace the meatspace entirely, a la Oculus?) 

Skip the mission statements and manifestations, and look at the operations of these pure tech companies. Let us, for a moment, just look at them as public companies who are legally obligated to maximize profits and provide returns for their investors.

Facebook office space

At first, the ad duopoly appear as hugely successful businesses, that throw money off at impressive 45% margins, and still growing. They manage this feat 3 major ways: a) automated their sales force, b) a free and unlimited source of content to put ads against (web in Google’s case, and both web and human activity in Facebook’s), c) and a hyper-efficient and automated price discovery mechanism (bidding).

A key is this: human intervention is kept to a minimum. This is partly matter of technical principle (“automation is good”), a shield against ethical criticism (“algorithms are less biased than people”) but is largely driven by good old financial motives. Like most other business, personnel costs are the biggest line items in pure tech. And once you take a hit of software margins, it’s hard to give it up.

At some level, you want to empathize with (but not too much) the tech employers. The folks who design, build, and maintain the technologies we take granted for are quite expensive to hire and retain. There are quite a few of them in the labor market, and since they all live in a tiny region that refuses to build housing, they are paid enormous sums. In Facebook’s case, the median yearly pay is almost a quarter million dollars

A Bay Area house (and a lot) with an asking price of $800,000.

If that sounds absurd, it’s because it is absurd. For me, though, it was also great. I too worked in that tiny region several, and competed in that labor market with those folks for quite a while, and I also got paid comparable sums. My salary went up by almost 25% on a happy Monday morning in 2011, thanks to a talent war between Facebook, Google (and in small part, LinkedIn).

But let’s go back to the question. What is all this for? How does this make internet better? Moral compass of capitalism is a hard one to read, but you’d think most of the gains of the market would be largely aligned with the human progress, and most of the skew be regulated away. But that’s not always what happens with an ad based business model.

Time for some nostalgia. I was born in 1986, and I built my first website when was 10. By 12, I was giving quotes to national papers about the possibilities of this new-fangled capital-i Internet thing. That’s not too old by internet standards, but early enough to remember a different internet, a cyber you could say, than what we have today. It was better, in some ways, but in many other ways, it was also worse. 

This might come controversial. But before Google, searching for stuff online wasn’t great. There were search engines, like AltaVista, and Lycos, and HotBot but looking things up online was more of a process, than a step. The first person to tell me about Google was my academic-turned-diplomat aunt, who used it to find legal papers online. Web developers got trained to keep things faster. 

And it was clear how it was all funded. There were ads, shown on result pages. The genius was when you Google your intent to buy something, say a pair of sandals, they could shop that around to people who could satisfy that. Occasionally you accidentally got an ad for Sandals islands, but it was good enough. Google minted many millionaires for years, even with the fraction of personalization it does today.

Something flipped, maybe around 2011 or so. Things on Google started to get out of hand. Search results and ads shown became personalized with no opt-out available. Google search became rich, with actual content as opposed to just links and vertical search became more integrated. 

2011 is also around the time when Facebook started to evolve from a social network to the behemoth it is today.

Interestingly enough, I remember the internet before Facebook too, and it was…OK. We didn’t have the ability to poke that cutie in AP Physics, but we managed. Keeping in touch with friends far and away was harder, but wasn’t impossible. It certainly wasn’t more expensive. 

That’s the uncomfortable truth behind the social networking companies, and especially Facebook. They can exist at the scale and profitability that they are at only by a pervasive, invasive surveillance mechanism, while shrugging away the negative externalities. Not only I am not convinced that a social networking company running off an ad network is a good model for other tech companies, I wonder if we really want this stuff to exist in the first place. That’s a different question than what we should do with Facebook today, but it’s an informative one for making plans and predictions for the future. What do we want our future internet to look like?

So on its performative existential drama blog “Hard Questions”, Facebook once itself asked this question: Is Facebook bad for us? The answer was a mumbled “yes”. Facebook found out that using Facebook the way Facebook designed itself for, made you you unhappy. The prognosis was less satisfactory, of course. After Facebook found out Facebook was bad for Facebook users, Facebook would fix Facebook so Facebooking wouldn’t be as bad for Facebook users. The entire saga was self-contained from the start to finish.

The necessity of Facebook has always been less clear, than many other companies. The way it so closely resembled real life, but fell short bothered me for years. Following the great techlash of 2017, news media folks like Matt Yglesias wrote cogent, incisive questioning whether Facebook should even exist. Maciej Ceglowski, back in 2011 compared the entire enterprise to a “bar ran by a mormon”, where the “social graph” of Facebook is neither social, nor a graph and also asked why it’d even be good for.

But maybe most striking was how vacuous Facebook’s explanation for its existence has been. While it made sense to put the entire world into a computer, if you are a computer person, it still lacked a moral principle, according to Zadie Smith:

The striking thing about the real Zuckerberg, in video and in print, is the relative banality of his ideas concerning the “Why” of Facebook. He uses the word “connect” as believers use the word “Jesus,” as if it were sacred in and of itself: “So the idea is really that, um, the site helps everyone connect with people and share information with the people they want to stay connected with….” Connection is the goal. The quality of that connection, the quality of the information that passes through it, the quality of the relationship that connection permits—none of this is important. That a lot of social networking software explicitly encourages people to make weak, superficial connections with each other (as Malcolm Gladwell has recently argued1), and that this might not be an entirely positive thing, seem to never have occurred to him.

Zuckerberg, to his credit, came to the same realization around 7 years later:

For the last decade or so we’ve been focusing on making the world more open and connected. But I used to think that if we just give people a voice and help some people connect that that would make the world a whole lot better by itself

Now, Facebook gears up to not just connecting people but actually putting its thumb on the scale on deciding whether a connection is good, or bad. The scale discussion is over, in some sense. The dots have been connected, it’s time to color between them. Of course, none of us gets a real, meaningful vote on if this is desirable. Should we keep going down this path, if our guides were previously so bad at their jobs that they sparked an outrage from United Nations for inciting genocide, as well as probably swinging an election in US by mistake?

Facebook might be a good thing, on net, as they say. Certainly the markets think that way. But even then, I think our calculations are currently too generous to Facebook partly because we aren’t factoring in the negative externalities. At least in the US, we are certainly not accounting for the real economical and social costs of privacy breaches. Put another way, as a global society, we might be subsidizing Facebook (and of course Bay Area landlords), for some dubious benefit. 

The untold danger of “ads pay for the internet” meme isn’t just the business model encourages mass surveillance, addictive behavior, creates moral hazards that aren’t accounted for, or helps some people in Bay Area make money off of other people’s work. It’s the idea that we are done. That we are at the end of the natural progression of technology where we stopped looking for other business models. That we are so enamored by not just the riches, but the vastness and magnificence of these businesses that we turn off our critical thinking. The real risk is us looking at what worked for these companies, and these companies only, and replicating for an easy win.

 Is this a good way to model the technology and internet for the future? 

Pewdiepie on right. He is the most subscribed star on YouTube.

I am not convinced. And neither should you be.

But I think there is light at the end of the tunnel. There are many other new businesses to be built online, where incentives of parties are better aligned. People will still mess things up, some people will still make more money than others, but we can do it all that while not subjecting ourselves to surveillance and accidentally throwing out democracy with the bath water.

Internet is relatively young. Internet technology companies are growing at a faster clip than most, which means there is more opportunity to experiment and build business models that don’t rely on just advertising and surveillance.

As more parts of the technology gets commodified and easier to use, different kinds of businesses, will emerge. Instead of simply optimizing for ad revenue, these tech businesses will try to maximize profits by aligning incentives of their users closer to theirs. Instead of defining the contours and colors of cultural output by advertising, we might optimize it for their inherent goodness. 

Netflix built a giant media company, sitting on the shoulders of Amazon for much of tech structure. And the business model that encourages both quantity and quality produced critically acclaimed shows like Orange is the New Black, and House of Cards. And on the other side of the scale, there are like Ben Thompson who have single-handedly built an highly influential analyst business, by wiring together a few other companies’ products. In between extremes are companies like Patreon, unsurprisingly founded by an artist, that fill in the gaps. Compare the output of these businesses to those coming out of Facebook, or the most successful YouTube stars. A digital farm and a shock footage of dead people shouldn’t be our high water marks for 21st century culture coming out of Bay Area.

Counterclockwise: Old, recent, and current London Trocadero

Advertising is a way of making money, much like selling cupcakes or cigarettes. It can be the difference between a small business thriving or closing up shop, like most do. It can be an outlet of creativity, and carrier of culture. But, just as easily, it can be eyesore. It can also be a scourge on our public spaces, an source of overconsumption, and the main driver of a surveillance capitalism. There is nothing about it that makes it one and only candidate for being the engine of the internet as we know. It is a result of choices that we all make, as regulators, as investors, as innovators, and as users. We can do better. Fear is the mind killer.