I have spent the better part of the last decade thinking about one thing: why do people watch what they watch?
Not in the algorithmic, A/B-tested, conversion-rate-optimized way that dominates most conversations in media and technology. I mean the deeper question. The one that sits underneath all the dashboards and metrics and quarterly earnings calls. Why does a particular piece of content — a game, a show, an article, a thirty-second ad — capture a human being’s attention, and why does another piece of content, perhaps objectively better by any measurable standard, get scrolled past without a second thought?
This question has informed nearly everything I have worked on. It informed how I think about why sports and television are so deeply intertwined. It informed how I think about the relationship between remote controls and subscription models. And it is, if I am being honest, the foundational insight behind the bet we are making on free, ad-supported streaming — that removing barriers between fans and the game is the single most important thing you can do to grow an audience.
But before any of that, before the business models and the platform strategies and the media rights deals, there is the brain. And the brain has rules.
The Only Finite Asset
Here is something that seems obvious but is routinely ignored in every boardroom I have ever sat in: human attention is finite.
Not finite in the way that money is finite, where you can always earn more or borrow more or print more. Finite in the way that time is finite. You get a fixed amount each day and when it is gone, it is gone. You cannot manufacture more of it. You cannot optimize your way to more of it. Every second of attention a person gives to your content is a second they are not giving to something else. Every impression that does not land is not a missed opportunity — it is a depleted resource that will never be recovered.
For publishers, producing content that holds a user’s attention is the key to monetization — whether via subscriptions, advertising, or both. For brands, accessing that attention and holding on to it via an ad and message that stays with a user is the difference between a campaign that works and one that does not. For platforms — and this is where I spend most of my time — the question is even more fundamental: can you build an environment where attention flows naturally toward the content, rather than being extracted from an audience that is already exhausted?
I think the media industry, broadly, has gotten this wrong. We have built an entire ecosystem around the assumption that attention can be captured through force — louder ads, more intrusive formats, autoplay videos, notification spam, dark patterns designed to prevent you from leaving. The logic is seductive: if attention is scarce, fight harder for it.
But the cognitive science tells a very different story. And three models, developed between the late 1950s and early 1970s, explain more about how attention actually works in media-rich environments than any modern analytics platform ever will.
Broadbent’s Filter: The Bouncer at the Door
In 1958, a British psychologist named Donald Broadbent proposed what would become one of the most influential models of human attention. His Filter Model — sometimes called the early selection model — suggested that the brain operates like a nightclub with a very strict bouncer at the door.
The idea is simple and, once you hear it, impossible to unsee. We are bombarded with sensory information every waking second. Sounds, images, smells, textures, temperatures — an unrelenting flood of data pouring in through every sense. Broadbent argued that we cannot possibly process all of it. So the brain applies a filter, early in the processing chain, that selects information based on physical characteristics — loudness, brightness, pitch, movement — and lets only the selected information through for deeper processing. Everything else gets discarded before you are even consciously aware of it.
This is why a bright red banner ad catches your eye even when you are reading an article. It is why a sudden change in volume during a commercial break makes you look up from your phone. It is why, in a crowded sports bar with six screens showing six different games, your eyes are drawn to the one where something just happened — a goal, a dunk, a collision — because movement and sudden visual change pass through the filter.
Broadbent’s filter explains the entire logic of the thirty-second television commercial. Make it loud. Make it bright. Make it move. Get through the bouncer.
The implications for media are profound and, I would argue, largely misunderstood. Most advertising is designed to pass through Broadbent’s filter by being more aggressive than the surrounding content. Louder. Flashier. More disruptive. And in the cable television era, when the audience was essentially captive — sitting on a couch, watching a channel they had already committed to, with a remote control that required physical effort to use — this worked well enough.
But in a streaming environment, where the audience has infinite options and zero switching cost, the Broadbent filter works against you. The same aggressiveness that was designed to capture attention in a low-choice environment now triggers a different behavior entirely: the user leaves. They close the tab. They swipe away. They switch apps. The filter still works — it still detects the loud, bright, disruptive stimulus — but the response is not engagement. It is avoidance.
This is one of the reasons I believe so strongly that the future of media advertising is not about interruption. It is about integration. But I am getting ahead of myself.
Treisman’s Attenuation: The Volume Knob
Six years after Broadbent published his filter model, Anne Treisman offered an elegant modification that, in my view, explains an enormous amount about how people actually consume media in the real world.
Treisman agreed with Broadbent that we cannot process everything. But she disagreed that the unattended information is simply blocked. Instead, she proposed that the brain works more like a volume knob than a bouncer. Unattended information is not rejected at the door — it is turned down. Attenuated. Weakened to the point where it usually does not reach conscious awareness, but not eliminated entirely.
The key insight is that certain stimuli have lower thresholds for detection. Your own name, for example. If someone says your name across a crowded room, you will hear it even if you were not paying attention to that conversation. A mother can hear her baby crying through a wall of noise. A sports fan will catch the name of their team on a television in a restaurant even while deep in conversation.
This is the cocktail party effect, and it is the reason background media works.
Think about how most people actually watch television today. They do not sit in rapt silence staring at the screen. They are on their phones. They are talking. They are cooking dinner or folding laundry or helping a child with homework. The television is on — the information is entering their sensory system — but it is attenuated. Turned down. Running in the background.
And then something happens. A name they recognize. A score from a game they care about. A dramatic shift in tone or music that signals something important. The attenuated signal suddenly crosses the threshold and pulls them back in.
This is exactly how I think about the relationship between live sports and casual fans. The game does not need your full attention for sixty or ninety continuous minutes. It needs to be on. It needs to be accessible. It needs to be in the room. And the moments of genuine intensity — the goals, the finishes, the controversies — will pull attention back through the attenuation layer because they cross the threshold of personal relevance.
This is also, I think, why the argument for free, accessible sports streaming is so much stronger than the argument for paywalled subscription services. A game behind a paywall never enters the attenuated background of a casual fan’s life. It does not exist in their sensory environment at all. There is nothing to attenuate. There is nothing that can cross a threshold. The game, for that person, simply does not exist.
But a game on a free platform that is already on the television — playing in the background while they scroll through their phone — is present. It is attenuated but alive. And when something extraordinary happens, when the moment arrives, it has a chance to break through.
A game behind a paywall never enters the attenuated background of a casual fan’s life. There is nothing to attenuate. The game simply does not exist.
You cannot convert a casual viewer into a fan if the casual viewer never encounters the content in the first place. Treisman’s model explains why reach matters more than extraction in the growth phase of any media product.
I think about this constantly. When we were building out our sports streaming strategy, the conventional wisdom was that you needed a paywall to prove the content had value. If it is free, the argument went, people will not respect it. They will not engage deeply. They will treat it as background noise.
But Treisman’s model says background noise is exactly what you want — at first. Because background noise is not silence. It is attenuated signal. It is content that is present in someone’s environment, waiting for the right moment to break through the threshold and become something they actually care about. Every fan starts as a casual viewer. And every casual viewer starts as someone who happened to leave the television on.
I grew up in a house where cricket was always on the television during the summer. Not because anyone had made a deliberate decision to watch every match. It was just there — attenuated, in the background, part of the ambient texture of the household. And because it was there, I absorbed it. I learned the rules without trying. I developed preferences for certain players without consciously choosing to. By the time I was old enough to have opinions, I was already a fan. The content had crossed the threshold a thousand times before I ever sat down and gave it my full attention.
That is the power of presence. That is what a paywall prevents.
Kahneman’s Capacity Model: The Finite Tank
Daniel Kahneman — who would later win the Nobel Prize for his work on behavioral economics and write the landmark Thinking, Fast and Slow — published Attention and Effort in 1973. His model of attention is, in some ways, the most important of the three for understanding the modern media landscape.
Kahneman proposed that attention is not a filter or a volume knob. It is a pool. A finite pool of cognitive resources that gets allocated across whatever tasks you are performing at any given moment. The more tasks you are doing simultaneously, the less of the pool is available for each one. And when the pool runs dry, performance on all tasks degrades.
This is the capacity model, and it describes with uncomfortable precision what it feels like to be a media consumer in 2024.
You are watching a show on your television. You are scrolling through social media on your phone. A notification pops up from a messaging app. Your smartwatch buzzes with an alert. There is music playing from a speaker in the next room. Each of these stimuli is drawing from the same finite pool of cognitive resources. And the pool is not bottomless.
The consequence, as Kahneman’s model predicts, is that the more media inputs you are processing simultaneously, the worse you process each one. You remember less of the show. You engage less deeply with the social media posts. The notification becomes a source of irritation rather than information. Everything gets a little bit thinner. A little bit shallower. A little bit less satisfying.
(Quick aside: this is, I think, one of the most underappreciated dynamics in the streaming wars. The proliferation of platforms and services and content libraries has not expanded the audience’s capacity for attention. It has just divided the same finite pool across more options. The result is not more engagement — it is more churn. People subscribe, feel overwhelmed, and cancel. Subscribe to something else, feel overwhelmed again, and cancel again. The pool does not grow. It just gets divided into smaller and smaller portions.)
Kahneman also introduced the concept of arousal — the idea that the size of the attention pool is not fixed but fluctuates with your level of engagement. When you are genuinely excited, genuinely invested, genuinely riveted by what you are experiencing, the pool expands. You can process more. You notice more. You remember more.
This is why live sports are so valuable. The unpredictability, the stakes, the communal experience of watching something unfold in real time — all of these increase arousal, which increases the available capacity for attention, which means the advertisements and brand messages delivered alongside the content are processed more deeply than they would be in a low-arousal environment like, say, scrolling through a social media feed at midnight.
The data supports this. Advertisers consistently pay a premium for live sports inventory, and the reason is not just audience size. It is audience quality. An engaged, aroused, attentive viewer in a high-stakes live environment processes an ad differently than a half-asleep viewer passively consuming background content. Kahneman’s model explains why.
What This Means for How We Build
I have laid out three models from three different decades — Broadbent in 1958, Treisman in 1964, Kahneman in 1973 — and each of them, developed long before the internet existed, describes something essential about how we experience media today.
Broadbent tells us that attention is selective and that only the most salient stimuli get through. In a world of infinite content, this means that subtlety is a luxury and clarity is a necessity. If your content does not immediately signal its value, it will not pass the filter.
Treisman tells us that unattended content is not gone — it is attenuated. This means that presence matters. Being in the room matters. Being available and accessible matters, even if the audience is not fully engaged at every moment. The breakthrough moment will come, but only if the content is there when it arrives.
Kahneman tells us that attention is a finite resource that is divided across tasks and expanded by arousal. This means that the quality of attention matters as much as the quantity. A thousand distracted impressions are worth less than a hundred engaged ones. And the environments that generate genuine engagement — live events, high-stakes narratives, communal experiences — are disproportionately valuable.
Together, these three models form a framework that has shaped how I think about building media products. They explain why we are seeing many times the audience on free streaming compared to paywalled alternatives. They explain why leapfrogging the cable era and going straight to accessible, frictionless distribution is the right strategy. They explain why the future of media is not about capturing attention through force, but about earning it through value.
The only truly finite asset is human attention. The companies and platforms and publishers that understand this — really understand it, at the cognitive level, not just the dashboard level — will be the ones that build the media landscape of the next decade.
Everyone else will keep making things louder.
I do not think this is a theoretical argument. I have watched it play out in real time. When we launched free sports streaming, every piece of conventional wisdom said we were wrong. That you needed a subscription to signal value. That advertisers would not pay for a free audience. That the economics could not work.
What actually happened was the opposite. The audience showed up — not because we shouted louder or spent more on marketing, but because we removed the barriers that prevented attention from forming in the first place. We let the content be present in people’s environments. We let the moments break through the attenuation threshold on their own. We let the arousal of live sports do what it has always done — expand the pool of attention and pull people in.
Broadbent, Treisman, and Kahneman published their models decades before the internet existed. They were studying Air Force pilots and cocktail party conversations and laboratory experiments with headphones. They could not have imagined a world of infinite streaming platforms and algorithmic feeds and smartphones that buzz two hundred times a day.
But they understood something that most of the media industry still does not: attention is not a problem to be solved. It is a phenomenon to be understood. And the companies that understand it — at the level of the brain, not the dashboard — will be the ones that earn it.
The cognitive models discussed in this essay draw from Donald Broadbent’s Perception and Communication (1958), Anne Treisman’s work on selective attention (1964), and Daniel Kahneman’s Attention and Effort (1973). Kahneman’s later work, Thinking, Fast and Slow (2011), expands on many of the ideas introduced in his capacity model.