Taste at Scale
By Amlan Das, Founder & Editor-in-Chief, DASThere’s a shelf in my office that makes people uncomfortable. A Theranos badge holder, the kind Elizabeth Holmes probably handed to someone who thought they were changing medicine. A WeWork coaster—the good ones, from the period when they were still pretending to be a tech company and not a sublease arbitrage scheme with kombucha. A Quibi notebook, still in the plastic, which feels appropriate because I’m not sure anyone ever actually used Quibi for its intended purpose either. I collect memorabilia from defunct companies. The ones that made magazine covers before they made bankruptcy filings. The ones where, for a period of time, smart people in expensive rooms convinced themselves and others that the story was the same thing as the substance. People ask why I do this. Morbid curiosity, partly. There’s something undeniably fascinating about holding an artifact from a $9 billion company that no longer exists—a physical object that outlasted the entire enterprise that produced it. The coaster survived. The company didn’t. There’s a lesson in there somewhere about what’s durable and what isn’t. But mostly I’m trying to figure out where it broke. Not the obvious stuff. The fraud, the overreach, the founders who got high on their own supply. That’s the easy narrative, the one that makes everyone feel better because it implies the failure was about bad actors rather than bad thinking. I’m interested in the subtler thing. The moment before the moment. When the story they were telling stopped matching the thing they were actually building—and nobody in the room could see it, or nobody wanted to. Because that gap—between narrative and reality, between the deck and the truth—is where everything falls apart. In startups. In companies. In careers. And I’ve spent 14 years watching it happen in slow motion across the marketing industry.
Merriam-Webster named “slop” the word of the year. The American Dialect Society agreed. So did The Economist. Three institutions that rarely align on anything looked at the state of AI-generated content and reached for the same word. Slop. The definition they settled on: low-quality, high-quantity output produced by artificial intelligence. The stuff flooding your feed, clogging your inbox, populating your search results with content that reads like it was written by someone who learned English from a thesaurus and has never had an actual experience. You know it when you see it. The LinkedIn post that opens with “In today’s fast-paced business environment.” The blog article that’s clearly 47 variations of the same sentence reshuffled for SEO. The email that addresses you by name but couldn’t possibly have been written by someone who knows anything about you, because it wasn’t written by someone at all. The discourse since the word-of-the-year announcements has been predictable. Thought leaders are thought-leading. Hot takes are being taken hotly. The proposed solutions range from the puritanical (reject AI entirely, return to artisanal content production, maybe churn your own butter while you’re at it) to the defeatist (this is the new normal, learn to love the slop, resistance is futile). None of this is serious. And almost all of it misses the point.
Here’s what bothers me about the slop conversation: it assumes the problem is AI. That before generative models showed up, we had quality. Thoughtfulness. Work that respected the audience. And now we have garbage. The machines came for our content and turned it into mush, and the solution is somehow returning to the before times when humans made things with care and intentionality. This is a comforting story. It’s also nonsense. Slop existed long before ChatGPT. It was just more expensive to produce. I’ve been in this industry for 14 years. I’ve seen things. Beautiful things and ugly things and things so confusing I still don’t understand what anyone was trying to accomplish. I’ve sat in rooms where people presented creative work with the gravity of unveiling a new scientific discovery, and I’ve watched that same work disappear into the market without leaving a trace. I’ve seen agencies deliver beautifully art-directed campaigns to customer segments that didn’t exist. Not “underperformed expectations”—didn’t exist. Personas conjured from vibes and assumptions, treated as real enough to build strategy around, targeted with precision at an audience that was never actually there. I’ve seen email programs with 40% open rates celebrated as massive wins while the finance team quietly noted that revenue hadn’t moved—because the campaigns were cannibalizing customers who would have purchased anyway. The marketing was successful. The business wasn’t. Nobody wanted to talk about it. I’ve seen creative that won awards and strategies that lost money, sometimes in the same quarter, for the same client. The trophies went on the shelf. The P&L told a different story. The trophies stayed on the shelf. That’s slop. Not because the work was ugly. A lot of it was gorgeous. Tasteful. The kind of stuff you’d put in a portfolio or show at a conference. It was slop because there was no intelligence behind it. No understanding of who it was for, whether they cared, or if any of it connected to an outcome that actually mattered. The content can be beautiful and still be slop. The strategy can be sophisticated and still be slop. The execution can be flawless and still be slop. If you don’t know who you’re talking to—really know, not “we created personas” know—it’s all slop. Just expensive, time-consuming, award-eligible slop. AI didn’t create this problem. AI just made it cheaper and faster to produce.
I watch more football than any reasonable adult should. Not just watch. Study. There’s a difference. Watching is sitting on the couch, reacting to what happens, feeling emotions about outcomes. Studying is trying to understand why things happen—the decisions and structures that precede the visible result. There’s an analyst named Greg Cosell who’s been breaking down NFL film for decades. What Cosell does is show you the game beneath the game. The broadcast shows you the quarterback throwing a touchdown pass. Cosell shows you the protection scheme that gave him time, the route concept that created the window, the pre-snap read that told him where to look before the ball was even snapped. Same play. Completely different understanding of what actually happened. The broadcast makes it look like talent and instinct. The film shows you the system—the preparation, the structure, the thousands of small decisions that made the moment possible. What looks like improvisation is usually architecture. Marketing has a similar gap between what it looks like and what actually happened. The campaign that “worked” might have worked for reasons that had nothing to do with the campaign. You launched during a competitor’s outage. Your product went viral on TikTok for unrelated reasons. Mercury was in retrograde and consumer sentiment shifted favorably. Who knows. The dashboard says success. The why remains mysterious. The channel that’s “underperforming” might be the only thing preventing a collapse somewhere else. It’s doing work you can’t see—building awareness that converts later, reaching customers who would otherwise churn, holding together a system that looks broken in isolation but functions in context. You can’t see any of this from the broadcast view. You have to go to the film. And the film, in marketing, is the customer file. The actual record of who your customers are, what they’ve done, what they’re worth, how they behave. Not personas. Not assumptions. Not the story you’ve been telling yourself because it’s easier than looking at the data. The source of truth that shows you what’s actually happening versus what the dashboards suggest is happening. Most brands don’t watch the film. They watch the broadcast, react to the highlights, call plays based on what worked last week, and wonder why the outcomes feel random.
Bill Belichick—before the coaching carousel, before the girlfriend discourse, back when he was just the most successful coach in NFL history—had this phrase he used constantly: “Discipline is consistent excellence.” Not discipline as punishment. Not discipline as rigidity or following rules. Discipline as the ability to execute at a high level, repeatedly, under varying conditions. The boring stuff done right, over and over, until it compounds into something that looks like genius from the outside. I say this as a Buffalo Bills fan, which means I’m genetically predisposed to resent New England and everything they represent. But I’m also honest. The Patriots weren’t more talented than everyone else. Most years they weren’t close. What they were was prepared. They understood their opponents better than their opponents understood themselves. They had a system that turned average players into productive ones and productive players into stars. They did the work that didn’t show up on highlight reels. And then, on Sundays, it looked like magic. It wasn’t magic. It was discipline. Consistent excellence in the things that don’t make magazine covers. Most marketing has no discipline. Not because the people are lazy. God, no. The people in marketing are usually working harder than they should be, running faster just to stay in place, producing more content than ever before while feeling less certain about whether any of it matters. The problem isn’t effort. It’s foundation. Every campaign starts from zero. Every quarter is a new hypothesis. The strategy depends on whoever’s in the room and what they believe today, which might be different from what they believed last quarter, which might be different from what the data would tell them if anyone had time to look at the data. The dashboards show activity. So much activity. Emails sent. Ads served. Impressions delivered. Engagement achieved. And then you sit down with the CFO and they have a different set of numbers. Revenue. Margin. Customer acquisition cost. Lifetime value. The numbers that actually pay for things. I’ve been in those meetings. The ones where the marketing scoreboard says you’re winning and the finance scoreboard says you’re not. The dissonance in the room is physical. You can feel it. Two realities, same company, and someone has to reconcile them. That gap—between activity metrics and business outcomes—is the slop gap. It doesn’t show up in the creative review. It shows up in the P&L, months later, when everyone’s moved on to the next campaign.
There’s a concept in probability theory called “ergodicity.” It sounds academic because it is, but the idea underneath it is practical and sort of devastating once you see it. The simple version: what works on average across a population might not work for any individual within it. Here’s the classic example. Imagine a game where you have a 50% chance of increasing your wealth by 50% and a 50% chance of losing 40%. Run the math on expected value—the average across all possible outcomes—and the game looks profitable. You should play. But you’re not the average. You’re one person, playing sequentially, experiencing outcomes in order. And if you play long enough, the losses compound in a way that the average doesn’t capture. The game is profitable in aggregate and ruinous for individuals. What’s true for the population isn’t true for you. Marketing has an ergodicity problem. The benchmarks. The best practices. The case studies. The “here’s what worked for Brand X” presentations. These are all population-level insights. Averages. Composites. What happened across a range of contexts, smoothed into a pattern that looks like a roadmap. And they’re applied to individual brands as if the average were a guarantee. As if your customers were average customers. As if your file were the composite file. As if what works “generally” will work for your specific situation, with your specific economics, in your specific competitive context. Sometimes it does. Plenty of times it doesn’t. Everyone wants to be Apple. I’ve seen it more times than I should have. In the earlier days of DBNY, my first agency, we would prompt a brand embarking on a rebrand or website redesign—genuine stupidity prompt, not AI prompt—and ask for references. Inspiration. Visual direction. A sweeping 80% of them referenced Apple in some way. The minimalism. The white space. The confidence of saying very little and letting the product speak. Here’s the thing about Apple: they can do that because they’re Apple. Because they have fifty years of brand equity. Because the product actually does speak—everyone already knows what an iPhone is. Because they’ve earned the right to be minimal through decades of maximum effort that came before. You are probably not Apple. Your customers probably don’t know who you are yet. The minimalism that reads as confidence for Apple reads as emptiness for you. The white space that signals premium for them signals “we couldn’t figure out what to say” for you. The playbook worked for Apple. The playbook might kill you. Same tactics, completely different context, opposite results. I’ve watched brands follow the playbook—the exact playbook that worked for the case study company—and get completely different results. Not because they executed poorly. Because they weren’t the average. Their customers behaved differently. Their margins worked differently. The assumptions baked into the best practice didn’t match their reality. The playbook was right in general and wrong in specific. Which, if you’re the specific, is just wrong. This is where judgment matters. The ability to look at population-level wisdom and ask: does this apply to my situation? To know your own customers well enough to recognize when the average is misleading. To have a point of view that’s grounded in your actual data rather than borrowed from someone else’s deck. AI can’t do this. AI is trained on the average. That’s the whole point—it learns patterns across massive datasets and gives you the most probable output given those patterns. Which is useful for some things and catastrophic for others. The slop flooding the internet isn’t just low-quality content. It’s average thinking applied indiscriminately. The same frameworks, the same structures, the same voice—replicated infinitely because the marginal cost of replication is zero. It’s ergodicity as content strategy. What works in aggregate, served to everyone, regardless of whether it works for anyone in particular.
When people talk about “taste,” they usually mean aesthetics. Font pairings. Color palettes. The inexplicable but real difference between a design that feels premium and one that feels like a template. Brand guidelines. Creative direction. The ineffable quality that separates a good ad from a great one and that nobody can quite define but everyone recognizes. That’s real. I’m not dismissing it. There’s a reason some work makes you stop scrolling and some work disappears into the visual noise of modern life. Aesthetic taste matters. But it’s not the taste that matters most. The taste that matters is judgment. Knowing which customers are actually valuable and which ones just look good in a cohort analysis. Understanding that a program optimized for acquisition volume might be destroying margin through channels that attract low-quality buyers. Recognizing that the strategy everyone agrees on might be the strategy no one has actually tested against the customer file. It’s the ability to see what’s actually happening—not what the dashboards suggest, not what the agency presented, not what everyone’s agreed to believe—and make decisions based on that reality rather than the more comfortable story. This is the taste AI can’t automate. Not because the technology isn’t sophisticated enough. Give it time; it’ll get more sophisticated. But because taste in this sense requires something upstream of execution: a point of view about what matters, developed through experience, applied with discretion. Everyone has access to the same tools now. The same platforms. The same AI models. The same dashboards and data sources and marketing clouds. Execution is a commodity. The cost of producing stuff approaches zero. What’s scarce is the judgment to know what stuff to produce. For whom. Based on what understanding. In service of what outcome. That’s taste at scale. Not aesthetic sensibility applied to more content. Judgment—real, grounded, specific judgment—operationalized efficiently. It’s the only moat left.
I keep the Theranos badge on my shelf because it reminds me of something I don’t want to forget. They had the story. God, did they have the story. A young founder with a vision. A technology that would democratize healthcare. A board stacked with serious people who’d seen enough pitches to know the difference between ambition and delusion. Magazine covers. Conference keynotes. The aesthetics of a company that was changing the world. What they didn’t have was the foundation. The actual capability that would make any of it real. The thing underneath the story that the story was supposedly about. For a while, it didn’t matter. The narrative was strong enough to carry itself. The belief was strong enough to create more belief. The room was full of people who wanted it to be true, and wanting is a powerful force. And then it mattered. It always eventually matters. The gap between story and substance can persist for a surprisingly long time, but it can’t persist forever. Reality has a way of asserting itself, usually at the most inconvenient moment, often after a lot of money has changed hands. The badge outlasted the company. A small plastic rectangle, sitting on my shelf, having successfully existed for longer than a $9 billion enterprise. I think about that when I see marketing strategies built on assumptions rather than understanding. When I see beautiful creative directed at fictional customers. When I see dashboards full of green arrows that somehow don’t translate into business results. The story can be strong. The execution can be flawless. The room can be full of believers. But if the foundation isn’t there—if you don’t actually know who your customers are, what they’re worth, how they behave—you’re just building a more elaborate version of the same gap. Narrative over here. Reality over there. And a growing distance between them that someone will eventually have to account for.
Audience is the only asset that truly compounds. Creative gets stale. What worked last year feels dated this year. The breakthrough campaign becomes the template everyone copies until it’s wallpaper. Channels get saturated. The arbitrage opportunity closes. The algorithm changes. The cost of attention rises until the math stops working. Tactics get copied. Your competitor sees what you’re doing, hires someone who knows how to do it, and suddenly your advantage is table stakes. But a deep understanding of your customer file—who they are, what they’ve done, what they’re worth, how they behave, what predicts their future actions—that appreciates over time. It gets more valuable, not less. Every decision you make gets a little smarter. Every campaign builds on the learning from the last one. The work compounds instead of resets. This is the opposite of slop. Slop is disconnected. Each piece exists in isolation, produced without understanding, thrown into the world to see what sticks. It’s volume as strategy, activity as substitute for insight. What compounds is connected. Grounded in understanding. Built on a foundation that makes the next thing better than the last thing. The word “slop” is useful because it captures something true: there’s a flood of garbage, it’s getting worse, and people are exhausted by it. Fair enough. True enough. Though I should note—as someone with roots in upstate New York—that slop isn’t always bad. Rochester has a dish called the Garbage Plate: a pile of home fries, macaroni salad, baked beans, and meat, covered in hot sauce and onions. It looks like chaos. It looks like someone lost a bet. And it’s absolutely beloved, the kind of thing people drive hours for and defend with genuine passion. The difference is intention. The Garbage Plate is slop on purpose. It knows what it is. It’s not pretending to be fine dining; it’s a late-night regional specialty that delivers exactly what it promises. The slop flooding the internet is the opposite—it’s pretending to be something it isn’t. It’s wearing the costume of insight without the substance. It’s the aesthetic of value without the actual value. Anyway. The conversation around slop keeps missing the point. The problem isn’t AI-generated content. AI is just a production method—faster, cheaper, more scalable than what came before. The problem is content of any origin, produced by any method, that’s disconnected from understanding. When you don’t know your customers, you produce slop by default. It might be beautiful slop. It might win awards. It might even perform well on the metrics you’ve chosen to measure, which may or may not be the metrics that matter. But it’s still slop. Activity without intelligence. Narrative without foundation. Story without substance.
I’m still collecting memorabilia. Last month I found a Juicero packet on eBay. The company that raised $120 million to build a $700 machine that squeezed juice out of proprietary bags—bags that, as Bloomberg famously demonstrated, you could squeeze just as effectively with your hands. It’s a perfect artifact. A physical reminder that smart people, with real resources, can build elaborate systems around assumptions that don’t survive contact with reality. That the gap between story and substance can persist long enough to raise nine figures before anyone checks whether the thing actually works. The packet sits on the shelf now, next to the Theranos badge and the WeWork coaster. A small museum of cautionary tales. Objects that outlasted the enterprises that made them. People still ask why I collect this stuff. I’ve gotten better at answering. Because the failure mode is always the same. And it’s never the one people expect. It’s not that the founders were stupid. They weren’t. It’s not that the investors were naive. They’d seen enough deals to know better. It’s not even that the products were obviously bad. Some of them were genuinely innovative in certain respects. The failure mode is the gap. The distance between what you’re saying and what’s actually true. Between the story you’re telling the market and the reality of who your customers are and what they actually need. That gap can hide for a while. Funding can cover it. Growth metrics can obscure it. Enthusiasm can paper over it. But it’s always there, accumulating interest, waiting for the moment when someone asks a question that the narrative can’t answer. The only defense is the foundation. Knowing what’s actually true. Understanding who you’re actually serving. Building on substance rather than story. Everything else is just slop—dressed up, well-funded, sometimes award-winning slop, but slop nonetheless. And the coaster will outlast it every time.
Amlan Das Founder & Editor-in-Chief, DAS — The Audience Development Agency madebydas.com