In 2019, dating apps looked like they’d killed professional matchmaking.
Tinder had some 50 million users. Meeting someone was frictionless, free, and infinite in supply. The human matchmaker seemed like a relic. It seemed, for a while, that a generation of suiters retreated from traditional methods of meeting partners (like through friends, clubs, and social gatherings) and leaned instead on frictionless digital matchmaking.
But the trends and numbers seem to be inverting. Between 2023 and 2024, Tinder lost 594,000 users. Bumble lost 368,000. In the UK alone, 1.4 million people left dating apps in a single year. Tinder subscriptions dropped 7% year over year in the third quarter of 2025. Forbes Health found 78% of remaining users reported emotional exhaustion.
Meanwhile, the matchmaking services market hit $4.37 billion and is projected to reach $10 billion by 2035. Elite services charge $25,000 to $250,000.
The thing apps were supposed to kill is thriving precisely, in part, because the apps exist. Anecdotally, I’ve heard a lot of folks talk about joining book clubs and run clubs and some about going back to church.
Sean Blanda captured the inversion in a recent LinkedIn post: dating apps actually increased the share of the market going to matchmakers, possibly due to exposure and the desire for a better version of the apps. The abundance of low-quality connections made curated, high-quality connections more valuable.
Here’s how he put it:
“The algorithmic, automated version exposed more people to the IDEA of matching with someone but in a kind of empty dehumanizing way. People were exposed to the concept, and that led to more people that were willing to pay for the human, bespoke version.
The same thing is happening to AI and editorial work. AI outputs showed more companies what editorial could do, but readers (and distro channels) are getting wise to slop. Now, companies are starting to come full circle: creative, thoughtful people creating things to serve an audience (and yes, make all of those top of funnel numbers sustainably go up and to the right).
If something is important to you, do you want the cheap, low quality version or the thing that actually works over the long term?”
It does seem to be a small but mighty trend, definitely buried beneath the cacophony of generic content and noise. But I’m starting to hear the beat of a new drum, one centered around unique data, insights, experience, and taste.
The Pattern of Abundance
There are examples everywhere once you start looking.
Spotify made recorded music nearly free to access. You can stream any song ever recorded for $0.003 per play. This is insane. I once, as a young lad, adamantly bought Pepsi products in hopes of getting a cap that contained a free $0.99 song on iTunes.
Now, people don’t consume less music. In fact, it seems like there’s an appetite to do so more intentionally. We see a 278% increase in concert ticket prices since 2000 (average ticket: $35.96 then, $135.93 now – three times the rate of inflation) and a substantial vinyl revival. When the digital copy becomes cheap, the embodied experience and the physical artifact become gold.
Smartphones made photography free, but the average wedding photographer now charges $3,000, up 58% in seven years, outpacing inflation.
I’ve been to several weddings in the last few years. You know what else we’re seeing? The rise of polaroids and digital cameras. Time is a flat circle.
Free online news crushed newspapers. Then Substack emerged. People wouldn’t pay for news, but they would pay for a specific writer’s judgement. Information may be incredibly abundant, but discernment and voice, a curatorial, editorial touch – well, that is still scarce (and thus valuable).
Lots of examples. The point is, when something becomes free or abundant, look elsewhere for where value is accumulating or reallocating.
The Paradox of Efficiency
In 1865, an English economist named William Stanley Jevons observed something counterintuitive.
James Watt had improved the efficiency of the coal-fired steam engine, reducing the coal needed per unit of power. Common sense said coal consumption would fall. Instead, it soared.
Better efficiency made coal viable for entirely new applications such as factories, railways, and ships, and total demand exploded.
“It is wholly a confusion of ideas,” Jevons wrote, “to suppose that the economical use of fuel is equivalent to a diminished consumption. The very contrary is the truth.”
Satya Nadella has invoked the Jevons Paradox to describe AI. The basic logic is that when AI makes software development cheaper, total demand for software increases, not decreases. When AI makes content creation cheaper, the market for content-related services expands, but the value migrates within it.
AI doesn’t shrink the market for marketing. It expands it. But it redistributes where the value accumulates. The execution layer becomes less defensible, as Ben, our Director of Organic Growth Strategy, has argued. But the judgment layer, the taste layer, the relationship layer… those expand.
The pie gets bigger, but the slices of the pie change.
What’s Becoming Cheap
The dust hasn’t settled on the effects of AI. Of course, we see some clear trends playing out – more LinkedIn posts being written, inboxes being filled, AI design watermarks all over advertisements, and a certain je ne sais quoi to Claude designed websites.
What I believe about AI is it has largely made obvious what was previously subtle – much of what we thought was valuable merely looked valuable on paper. Think of the “ultimate guide” from years past – the veritable book report of SEO content. Certainly, it drove traffic. But what kind of traffic? And the content quality on most of these was passable, but they were effectively rehashed Wikipedia articles with a few paragraphs of exposition.
Basic SEO analysis is facile with AI – cursory keyword research (the kind where you run a content gap analysis and order by volume), competitive audits, basic reporting. The stuff agencies would charge a few thousand dollars for while clients ask in internal meetings, “hey, what is our SEO agency actually doing?”
Design execution, even for non-designers, is smooth now – layouts, templates, stock-quality visuals.
Distribution mechanics related to content are at the push of a button – scheduling, formatting, repurposing across channels.
Information synthesis is so much faster and easier than it used to be – summarization, research compilation, Q&A. Note: I find this latter category to be absolutely valuable when leveraged the right way, bringing us to new spheres of possibility. A topic for another essay…
These are all still essential, just no longer a real source of competitive advantage. Electricity is necessary for every business, but nobody competes on their electricity.
The mistake most companies are making right now: treating AI as a way to do more of the cheap thing, faster. More blog posts. More social variants. More email permutations. This is the Spotify-era record label pressing more CDs.
What’s Becoming Valuable
Joel Spolsky wrote an amazing essay on value in 2002, summarizing lessons from microeconomics and applying them to software businesses:
“All else being equal, demand for a product increases when the prices of its complements decrease.”
As he wrote, if flights to Miami get cheaper, demand for Miami hotels increases (more people need a place to stay in Miami).
While I don’t have a crystal ball or an Omniscient (ha) view of the market, I do have a pretty good angle on what’s happening with regards to SEO, content, and growth marketing.
Let’s take two fairly banal observations: the amount of content and the amount of websites/web pages is increasing.
What, then, would that mean about the value of their complements?
Technical SEO. The sheer amount of pages and unbridled creation often creates an externality in the form of technical debt and disorganization. If a company wants to run AirOps and publish 20,000 pages, someone needs to either filter that up front or clean it up and organize it on the tail-end. Same goes for pages. We can also assume (and we already see) that LLMs present their own new layer of technical SEO (technical AEO) challenges, especially with agent-ready websites.
Distribution. Content supply increases, demand remains more or less stable. There’s still a finite quantity of attention or pixel space on a LinkedIn feed or AI answer. An emergent value, then, is clearly distribution (broadly speaking). This could include earned media and promotion, but also leveraging influential people and surface areas to drive attention for assets.
Judgment and taste. When anyone can produce a passable blog post, the ability to decide which blog post to produce and which to kill becomes a scarce skill. Strategy, editorial sensibility, knowing when to say no. This is the Substack lesson: readers don’t pay for information, they pay for a mind they trust to filter and interpret information.
Original data. AI can synthesize existing information instantly, but it needs something to synthesize. On the user side, the supply of noise is high, which means the value of signal increases. The answer to the question, “what the hell is actually going on?” First-party research, proprietary datasets, novel findings are the raw materials that can’t be copied and therefore can’t be commoditized, but can be fuel for distribution, repurposing, and a unique market point of view, all of which can create flywheels. When everyone’s drawing from the same well, the company with its own well has a strong advantage.
Relationships and trust. The matchmaker lesson, right? When AI can write the outreach email, the cold DM, the partnership pitch, what matters is whether the recipient trusts the sender. Brand relationships, creator partnerships, and community credibility are human-to-human signals that resist automation and are perhaps even antifragile in the face of a dead internet.
Experience and embodiment. The concert ticket lesson, yeah? In-person events, intimate communities, physical products, IRL encounters. Anything that requires presence. We wrote about this recently in an essay on costly signaling and unreasonable hospitality. Don’t be surprised if you’re a CMO in NYC and you eventually get an invite to a Mets game from me (or at least a nice dinner).
Speed of application. Kevin Kelly identified “immediacy” as a generative, something that gains value when copies are free. We look at speed as a structural advantage, especially when set to the beat of experimentation and learning.
Systems thinking and risk management. In a changing environment, it helps not only to have an eye on a unit, but on the system itself. Not only thinking about how AI changes how you write a brief, but how it changes the system of content production for the sake of revenue growth when competitors are also changing their work processes. Not only thinking about how to structure a page for citation, but identifying risk thresholds and engineering systems to capture upside without capsizing the boat.
The Strategic Inversion
Doug Shapiro framed the strategic imperative well:
“The biggest beneficiaries of technological change are those who can anticipate which resources will become abundant and which will become scarce and are able to squander the abundant resource to corner the scarce one.”
“Squander the abundant resource” is a wild thing to really consider.
Where many marketers are focusing their attention – on increasing efficiency of production, of execution optimizations, of sprinkling AI on an outdated mode of customer discovery – is effectively the entry point, not the point at which value is accumulating.
The measurement layer needs to reflect this, too. Output metrics (posts published, keywords targeted, pages created) measure the commodity. The complements require different instruments: share of model, brand recall, relationship density, pipeline influenced. Harder to measure, which is exactly why they’re harder to fake, which is exactly why they’re more valuable.
The Starbucks Lesson
I’m now a coffee snob.
Shade-grown, organic, fairtrade, single-source beans hand-ground and perfectly steeped in a French Press.
But there’s an alternative universe where I’m drinking Folger’s from the coffee pot. And that alternative universe is likely one where Starbucks was never founded.
That’s the lore anyway. The theory goes that Starbucks, with its ubiquity, options, and increased quality from the previous standard, created the consumer palate for craft coffee in the US.
By making expensive lattes ubiquitous, it taught a hundred million Americans that coffee could be something you cared about, something you’d pay a premium for, something with origin and craft and ritual. Third-wave shops like Blue Bottle exist because Starbucks did the hard work of mass education first.
AI is doing the same thing for marketing. It’s educating the market about what’s possible at the median level. Personally, I am now more ambitious than ever about producing original research, high quality video, and design (and I’m hiring for each of those roles – money where my megaphone is).
A Note on the Scarcest Resource
Underlying this essay is a belief of mine that much of what we see and hear and consume is noise.
You quickly learn that when you go on vacation, stop checking email and LinkedIn, and read East of Eden.
Noise is increasing, and the rate at which it is increasing is also increasing.
Thus, we can apply the same model to noise and attention that we can to any of the above tactics.
Noise is cheap or free, so what is valuable? Your ability to direct your attention, maintain emotional and intellectual sovereignty, and be resilient against mimesis.
I wanted to watch a World Cup match recently, and my girlfriend and I tried to go to a very popular soccer bar in the city. It has been promoted by nearly every NYC Instagram account (and I’m not on TikTok, but I’m sure there as well). It’s trendy.
I’m sure it’s also a great bar. But simply put, we were two in a hoard of people who were swayed by the megaphone of social media and influencer marketing.
Value, here, ended up being ignoring what was popular and instead finding an underrated Mexican restaurant that was showing the games where we were able to find a seat, enjoy great food, and perhaps have a little JOMO because we weren’t pushed into a sweaty, trendy spot like little sardines.
My prediction is you’ll increasingly see a premium on your ability to own your thoughts, attention, and focus in a world that is increasingly trying to take those away from you. Perhaps the biggest reallocation of value isn’t just finding the things that are valuable, but the ability to find the things that are valuable. This requires a fundamental shift in how you consume and filter information.
Just a thought. Carry on.
Want more insights like this? Subscribe to Field Notes
