It became so ubiquitous I assumed it was just how things had always been — that getting the optimal output out of every decision was the human condition. There’s an appeal to the fantasy of the right answer, which is what drew me to math and engineering in the first place. Later, it drew me to the psychology of choice and research by Daniel Kahneman, Amos Tversky, and Richard Thaler. Reading their work inspired me to pursue a PhD in judgment and decision-making, with the hope that I’d learn to make optimal decisions with ease.

I came out the other side with a different conclusion: difficulty in decision-making is the human condition. And yes, we want to ‘maximize’ what we can out of life (I’ll get to the psychology), but optimization is not the only strategy — just the one I and many others have absorbed as the default. It’s an old concept in math, but it’s only in the last hundred years or so that we’ve started applying it to computer science and then to human lives. Somewhere along the way, optimizing your health, sleep, or diet became its own form of self-care — one aimed, notably, at people who never thought of themselves as “self-care” people to begin with.

This is the first article in a series tracing where the notion of optimization comes from, why it’s so popular right now, and when (and for whom) it can help or lead astray.

Things We Optimize

Optimization has permeated all facets of modern society and culture. We optimize our morning routine, our looks (maxxing), daily purchases, and definitely our health. Equinox is rolling out a $40,000/year fitness and wellness program literally called “Optimize” and we use biometrics tracking and fitness wearables to get even more out of every activity.

We optimize our time with productivity tools (and more recently AI agents) to make more money and get more done. Popular self-help author, Oliver Burkeman writes, “For years now, we’ve been deluged with advice on living the fully optimized life, in books with titles such as Extreme Productivity, and the 4 Hour Workweek and Smarter Faster Better, plus websites full of ‘life hacks.’” The idea of “hacking” our lives or bodies seems almost commonplace now. Relatedly, Mark Manson talks about how “Our culture today is obsessively focused on unrealistically positive expectations.”

None of this is new to point out, and people like Burkeman, Manson, and many others are tiring of the prevalence of optimization. The Global Wellness Institute has declared an “over-optimization backlash.” Reddit threads show people turning away from optimization, writers create satire and mock the turning yourself into a “robot” approach to health. “Optimization as neurosis” has become a widespread criticism of the self-help and self-care movements. But most of that criticism treats optimization as self-evidently absurd, without asking what it actually is and where it comes from.

The Origins of Optimization

Merriam Webster defines optimization as “an act, process, or methodology of making something (such as a design, system, or decision) as fully perfect, functional, or effective as possible.” Optimization is a concept originally from math and in that context has three parts:

• An objective function — the thing you’re trying to maximize or minimize (cost, time, distance)

• Decision variables — the choices you make in pursuit of the objective (the levels of variables other than your objective)

• Constraints — the limits on those choices (budget, time, resources)

To optimize, you have to translate the real world into equations and metrics, and treat it as a closed system with a single best solution waiting to be found. This is the logic behind its use in operations research, engineering, manufacturing, economics, and computer science. The more variables, metrics, and constraints in play, the harder that solution is to find.

The Traveling Salesman Problem

One of the most famous optimization problems is the traveling salesman problem (TSP). The problem is one that we face every day: given a list of destinations, find the shortest possible route (objective) that visits each one exactly once (constraint). The decisions then become which city to go to next.

When trying to solve this type of problem, mathematicians can select from a variety of methods and heuristics. An exact solution can be found through brute force, where you try all permutations. This is possible on simpler problems, but for more complex problems heuristics like “nearest neighbor” (branching out to the closest next city one at a time) can approximate a pretty good solution much faster.

A real life TSP for visiting all 24,978 cities in Sweden was solved in 2004. This was the record for the longest exact solution until an 85,900 point problem for a Bell Labs computer chip was solved, taking over 136 CPU-years (one single computer core running constantly at full power for 136 years).

Optimization involves finding increasingly small gains to converge on some theoretical limit. TSP optimization made news when researchers beat a theoretical limit by 0.2 billionth of a trillionth of a trillionth of a percent! Fortunately, researchers in the space seem good humored about these small gains. Christos Papadimitriou said of the TSP: “[It] isn’t a problem, it’s an addiction.”

New methods for improving optimization are important for computing efficiency and are used in everyday applications like delivery tracking, ride-sharing, and mapping tools to find the shortest path visiting multiple stops.

AI and LLMs (chatbots like ChatGPT and Claude) are also the outputs of many iterations of optimization. The models are trained on huge databases (i.e. the internet) to mimic the patterns in language by minimizing the loss function that quantifies the difference between what the model would predict the next word is and what the actual text says. The computing required dwarfs the amount needed to solve the Bell Labs TSP. GPT-4’s training run alone is estimated to have used roughly 25,000 GPUs running for over three months straight — about 6,500 GPU-years, or roughly 48 times the entire computational effort behind Concorde’s Bell Labs TSP record.

Complexity increases when going from optimized routes between cities to predicting language, and takes another jump when applied to human behavior. Millions of people are using AI tools to help optimize their own lives. The computing cost to use one month of ChatGPT queries at ~0.3 Wh/query and roughly 2.5 billion prompts a day is ~22.5 GWh — about 189x the energy of the entire Concorde feat, every month, from one chatbot. All of this energy usage begs the question, is human optimization worth the cost of compute?

The Human Optimization Problem

Bryan Johnson is the quintessential example of optimization of the human body and health. He is the world’s most measured human and on a quest to defy aging and pioneer new practices in longevity. He epitomizes the biohacking movement as he incorporates staples of computational optimization: measurement, analysis, and fine-tuning of protocols to systematically improve the mechanical outputs of his body (a motivational mindset or philosophy I call viewing the body like a “machine”).

His objective function is “Don’t Die,” and he is unambiguously optimizing for the length of his life by making small improvements and chipping away towards some theoretical limit (146-194 years of maximum human lifespan by one recent model). His decision variables relate to diet (vegan only, supplements, pills, no caffeine), sleep (timing, length, quality), and physical interventions (exercise, red-light therapy, experimental treatments like plasma exchange, diagnostic testing). Importantly he has far fewer constraints than the typical person, with the time to focus solely on his health and money to spend millions of dollars a year to hire a staff of medical professionals.

He’s hoping that his self-optimization experiment (N = 1) will yield health benefits for others too. And in general, we are making progress on extending lifespan in the U.S. and across the globe, as life expectancy continues to rise.

Despite this progress though, perceptions of health in the U.S. and globally are going in the opposite direction. Data from the World Values Survey (graphed by Claude) shows a slight decreasing trend in perceived health over the last 30 years, where self-rated health perceptions have dropped below the line from “Good” toward “Fair.”

There’s also a trend of declining perceptions of healthcare quality and coverage in the U.S. since 2010. And happiness is on the decline too (especially among young adults).

So why the discrepancy between increasing life span and decreasing perceived health, care, and happiness? Some thought leaders like Peter Attia might point out the difference between life span (how long you live) and health span (how many healthy years you live), but I think that’s missing the deeper point.

I would argue that for many people, optimizing for health or life span is trying to solve a very complex problem with the wrong methods.

Optimization as a Method, not an End Goal

When people employ optimization for a traveling salesman or operations research problem, they’re using it as a method to try to achieve a better outcome — not as an end in and of itself. Product designers use a similar idea in the jobs to be done framework, which reminds them what the customer actually needs solved. A customer pays for the hole, not the Dewalt power drill; for the ability to connect remotely with people, not for the Zoom meeting itself.

Within general health and wellness, let’s say that the customer need or objective function is to feel better. Part of why we reach for optimization at all is that health has generally improved, which raises the stakes on the objective function itself. We’re no longer aiming for “not sick.” We’re aiming for high performance, longevity, and optimized everything. Expectations have climbed faster than most people’s ability to meet them, which may explain why perceived health has fallen even as longevity keeps rising.

Another problem is that health is a complicated product of your unique biology, environment, behaviors, and choices. And the health and wellness marketplaces continue to show huge growth. There are way more decisions to make about what to buy and do in pursuit of wellness now.

And there are also more constraints working against making any of them well — rising healthcare costs, longer work hours, more stress.

Therefore the perfect health choice or routine is only a fantasy. Instead we must all use heuristics to make difficult choices and approximate a solution. For a given person, your morning routine, sleep, fitness, or health can be “optimized,” but it will never be truly “optimal.” Optimization, in that sense, is really just a method for achieving better outcomes: better for some types of problems, and worse for others.

Limits of Optimization

Some of the limitations of optimization are rooted in our own limited knowledge and technology — what Bryan Johnson is working on. But here’s a thought experiment. Suppose we solved every practical obstacle preventing human optimization: perfect data, every variable mapped, unlimited processing power. Would optimizing a human life actually work?

What it’s good for:

• Well-defined problems, where the components are clear and there can be consensus on which direction is better. This includes objective metrics with a direct, immediate feedback loop like heart rate, VO2 max, or sleep duration.

• Small-scale decisions with a single solution based on a simplified set of constraints and decision variables, like a toaster or USB cable. Or maybe you just don’t care about the outcome enough and can settle for an imperfect choice.

• Stable problems that don’t change while you’re solving them. A marathon is the same distance today as it was when it was standardized in 1921. If the distance kept increasing, you’d have no way to know if you were making progress.

• Small, incremental gains on an existing system — refining what’s already there rather than throwing it out and starting over. One of my PhD professors drew a distinction between two approaches to solving hard problems: engineering and physics. Engineering optimizes within a known set of constraints — better materials, tighter tolerances, a faster chip.

What it’s not good for:

• Poorly defined problems: There’s no agreed upon definition of “health” or “wellness,” a recurring topic in my research. These are nebulous concepts, with many interpretations of the outcome and goal.

• Incompatible variables and multiple objectives: a human life includes health, relationships, money, time, meaning, pleasure. These trade off against each other in ways that don’t net out to a single number. Even if we could reliably measure them all, people would be unable to agree on the conversion rate between them; therefore representing them as equations would be impossible. Some variables are also incompatible, like money and human life, or extrinsic and intrinsic motivators (which tend to get crowded out when you begin to quantify).

• Relatedly, subjective outputs like happiness or productivity are impossible to optimize for because the target is constantly moving. Researchers have found that being too concerned with happiness introduces negative judgment which makes it even more unattainable. Manson writes that happiness “is not a solvable equation.” Burkeman describes productivity as a “trap” because getting more efficient at something only gives you more bandwidth to take on more.

• Problems that shift while you’re solving for them — the goal, the constraints, or the inputs change and the problem you’re optimizing for isn’t the problem you actually have anymore. The popular business phrase “what gets measured gets managed” is actually a cautionary message that measurement can mislead organizations to focus on the wrong things. Said another way: “When a measure becomes a target, it ceases to be a good measure” (Goodhart’s Law).

Which brings us back to the man determined not to die. Bryan Johnson and others in the longevity space are doing the work of engineering: combining the best tools, science, and technology to make incremental gains. It works, if length of life is your objective function, you have the resources to chase it, and are willing to sacrifice other experiential aspects of a good life.

Most people don't, and more importantly, most of what makes a life good isn't reducible to years lived anyway. It's a physics problem — one where the answer isn't a better protocol, but a willingness to question the premise itself.

Conclusion

It’s easy to judge Bryan Johnson and suggest that what he’s doing is weird. He knows this, and he doesn’t care. It’s an extreme expression of where the health and wellness space is now, but it isn’t so far off in terms of the popular approach to self-care through optimization. Plenty of people share this approach, and they shouldn’t be judged either. What works for them may not work for you, and it’s on each of us to educate ourselves and find our own path to the good life.

Someone who’s been writing about consumer wellness trends, Julia Klim, asks a good question: “What are we optimizing for?” I’d push it further — why are we so driven to optimize in the first place, and might we be better off, at least for some things, not optimizing at all?

I still haven’t found a way to make every decision with ease. The PhD didn’t deliver what I was hoping for on that front. What it gave me instead was a better understanding of the different approaches people take for health and wellness, and a lens for evaluating them.

If optimized gains are mostly small and incremental, why are we investing so heavily in them — individually and as an industry? Why are we “hacking” our lives and our bodies and what might other approaches look like?

That’s what the rest of this series sets out to answer — starting with how optimization became one of the dominant forms of self-care in modern life, and how many benefit from the approach while others risk burnout. Pursuing a healthy life only continues to increase in complexity. Optimize if you want, but make sure you've got the computing power to spare.

Subscribe now

Share

Leave a comment

I find this surprising for two reasons. I’d always assumed human connection was the primary driver behind effective therapy — CBT, psychodynamic, and ACT differ in approach, but all of them involve a human delivering care, and a lot of emphasis gets put on the therapeutic relationship itself. A loneliness epidemic would seemingly suggest a greater need for human companionship than ever, not less. And yet consumers keep gravitating toward AI companions and chatbots, as if mental health care needs to build beyond human interaction to actually reach people.

My early experiences with AI chatbots didn’t help my skepticism. They were full of hallucinations and inaccuracies, more frustrating than impressive on anything deeper than a glorified Google search, and there were real horror stories of chatbots contributing to user harm.

The models are getting better, and the risks are still there (more on that below), but early examples are starting to show how effective a well-designed AI-augmented product can be for mental health.

When Users Build the Product Before the Industry Does

It’s easy to be cynical about AI right now. It’s been crammed into every far-flung product category you can think of — golf clubs, whisky, pet translation, toothbrushes. It feels performative, like companies need to demonstrate they’re modernizing, and it can become a turn-off. I’d rather drink whisky made from a legacy of Scottish artisans than one distilled from a model’s guess at my preferences — reducing my taste to 1s and 0s neglects something uniquely human about the process. I’ve been brushing my teeth the same way for years and been just fine.

It’s also backwards as product design. Instead of starting with a user need and building toward it, most of these launches start with a technology and go looking for a problem to attach it to. That’s research, not design. Relatedly, I’ve always felt it is weird how pharmaceutical drug development usually starts with discovering a new physiological effect and then afterwards finding a need or market to target.

But mental health AI doesn’t fit that pattern. People are already going to AI (in the form of chatbots or large-language models, LLMs) to talk through their problems, on their own, without any company designing for it. That’s an prototypical example of lead user innovation — where users build or adopt their own solution to a problem well ahead of the marketplace catching up. This can be a positive sign of innovation if companies can observe and build on the user momentum, just as they did with the origin of mountain biking.

Mountain Biking as Lead User Innovation

Mountain biking wasn’t invented by a bike company. In the 1970s, a group of riders in Marin County started taking modified cruiser bikes down fire roads for informal downhill races. Riders like Gary Fisher and Joe Breeze kept modifying their own bikes with fatter tires, better brakes, and different gearing because nothing on the market was built for what they were actually doing. It took years before manufacturers caught up and started building purpose-made mountain bikes. Now the industry is approaching nearly $10B in size.

Users taking things into their own hands told manufacturers there was a real, underserved need. This was probably a pretty reckless and dangerous thing to be doing, and once companies got involved they were able to make the bikes even safer and better.

That’s the position mental health AI is in right now. People adopting general-purpose chatbots for therapy-like conversations is a strong signal that there’s a real, unmet need. But a general LLM is a road bike on a fire trail — it can get you down the hill, but it wasn’t built for it, and the parts that make purpose-built off-road machines actually safe (suspension, frame strength, brakes) is still missing for mental health.

The Driving Marketplace Gaps

So what’s actually generating this lead-user behavior? We know there are structural problems with traditional mental health care given that only about half of people get the mental health support they need. But how does AI stack up against each?

1. There’s a shortage of therapists. The bureau of health workforce estimates that 40% of people in the U.S. live in an area with a shortage of behavioral health providers. The average wait time to see a behavioral health specialist is 6 weeks.

AI resources, on the other hand, are everywhere. The list of chatbots keeps growing and half of people in the U.S. have now used an LLM and major AI players like OpenAI, Anthropic, Perplexity, Google, Amazon, and Microsoft are all beginning to build health-specific models (although I’m not sure if this includes mental health or is just physical).

2. Therapy has limited hours. Most people who see a therapist do so for about an hour, once a week or a few times a month — roughly 0.25-0.5% of total hours in a week. Crises don’t wait for that slot; they show up whenever they show up, and that hour gets used to process what already happened. AI is available 24/7, which is a core marketing pitch for newer mental health AI companies like Dawn. It can also be easier to disclose something difficult to a chatbot that already “knows” everything else about you than to open up to a new person.

3. Therapy is expensive. The average cost runs $100-250 an hour. Roughly ⅓ of therapists don’t accept insurance, and the cash-pay model seems to be growing, with big players like BetterHelp and Talkspace leading the charge. AI, by contrast, is cheap — the cost of compute, and most LLMs have a free base model that’s currently being leveraged for this use case.

4. Therapy speaks one language, and a lot of consumers speak another. I’ve written before about how health and wellness consumers vary in their approaches to taking care of themselves – treating the body (and mind) like a machine to optimize, a temple to nurture, or some combination of the two. When I applied that framework to the mental health space I found it overwhelmingly oriented toward temple language — balance, harmony, self-compassion — while performance and optimization framing was almost entirely absent. That’s a real mismatch, given how many consumers actually want the latter.

Some of this may be a byproduct of who goes into the field: colleagues in sociology point to psychology’s strong liberal skew, both in who becomes a provider and who seeks one out, with people gravitating toward like-minded providers. There’s likely more stigma toward mental health care among conservatives (a finding we’ve replicated in research with UNH professor Justin Pomerance), but that doesn’t mean the demand isn’t there — it means it’s going unmet. “Therapy speak” (“triggered”, “trauma”, “boundaries”) has also seeped so far into everyday language that it’s started to grate, and a real opening exists for new framing.

An AI tool doesn’t have this problem, because it doesn’t have a fixed identity or ideology to project. It can adapt its language style to you — probably more easily than a human provider can adapt theirs to a client. Using a tool better matched to your needs can reduce stigma and improve adoption and retention.

Beyond mental health specifically, I’ve written before about how people feel overwhelmed by the sheer amount of tracking baked into modern wearables and apps. What excites me about AI is the possibility of a more passive read on how you’re doing. I’ve started experimenting with feeding my own journal entries to an AI tool to see what it can infer over time, without the friction of active tracking.

Does Any of This Show Up in Outcomes?

The early research is directionally positive but thin — small samples, not yet replicated, and mostly published in lower-tier journals rather than flagship outlets like Nature and Science. A few consistent signals so far though: AI companions can reduce loneliness, self-guided tools like flourish have shown modest reductions in depression and anxiety symptoms among college students, and AI-augmented coaching apps like Bloom show higher engagement than non-AI versions. One fully AI-delivered therapy tool, Therabot, showed real clinical improvement, though it hasn’t yet been tested head-to-head against human therapists. These are worth watching.

What about the risks?

The risks are ever present, as it’s still a “road bike on a fire trail.” Kids are using AI chatbots for mental health support, but more often than not without telling their parents — and general-purpose tools aren’t built with the guardrails a clinical use case demands. This can cause them to respond inappropriately, and there have already been extreme cases of chatbots responding to a mental health crisis in ways that made things worse, including at least one documented suicide. Even where models are good at detecting suicide risk, they’re not necessarily good at intervening on it. And when researchers have compared chatbots responses to therapists’, chatbots tend to over-validate and jump to generic advice, without the probing and tailoring a real clinical conversation requires.

It’s not surprising, then, that 85% of Americans surveyed report some hesitation about AI in mental health care. The top reasons: it can’t really understand you, people want to talk things through with someone, and they don’t trust it to handle an actual crisis. Younger generations are considerably more open to it, and there’s more comfort with AI as a supplement to therapy than as a replacement for it. Interestingly, privacy doesn’t come up as a top concern yet — though it’s likely to become one as people better understand how these tools use their data.

Providers have their own hesitations: a belief that AI can’t really help, fear of being replaced, uncertainty about how to use it well. But there’s a case for AI helping providers too — better client insight through pattern-spotting across sessions, more engaged clients between visits, and more efficient documentation and note-taking.

Closing Thoughts

Riders didn’t wait for Trek or Specialized to build them a proper trail bike — they built their own out of what was on hand, and the market eventually caught up with something engineered for the terrain. That’s roughly where mental health AI is right now: real, well-documented gaps in cost, access, cadence, and tone are driving people to a general-purpose tool that wasn’t built for the job, and doing it well enough to prove there’s a real market underneath it.

Only about half of people who need mental health care actually get it. Roughly half of Americans have now used an LLM. But do two halves make a whole?

The unresolved question isn’t whether AI belongs in mental health – the lead users already engaging with general LLM’s suggests that there’s something important to follow here. It’s whether the industry does the science and engineering work fast enough: purpose-built models, humans in the loop as guardrails, and real design attention to the moments where a general chatbot’s generic reassurance isn’t enough. If we get that part right, this looks less like AI in a toothbrush and more like the rush of riding a top of the line Trek mountain bike down a Marin fire trail.

Subscribe now

I'd love to hear how you're actually using AI for this — whether it's you, a friend, or just something you've noticed in the wild. And if you work in mental health or product, I'm curious about your vision for these tools.

Share

Leave a comment

I want to suggest a different approach, borrowed from nutritional science and using one of my favorite rhetorical devices: the metaphor (no machines or temples in this article though).

In virtually every American classroom, kids learn about the basic food groups. The message is generally that a healthy diet requires balance: a variety of the right inputs, in the right proportions, consistently over time. We have no equivalent framework for mental health, and children are not educated about good mental hygiene. Following up on a post about the stream of consciousness and importance of mind-wandering, I want to sketch an idea out here, inspired by conversations with UBC cognitive neuroscience professor Kalina Christoff and her graduate student Doug Forrest.

What we’re feeding our minds matter and we should be thinking more about where we’re getting our mental “food” from and how it is processed. We need to revise our mental dietary guidelines.

A brief history of processed food

One of Michael Pollan’s ideas from his book Cooked that has really stuck with me is how much cooking with fire did for human nutrition and evolution. The heat breaks down the food and provides the front end digestive work, allowing us access to more calories and nutrients with less effort.

Modern food processing and mass manufacturing has taken that to an extreme though, fueled by clever marketing campaigns and profit-maximizing companies. The U.S. started canning foods around the civil war to help feed soldiers. After World War II, companies advertised shelf-stable and prepared foods and their convenience made them highly popular, establishing a new comfort level with consuming processed foods. At one point in the 1980s, tobacco companies like Philip Morris bought up snacks and breakfast cereal brands (e.g., Fruity Pebbles, Oreos, Kraft Mac & Cheese, Jell-O, Kool-Aid) and leveraged the same tactics used to sell cigarettes. “Ultra-processed” foods deliver sugar, salt, and engineered flavor with almost no effort required by the consumer but unwrapping the packaging.

We now understand better how these ultra-processed foods negatively impact consumer metabolic and overall health. The 2025–2030 Dietary Guidelines for Americans represent a notable course correction on processed foods. After decades of guidance organized around fat content and calorie counting, the new framework focuses on food quality and processing level. The central message is to eat real food, in reasonable variety, and limit the things that have been engineered to be more palatable than nutritious. The food pyramid has been revised and this latest version reflects an emerging consensus that the composition of the diet matters more than any single nutrient.

The same logic, I’d argue, applies to our diets of mental experience.

Three sources for the mental diet

In a previous post I wrote about the wonderful stream of consciousness and mind-wandering — the default mental state we spend roughly half our waking lives in. The argument there was that this spontaneous inner life is more meaningful and more influential on our behavior than we tend to assume. Here I want to take that a step further and ask: what are we feeding our minds? And how does mind-wandering fit into a healthy mental diet?

The mind is always consuming information. Even when you’re doing “nothing,” your brain isn’t offline but processing, integrating, and constructing the internal narrative of your life. What varies is the source of the input, and how much pre-processing has already been done before the content reaches your conscious attention. Mind-wandering is the mind’s natural propensity to wander to a new topic of thought. And if it’s not fed external stimuli to attend to, it will by default turn towards internally generated content (imaginations, daydreams, fantasies, etc.).

It’s useful to think about the sources for the stream of consciousness as being fed by three different smaller streams. The three streams — exteroceptive, interoceptive, imaginative — correspond roughly to different levels of mental processing by either a manufacturer or by yourself as the producer, in the same way that food differs in how much it was prepared, modified, and transported before it reaches you.

The exteroceptive stream

On one end is the exteroceptive stream: stimulation coming from the outside world through your senses. This ultra-processed exteroceptive content that is algorithmically-curated to command attention is the junk or fast food of the mental diet. It’s good on the go and can give you a quick jolt of calories and nutrients. Things like scrolling TikTok, watching a Netflix series, or commenting on a friend’s Instagram post contain important basic experiential elements like social, fantasy, positive emotions, but in highly concentrated form.

Someone else (usually a company or business) is doing most of the production work to engineer content for maximum palatability and rapid consumption (and fast metabolizing). Because it’s mass produced, the makeup of the food is also the same wherever you go.

Higher-quality exteroceptive content like a good book, a documentary, or a live performance is more like fast casual dining: Chipotle, Sweetgreen. Still produced by someone else, but with more involvement from you in choosing among the fresher, purer ingredients, which makes each experience a little different.

The interoceptive and imaginative streams

In the interoceptive stream you’re attending to signals from inside your own body. Exercise, therapy, journaling, yoga, sitting with an emotion, counting breaths (a proxy for mindfulness), or engaging in conversation or consuming art that requires more work for you to co-produce the experience.

And on the other end is the imaginative stream involved in experiences like daydreaming, creative hobbies, prayer, or deep meditation. The mind generates its own content largely from scratch.

Interoceptive and imaginative experiences ask more from the consumer, and are more effortful and slower to process: your local farm to table restaurant, or a meal produced from your own garden. In imaginative experiences in particular you are not only the receiver but also the producer. These experiences can be highly enriching and meaningful because they are loaded with sensory experience, emotions, and exploration of alternative worlds and realities — what we typically associate with mind-wandering and the inner experience.

These streams aren’t perfectly discrete categories. For example, a great film can trigger imaginative reflection; a yoga class involves both a teacher’s cues and interoceptive attention. But they are experientially distinct and the ratio in which we consume them relates to a healthy mental diet.

Mental diets skewed towards exteroceptive

The average American adult spends roughly seven hours per day looking at screens (the ultimate source of exteroceptive stimulation), which is more than 40% of all waking hours and an increase in 8.4% since 2013. Social media alone accounts for about 2.5 hours daily (approximately a ~25% increase since 2015), a figure that climbs past three hours for adults aged 18–24. Using Claude to visualize the time spent on social media versus socializing and communicating (including phone calls, texts, email, face-to-face contact — examples of more interoceptive content) shows a widening gap.

Creative hobbies like arts and crafts, writing, playing music which require more active processing of the interoceptive and imaginative streams pale in time spent relative to watching television.

Mind-body practices like meditation, yoga, and guided imagery/progressive relaxation are up over time.

The trends suggest that the composition of the average American mental diet is, by a wide margin, ultra-processed. Modern consumers trend towards filling their free time with mental stimulation. Research found that people would rather administer mild electric shocks to themselves than sit alone with their thoughts for fifteen minutes. The impulse to fill every quiet moment is understandable, but it comes at a cost. Default mode network (DMN) researchers suggests this state of mind is precisely when a lot of the integrative, generative cognitive work happens. When the DMN is perpetually suppressed by external stimulation, important thoughts and experiences go unprocessed.

Limiting an ultra-processed mental diet

The skew toward exteroceptive consumption isn’t accidental, and it isn’t a failure of individual willpower. The attention economy is very good at producing highly palatable mental content. Every major social platform and streaming service has been engineered to keep you consuming. The highly engaging content is available everywhere, all the time.

Parents have remarked on how much more visually intense children’s programming is now than when they were kids: the high-contrast colors and rapid cuts of something like Cocomelon compared to older Sesame Street. It’s the same across television and streaming: try watching an early-2000s drama from start to finish compared to something made for the current streaming environment. Watching Lost one episode a week on ABC and then listening to the recap/analysis on the radio is a much different experience than binge-watching it on Hulu.

Meanwhile the less processed alternatives (reading, creative writing, sitting quietly) require more sustained effort and attention. Maryanne Wolf, who studies the neuroscience of reading, has written about how even the act of reading on a screen versus a physical book changes how deeply we process what we’ve encountered.

The physical dietary guidelines understood how structurally engrained our relationship to food is, which is why food policy pertains to more than food labeling. It includes education, subsidies, and cafeteria options in schools, hospitals, etc. A mental diet conversation needs the same structural lens to consider how we feed our minds.

What a balanced diet might look like

This framework isn’t an argument for eliminating exteroceptive consumption, or for a romanticized return to pre-smartphone life (although I’ve at times considered how much lighter I might feel with a “dumb” phone). Instead, I argue for more balance of mental processing: how much you’re producing versus receiving pre-processed content. In the same way that we should be limiting ultra-processed food in our physical diets.

Two hours of reading and two hours of social media scrolling are not equivalent experiences even if the content of both is interesting, because they make different demands on your brain to co-produce the experience. One tends to draw from the imaginative and interoceptive streams; the other tends to suppress them. Noticing which stream you’re actually in and where the content is coming from will signal how much it is processed before it reaches you.

We would all probably benefit from reducing the pre-processing in a little more of our leisure time. Small shifts in medium can move you in the right direction — a phone call instead of a scroll, a physical book instead of an audiobook, a live show instead of a stream — without requiring any dramatic change in how much you consume overall.

William James, whose phrase the “wonderful stream of consciousness” kicked off my last post on the topic, was writing at the height of the second industrial revolution. The threats to inner life then were physical — monotony, exhaustion, a Protestant suspicion of leisure itself.

Amid what some are calling a fourth industrial revolution, we face a different version of the same problem. Tools like AI promise to make us more efficient and free up more time to do other things. The attention economy doesn’t demand your labor but your attention in your free time. It offers you something you want but engineered to be so engaging that it’s hard to stop and easy to overdo it. The stream gets fed with chemical runoff from mass production. What you put in shapes what it gives back.

Thanks for reading. I’m always curious to hear your thoughts and ideas about modern consumer health and wellness.

Leave a comment

Subscribe now

These frameworks are useful for understanding more than just fitness routines though. The philosophies apply broadly to how people think about their health, wellness, and their lives more broadly. Curious about how the framework might represent a different area of the health and wellness marketplace, I sought to apply it to the mental health marketplace.

Using consumer survey data and a social media analysis of 50 mental health companies, I compared why people actually seek mental health support with how companies market it.

What I found was a totally different marketplace structure than physical fitness, with overwhelming emphasis on balance and harmony, while almost entirely ignoring performance and optimization. That gap is costing companies in terms of engagement and credibility, and dissuading consumers out of using good products.

Why People Seek Mental Health Support

Meditation and therapy have become the primary tools for consumer mental health and wellness. There are thousands of meditation apps, including Headspace, Calm, Insight Timer, Simple Habits, and Happier. They offer a range of benefits, including helping reduce anxiety, deal with depression, and improve sleep. There is ample evidence to show that meditation and therapy are generally effective in improving mental health.

In a previous study, we surveyed 201 people about why they do or don’t meditate. They wrote in their responses and also ranked the top 6 reasons. The most common reasons to meditate were:

These fell into two broad clusters:

Mental performance reasons: improving focus and concentration, enhancing productivity and work performance, sleeping better, and improving healthy habits

Mental balance reasons: managing stress and anxiety to find emotional stability, increasing self-awareness, and being more present in daily life

These map nicely onto the machine/temple framework, with people using meditation for more intrinsic reasons (i.e., done for more internal reasons like self-awareness or presence) or instrumental ones (i.e., done for a specific reason to boost some downstream outcome like sleep, productivity, or exercise).

Neither is necessarily better*; they reflect different motivations and needs. And in a well-functioning marketplace, product features and marketing would speak to both.

A quick note that the body-as-a-machine or temple metaphors don’t perfectly translate to the discussion of the mind and care. The mind has been referred to as a machine (such as in Lakoff and Johnson’s 1980 book about metaphors) and the study of cognitive science and neuroscience is largely based on the metaphor of the brain as a computer, which emphasizes cold, rational, rule-based thought. Michael Pollan talks about this in his new book A World Appears. I haven’t come across the mind referred to as a temple. But I believe that the underlying meanings and orientations still apply.

*My data shows that mental, social, and emotional reasons were directionally positive predictors of self-reported meditation interest, frequency, and experience, while cognitive reasons were significant negative predictors. I think this is partially an artifact of the ranking style question and more work should be done to test the relative effect of different reasons.

How The Marketplace Represents Mental Health

So how does the marketplace tend to appeal to consumers? Does it leverage performance or balance, both, or neither?

To analyze the mental health marketplace (I’ll refer to it as mental health, but know that I also mean wellness) I looked at the social media marketing (Instagram profiles and Meta ads) for 50 companies. I included companies from different sub-industries including AI chatbots, meditation and mindfulness, general talk therapy, therapy specialties, and workplace wellness.

I used my LLM model to generate temple and machine ratings for each company.

Scores primarily fell within the high-temple, low-machine quadrant, with only 2 scoring higher than the midpoint in machine, graphed below. As with the fitness marketplace, temple and machine scores were negatively correlated (r = -.53), although the effect wasn’t quite as large (r = -.89). Still, companies are trading off between temple versus machine-oriented strategies (and mostly choosing temple).

Across nearly all sub-categories, the emotional and experiential benefits of mental health products were emphasized: calm, balance, presence, self-compassion (temple-oriented), while cognitive, functional, and performance (machine-oriented) benefits were passed over. The lone exception was workplace mental health, where companies were more likely to orient towards performance and productivity.

The Missing Machine

The extent to which the marketplace messages mental health for balance and harmony, as opposed to performance and optimization, was greater than I expected. Only 2 out of the 50 companies — BetterUp (“powering performance-ready workforces in the AI era”) and Kick Health (“performance medicine for restful sleep & confident presentations”) — lean in to the machine. And you could reasonably argue that neither is a mental health company in the traditional sense.

This is a significant missed opportunity because the consumer base is broader than the messaging suggests. My research shows that people who view the body as a machine still care about mental health. They're just motivated differently. They want to know that therapy or meditation will make them sharper, more resilient, more productive (the software needs to be great too in order to run the hardware). The marketplace is missing the messaging to reach them.

Leveraging Machine Messaging

Performance framing works in adjacent spaces. Consider Andrew Huberman’s popular health podcast, where he often refers to the body, and even the mind, in terms related to optimization and performance. In a recent episode with the prominent Mindfulness researcher, Dr. Richie Davidson (“Science-Based Meditation Tools to Improve Your Brain & Health”), they contextualize the experience and benefits of meditation through language oriented toward the body-as-a-machine.

Huberman highlights the benefits for improving focus and helping with sleep, and skips over some of the more emotional, introspective, and existential benefits. Davidson does frame the benefits in a temple-manner at times, such as when he describes the feeling of “oneness” following from some meditation as a spiritual or transcendent experience (I’m a big fan of that part too).

They also compare how meditation can initially shake things up mentally and emotionally (e.g., a 1 week uptick of anxiety after you start) to how you might feel sore after beginning a new exercise routine (from the lactic acid building up). The implication is that a “no-pain no-gain” mindset that helps maximize performance in physical health should also be applied to mental health. You don’t see that in current mental health products or marketing.

The analogs between mental and physical health are interesting and potentially fruitful to connect with and motivate people that opt for the optimization and performance often found in fitness.

I also thought I’d find that AI mental health chatbots, which literally leverage machines to deliver care, would lean more into the machine mindset but they largely don't. In the physical fitness space, companies that used more equipment and technology leaned more into the machine orientation. Woebot, Headspace's Ebb feature, and similar products market themselves in almost entirely temple terms. A mismatch between technology and marketing may turn off the very consumers most naturally drawn to AI-assisted solutions.

Closing Thoughts

As a research collaborator and I have often said, “mental health has a marketing problem.” There’s a large body of evidence that demonstrates that mental health solutions like meditation and therapy are effective. There is a huge consumer population with established need. Yet people continue not to seek help, which means that the marketing is generally ineffective (I’ll probably talk more about this, including the overestimation of mental health stigma, in a later article).

While most research on mental health has gone toward understanding the benefits of tools like meditation and therapy, less is known about who these solutions work best for, how best to tailor the options to the individual, and how to solve an ongoing engagement issue. 95% of people that download an app aren’t using it within 30 days. It’s easy to download an app, but much harder to build the healthy habit.

Part of the engagement problem is a motivational mismatch. The mental health marketplace is lopsided in its attempts to appeal to consumers. Unlike the physical fitness marketplace, which offers balanced, albeit stratified options, mental health companies overemphasize balance and harmony, at the expense of performance and optimization.

There’s good reason to believe that what works in physical fitness will work for mental health. There’s a growing body of evidence including research linking mental and physical health and a trend of marketplace consolidation as part of the broadening of consumer health through the push for wellness. Peloton acquired the mindfulness app Breathwrk. Universal Health Services just bought the therapy company TalkSpace. Telehealth platform Teladoc owns the therapy app BetterHelp.

The infrastructure is being built, but the messaging hasn't caught up and neither has the product philosophy. In the fitness wearables space I found that most companies used either a general consumer (Samsung, Google, Apple Watch: connectivity, fashion, convenience), pure machine (WHOOP: strain, recovery, optimization) or pure temple marketing strategy (Bellabeat: stress regulation, sleep, hormonal health). The standout was Oura, which uniquely positions itself as high in both and has built a broad, loyal user base as a result. The mental health marketplace has no Oura yet.

It’s easy to forget that the digital mental health industry is still young and rapidly changing, with the oldest companies in my list being Teladoc (est. 2002) and Headspace (2010). We need new approaches that hit on what people actually care about. The “meets you where you are” messaging that has dominated mental health no longer rings true. Meeting people where they are might look more like pushing and challenging them— to build their mental fitness, not just balance their mental health.

Subscribe now

Leave a comment

Wearables are powerful tools for behavior change. They increase physical activity, improve sleep, reduce alcohol consumption, and in general make people more aware of their health. And yet, for many users, they have a surprisingly short lifespan.

Wearables don’t fail because people stop caring about their health. They fail because they rely too heavily on extrinsic motivation and don’t evolve as users’ motivation changes.

The novel insights about the body tend to dwindle and the extrinsic motivation of tracking can crowd out the intrinsic motivation and enjoyment of fitness. Over time, what once felt motivating starts to feel effortful—or even burdensome. Many ultimately opt out of tracking in favor of their own subjective monitoring and intuition. To understand why this happens, it helps to look at the psychology of motivation and how motivation changes over time.

Underlying Psychology

There’s not a lot of research on the long-term motivational dynamics of wearable fitness trackers. But at a high level, we know they work—at least initially.

Wearables usage is associated with increased physical activity and other positive behavior change through helping users form clear, trackable goals and raising bodily awareness through objective metrics like steps, mile time, strain, recovery, sleep scores, etc. Add in personal records, social comparison, and streaks, and these metrics become powerful extrinsic motivators.

An example of the motivating power of metrics tracking is from the Boston marathon. People often set goals for race times around targets like 3:30, 4, 4:30, or 5:00. You can see the big spikes in the number of finishers at those times because those goals serve as very effective extrinsic motivators. Or people set goals to simply finish the race.

But there’s a catch. The motivation underlying these types of goals is often short-lived. It might get people over the hump, but does not mean they’ll become long time marathon runners. This is because what once felt like an achievement becomes the new baseline, as reference points shift. To stay motivated, you have to keep pushing further—and eventually, that becomes unsustainable. I have quite a few friends that ran marathons this past year and some have told me they’ll never run again, let alone another marathon.

To sustain motivation in health and wellness, understanding psychology is important. Self-determination theory or SDT (Ryan and Deci 2000) distinguishes between extrinsic (driven by outcomes, metrics, rewards) and intrinsic motivation (driven by enjoyment, interest, and internal satisfaction). These exist along a spectrum of internalization, with differing levels of perceived locus of causality—that is, the extent to which people feel their behavior is self-directed versus externally driven—as well as different regulatory processes.

As examples of more intrinsically motivated running, people might run because it helps clear the mind, because it’s a beautiful day outside, or simply because they enjoy running. Obviously not everyone feels intrinsically motivated to run, which is why extrinsic motivators to spur or bolster motivation are so important.

Wearables are exceptionally good at generating extrinsic motivation. But intrinsic motivation is what sustains behavior over the long term. And as research by Etkin (2016) shows, there’s a “hidden cost” to tracking: extrinsic motivators can crowd out intrinsic ones.

So if wearables tend to lose engagement over time, the question becomes: how should they evolve as users’ motivation evolves? Part of the answer lies in recognizing that motivation itself follows a pattern.

Across users and segments, there is a consistent motivational lifecycle underlying wearable usage that looks something like this: onboarding → excitement → optimization → burnout → intuition and self-reliance.

At first, users rely heavily on external feedback. Over time, they internalize what they’ve learned and eventually, many no longer need the device in the same way—or at all.

Wearable Consumer Journeys

If you look at typical consumer journeys through the wearables space, they might look a lot like this LinkedIn post:

People might go from no wearable, to a mass market device like an Apple Watch (for convenience and connectivity), and then to a more specialized performance tracker like a Whoop (for tracking daily recovery) or a Garmin (for more specific running workouts and tracking, says Reddit users), only to go back to no wearable.

My own journey looks like: No watch → Apple Watch→ Fitbit → No watch → Apple Watch → Whoop → Oura → ???

What looks like a product journey is actually a motivational one. My philosophies of the body framework suggests that people differ in how they approach bodily care: as more performance-oriented (body-as-a-machine) and/or more holistic (body-as-a-temple). In a previous article I illustrated how four segments emerge: low machine/low temple (tech convenience and connectivity), high machine/low temple (high performance), high machine/high temple (balanced performance and holistic wellness), and low machine/high temple (holistic wellness).

Over time, I’ve shifted in terms of my beliefs and approaches to care of my body for performance versus balance, and in what wearables I’ve used to help reinforce and motivate that care. Illustrated across the segments, my journey has looked like this:

Seen this way, movement across products reflects changes in motivation, not just preferences for different devices. This raises an important question: if users are moving through different motivational states over time, how should products adapt? And if wearables lose engagement over time, the question becomes: how should they evolve as users move through this lifecycle?

How do you maintain engagement?

Companies have a strong incentive to solve this problem, as they make money from continued usage. But not all users move through the wearable journey in the same way, or disengage for the same reasons.

Some users might double down on optimization with advanced lab testing. Others opt out entirely. Still others look for complementary approaches for managing their health that may not have been innovated yet. Understanding these different paths, illustrated below, is key to designing products that sustain engagement.

Appealing to super-machine users

One clear direction companies are pursuing is more data and more tracking. Whoop, Oura, Equinox, Function Health, and Superpower are pushing into full lab testing and offering insights across hundreds of biomarkers.

But this approach doubles down on a narrow segment of high-performance athletes that are already highly engaged, looking to optimize further, and able to afford it (the Optimize by Equinox program is in partnership with Function and costs $40k/year). For context, only a small fraction of the population will ever run a marathon (~1% of the global population). Designing for this group alone won’t solve the broader engagement problem.

Appealing to machine users

Many body-as-machine users are comfortable with sustained tracking. Objective measurement aligns with how they conceptualize the body, and they have the knowledge and tools to act on it. For this group, structure and metrics can support long-term engagement.

Appealing to temple users

By contrast, body-as-temple users may never fully embrace tracking. Objective measurement and extrinsic goal pursuit often conflict with their preference for subjective, intuitive, enjoyable care. For them, wearables may always feel somewhat misaligned, although select holistic features (like sleep or cycle tracking) might hold some appeal.

Appealing to hybrid users

This is the most important group, and the biggest opportunity for companies. Most consumers don’t see the body solely as a machine or as a temple. They see it as both. They want structure, but not rigidity. Optimization, but not burnout. Data, but only when it’s helpful.

The problem is that most wearables are static—while users are dynamic. For this group (myself included), sustained engagement depends on maintaining self-efficacy, supporting intrinsic motivation, and preventing data from becoming overwhelming or demotivating. Instead of being one static product, wearables should adapt to who you are today.

Keep self-efficacy high by complementing data with tools and guidance

Many theories of motivation (Atkinson, Bandura) emphasize the importance of perceived likelihood of success. We pursue goals when we believe we can achieve them and have self-efficacy. This means that people feel they are able to positively affect change through their own actions. Translated to wearables: Data only motivates if users know what to do with it.

When using the Whoop or Oura app, I don’t always understand my scores or how to improve them. The apps offer general suggestions, but often lack clear, personalized, and actionable guidance. Without that clarity, the data can become frustrating. The same applies at an even larger scale to lab testing. Knowing your scores on hundreds of biomarkers may sound helpful, but without interpretation and support, they risk becoming overwhelming.

This is where AI guidance (Anthropic, Microsoft, and OpenAI have all recently announced LLM products for consumer healthcare), or support through coaching, personal training, nutrition, or doctor recommendations becomes critical as part of the product experience.

Meet users where they are by designing in flexible modes

Health motivation isn’t static but changes with injury, illness, life events, environmental factors, and generally shifting priorities. But most wearables (especially those that represent the body as a machine for optimization) assume consistency.

Moments when routines are disrupted—like getting sick, traveling, or going through a busy period—are often when disengagement begins. Or for example, last weekend a friend told me she took off her Oura Ring off for a night out because she didn’t want to see the score the next day. Moments like this, if repeated, turn into long-term disengagement.

So how can the app and data experience change to reflect differing usage contexts and needs across days, weeks, months, and years? One solution is to expand on flexible modes that reflect changing contexts and goals.

I like Oura’s Rest Mode that you can turn on when sick and it stops reporting on activity goals. WHOOP and Oura provide pregnancy tracking features that change targets and provide context for changing metrics. Adding more custom modes can allow users to adjust targets and expectations to match changing goals and environments and avoid making them feel badly about their data. That may mean, counterintuitively for products whose primary value-proposition is the data, providing less data. Some other modes could be:

• “It’s My Birthday Mode” to give users a few cheat days per month without feeling guilty about it. Research shows how having some slack, emergency reserves, or cheat days available is both an attractive and effective tool for goal pursuit.

• Some users may still want to drink and go out more than a few days a month and may not be ready to reduce their alcohol consumption. Give them an “Everything But Drinking Mode” that helps them pursue behaviors they do want to change.

• “I’m Injured,” “I’m Sick,” or “It’s Busy at Work Mode” could allow users to shift goals and tracking for a defined window of time.

• I want a “Maintenance Mode” that only shows me updates once or twice a week, and alerts me if there’s a big change in my data that I need to look at more closely. Otherwise I don’t want to be checking my app every day and don’t necessarily have the self-control or habit formed to drop down to more infrequent check-ins.

Final Thoughts

Wearable fitness trackers are incredibly effective at starting behavior change. They provide structure, accountability, and insight. But sustaining behavior requires something different.

As progress slows, priorities shift, and life intervenes, a daily performance score can become less motivating, or even demotivating. Many users—myself included—eventually feel a tension between wanting to care for their health and wanting to disengage from constant tracking and go back to reliance on their own intuition and knowledge. The irony is that technology designed to improve well-being can sometimes get in the way of it, so many users end up dropping their devices after the useful insights and extrinsic motivation begin to wane.

As an industry though, wearables are here to stay—even though they have a long-term engagement problem. The next phase of innovation won’t come from adding more sensors or more data, but from giving users just enough data at the right time to sustain motivation. It will come from understanding how motivation evolves—and designing products that evolve with it.

These companies can also provide value beyond the data. The trend towards lab testing and holistic care suggests that users want experiences that go beyond the app. Coaching, personal training, meditation, and social features are now all within scope for a fitness app. The most successful wearables won’t just track behavior, but will adapt to how people’s motivation and goals change over time.

Subscribe now

Share

Leave a comment

The evolution of the food pyramid is a great metaphorical illustration of Joseph Campbell’s point. Foods once considered dietary virtues: bread, pasta, and cereal that formed the foundation of the pyramid, are now often treated with suspicion in an era of low-carb and high-protein diets.

New technologies (e.g., AI, GLP-1s, wearable fitness trackers, biometrics testing) and holistic lifestyles (e.g., wellness movements and trends like MAHA that sometimes border on new religions) support changing beliefs about everyday well-being. Companies increasingly acknowledge the gray areas between good and bad and entice consumers to indulge their cravings.

Vice and virtue

Plato and Aristotle referred to good and bad things as virtues and vices. Christian theology also draws a clear distinction between vice and virtue. The seven deadly sins or capital vices are: pride, greed, lust, envy, gluttony, wrath, and sloth. The seven corresponding heavenly virtues to counteract the vices are: humility, generosity, chastity, kindness, patience, diligence, and temperance.

Human nature seems to favor simplifying the world into binaries: us versus them, rich or poor, male or female, right versus wrong. This tendency toward binaries also appears in the study of psychology. There are many dual-process models, especially within the areas of self-control and self-regulation: System 1 versus 2, ego and id, following your heart versus your head, self-discrepancy theory, and current versus future selves (one of my favorite philosophical conceptualizations: “Not gonna fold the laundry? That’s tomorrow Daniel’s problem”).

In the consumer behavior and psychology literatures, vice and virtue are used to refer to what is unhealthy versus healthy. Often times this appears as indulging in pleasure versus sacrificing it.

Behavioral economics definitions often describe vice as maximizing short-term pleasure and virtue as maximizing long-term utility. There is some acknowledgement that these are not hard and fast categories, but relative. One of the more complete accounts of vice and virtue by Wertenbroch (1998) suggests, for example, that “Dunkin’ Donuts Munchkins” versus muffins represent relative vice or virtue. Some of the things on the list are still considered unhealthy.

The problem with these conceptualizations of vice and virtue is that there is no agreed upon definition of “healthy,” because there are so many dimensions to health. The line constantly shifts depending on the individual, the situation, and broader societal perspectives. As a result, relative notions of vice and virtue offer disappointingly little practical guidance for people trying to live healthier lives.

In the last century we’ve seen too many shifts in health trends to count, as scientific research and education proliferates. Cocaine, heroin, and cigarettes went from having medicinal benefits to being firmly in the vice zone. It seems like alcohol is soon following, with guidance that no alcohol is safest for our health. Biometrics and sleep tracking is great, until it gives you anxiety about the state of your body. Interview participants talked about how more women are embracing strength training for the longevity benefits, however talked about how it’s also easy to overdo it and added caution around the effects of high intensity exercise on hormonal health. The protestant work ethic views taking a personal day as a sin, but mental health or personal days are becoming more common. Processed foods are bad, but when you’ve had a bad week, sometimes you just need a pint of (Van Leeuwen Salt + Straw) ice cream.

Understanding how to balance between vice and virtue becomes a difficult juggling act of modern living. However, there does seem to be some consensus around the things people try to limit consuming too much of. One constant may be the presence of regret when engaging in vice.

We recruited 201 people for a survey and asked them which behaviors they felt they needed to limit to avoid overindulging in. The top things were: junk foods and treats, social media, video games, television/streaming, drugs and alcohol, sleeping in/extra rest, spending/shopping, procrastination. These examples also frequently appear in popular and social media discussions.

Much of what we consider healthy depends on which dimension of health we prioritize.

Multiple dimensions of well-being

The default dimension of health people think of is physical. Even within physical health, there are tradeoffs between exercise and rest, performance and injury prevention, speed and endurance, that you have to manage. And when you consider the broader aspects of health, wellness or well-being has been conceptualized with anywhere between 6 and 12 dimensions (e.g., physical, mental, spiritual, emotional, social, financial, environmental, etc.). Making any one decision has potential ramifications across multiple dimensions: Do I work extra towards a job promotion and more money (financial), or do I take a trip to relax with friends (mental, emotional, social)? Do I treat myself after an exhausting day (physical, mental, emotional) or stick to my diet plan (physical)?

Choosing a vice may even be considered virtuous when considering the broader dimensions of well-being.

Virtuous vices

Across those 201 responses, participants described situational and psychological drivers that make vice appealing in the moment despite virtuous intentions. The most common reason to consume vice is for stress relief and coping. People also want immediate gratification, are tired (physically or mentally), or there are environmental triggers or cues, such as in social contexts. Some people also talked about having bad habits or addictions: a certain behavioral momentum that leads them to consume vice when they’d rather not.

In another study, we tested 4 of the top reasons for consuming vice: mental health, physical exhaustion, social pressure, instant gratification, between subjects across two scenarios (eating chocolate cake after a tough day and skipping out on work to take a trip with friends). Mental health was the only reason that reliably reduced perceptions of the vice rating and increased long term perceived virtue goal success relative to control.

I find the role of mental health in consuming vice interesting and suggestive of modern trends of moving towards a more holistic view of well-being. More people are embracing the multi-faceted nature of a healthy life, and a healthy body. My research on viewing the body as a temple is also demonstrative of this trend. The tension between vice and virtue does not just shape individual decisions. It also shapes how companies market indulgent products.

Marketing virtuous vices

Companies don’t appear to shy away from sitting in the gray area between vice and virtue and some lean directly in to the indulgence. I was surprised to see frequent instances of use of “-holic” in company branding in Seoul, Tokyo, and Budapest to sell melon bowls, American clothing (they had a sweatshirt I almost bought that said “Basketball Stadium University”), and beer. At first I assumed this was just an imprecise translation from English, because why would companies want consumers to give in to their cravings? Especially in alcohol sales.

Psychologist Nick Haslam talks about “concept creep” to describe the expanding usage of words like addiction or obsession. He suggests that addiction is now used to describe behavioral addictions like gambling (new to the DSM), as well as the “realm of bad habits and other pleasurable behaviors,” for which we might feel embarrassed or ashamed about. But it can even signal positive behaviors and commitments.

This trend appears in the U.S. too. I remember playing a lot on AddictingGames.com when I was a kid. In addition to Van Leeuwen Ice Cream, I saw a dessert restaurant called Cravings in Ann Arbor and recently got an Instagram ad for an “addictive” game to learn buddhist philosophy. This last one seems quite contradictory given Buddhist concepts of non-attachment and the middle way leave little room for consumer dependencies and black and white binaries.

In my analysis of the fitness class marketplace I came across several companies called: Pilates Addiction, Gym Addict Fitness, and an instagram post for Para Performance encouraging followers to “be addicted to bettering yourself.”

The concept creep of temptation, vice, and even addiction into consumer marketing may pose as a trivial invitation to indulge, but for certain individuals and domains with a history of self-control challenges (e.g., gambling, weight management), the sales tactic verges on inappropriate or even manipulative.

The virtues of the past

Vice and virtue might be useful guardrails to provide a relative orientation toward what is unhealthy or healthy, however they have their limitations for marketing and for the individual. Considering broader views of well-being, people sometimes opt for virtuous vices, especially when sacrificing physical or financial for mental well-being. Beyond individual choices, societies’ priorities and knowledge around health evolve, making older vice-virtue binaries less relevant.

Companies recognize consumer appetite for indulgence (and let’s face it, there will always be appetite), and incorporate appeals to guilty pleasures. Some of these might be cutesy low-stakes offerings that remind people they can treat themselves for their holistic well-being. However this language should be avoided in higher stakes scenarios where people struggle with self-control and setting limits.

The GLP-1 ad in my morning newsletter suggests a new trend regarding viceful temptation. Instead of encouraging people to consume more, the medical intervention acts directly on the wanting and craving neural circuitry to reduce rather than reinforce those cravings. These effects go beyond food and may also help reduce cravings for other behaviors and substances like alcohol, drugs, and gambling. We still need to learn much more about GLP-1s, but early evidence suggests that these are solutions to help limit overindulgence that go beyond behavioral intervention.

A lot is asked of the individual these days in order to manage their health. Consumers need to sort through fact and fiction, vice and virtue, and navigate the gray areas. When it comes to offering clear guidance around consumer health and the “necessities of today,” the old adage comes to mind: “everything in moderation. Including moderation.” The strategies people use to limit the bad stuff and do more of the good basically boil down to two: moderation and abstinence.

Bryan Johnson of Don’t Die notoriety is a big abstainer. Other thought leaders like Michael Pollan or Oliver Burkman suggest moderation and “dailyish” habits. Some people and products seem more apt for leveraging one or the other. I’ll write a follow-up article on the common psychological strategies people use to help limit overindulgence.

Subscribe now

Share

ICYMI from the first post:

Fitness class companies represent the body-as-a-machine, emphasizing high-intensity physical exertion for optimal results (think: HIIT, Barry’s or Solidcore), or the body-as-a-temple, emphasizing holistic balance for overall well-being (think: yoga and wellness, CorePower). In the marketplace these are product and branding strategies and treated as an either-or tradeoff.

And from the second post:

Although the temple representation is more popular in the local boutique fitness class marketplace, the machine is still the default for national fitness brands. I use the example of the fitness wearables marketplace to illustrate how the framework highlights four distinct consumer segments and how a hybrid strategy of temple AND machine can be innovative.

The big questions I’ve been building to are: Does the structure of the health and wellness marketplace reflect consumers’ psychology? Do companies actually understand their consumers? What might be missing and what does this tell us about future consumer trends?

Bottom-Up Theory Building

For a proper human-centered theory or product, you need to start with the user and their underlying needs. So let’s start with you, the reader—that is, the consumer. And let’s get philosophical for a second. When you think about the makeup of your body are you more of a mind-body dualist like Descartes or do you think about your body as an integrated whole? What’s the primary purpose or function of your body: doing or being? (see: Maslow, or Kant: means vs. ends) And what does the body feel like when you’re taking care of it? Should workouts be ultra-intense or gentle and easy?

We designed 10 items to measure metaphorical beliefs about the body. How would you respond to the following statements from 1 (Strongly Disagree) to 7 (Strongly Agree)?

• My body is mainly a mechanical system where parts work together efficiently

• I think of my body as a machine, with each component enhancing its overall functionality

• Optimizing physical function and performance is fundamental for my body

• I can measure most aspects of my body, like reading the gauges on a car’s dashboard

• My body is a powerful tool, used to accomplish my objectives

• My body is mainly a holistic ecosystem where parts coexist in harmony

• I think of my body as a garden, with each feature nurturing and supporting the whole

• Cultivating physical, mental, and spiritual balance is essential for my body

• I believe that my body receives energy and vitality from sources beyond my awareness

• My body is a temple, respected as a sanctuary for my mind and spirit

So given all of that philosophizing, would you say your body is more like a Machine? Temple? Both? Or neither? Here’s how 4,404 participants across 16 months of data collection scored on our measures:

Consumer Trends

To better understand who views their body like a machine and/or temple I conducted an internal meta-analysis across all my studies and looked at demographic predictors of philosophies of body including age, gender, and race. Here are some of the findings.

Increasing Popularity of the Body-As-A-Temple

After controlling for demographic variables, the default view among consumers was the body as a machine. However, there was a small time-based trend across the 16-month window, where scores on the body-as-a-temple philosophy increased relative to the body-as-a-machine. A one-year difference in when the study was fielded equated to approximately .7 standard deviations increase in the temple relative to the machine orientation (y-axis). You can see the later data points creep just above the neutral 0.0 line, suggesting the body-as-a-temple may be the default consumer approach to care sometime in the near future.

Machine is Masculine

Men were significantly higher in machine-oriented beliefs (.22 Standard Deviations) while women were significantly higher in temple-oriented beliefs (0.13 SDs). In the figure below you can see how the cyan line (temple) is higher for women and the red line (machine) is higher for men.

The Temple of Youth

Age showed a curvilinear association with the temple orientation, such that younger consumers were higher in the temple philosophy of body (left tails on the blue lines and then a general negative trend with age above). There were no significant trends with the machine philosophy and age, and there were no significant interactions by age and gender. There’s an interesting bump in machine-orientation for both men and women from the age of ~28-45.

Cultural Roots of the Body-as-a-Temple

Non-White participants were more likely to think of the body as a temple, including Asian-American (0.11 SDs), Black (0.60 SDs), and mixed-race participants (0.19 SDs). Black (0.48 SDs) and mixed-race participants (.20 SDs) were significantly more likely to think of their bodies as temples relative to machines (effect for Asian-Americans was directional, .066 SDs, but not significant).

Political Party, Employment, and Income

On a sub-sample of 596 participants I looked at political party and found that democrats were less likely than republicans (-.38 SDs) to view the body as a temple (likely driven by religiosity or spirituality). There was a small effect that people that were unemployed (-.28 SDs) or not in paid work (-.27 SDs) were less likely to view the body as a temple, relative to those in full-time employment. There was no significant effect of income. It’s less of an economic phenomenon and more culturally driven.

Squaring the Consumer Results with the Marketplace

These consumer trends are relatively interesting and informative, although generally not surprising. What is most surprising though, is that these trends didn’t always align with how the marketplace represented the body based on what kinds of people were featured in their social media ads. In the marketplace analysis I found that companies use older, more female customers in social media branding that is oriented towards the body-as-a-temple versus the body-as-a-machine. There was not a significant relationship between racial/ethnic diversity and representations of body in the ads. The marketplace matches the consumer analysis on gender, but not on age or ethnicity.

Confusing Age with Generation

It’s the younger consumers in Gen-Z (~18-28) that identified the most with the body-as-a-temple and that are likely driving the societal trend of interest in holistic wellness. That age window coincides with the target ClassPass consumers (my model calculated the average age at 29.9).

I understand why companies might think that older consumers want to treat their bodies more like temples. Yoga classes are generally easier to do than HIIT for example, and are going to be more open to all ages and skill levels. Older consumers are not looking to push the limits of their bodies to achieve new gains that those in the age window of ~28-45 are.

However companies are confusing age with generation. It’s easy to forget with all of the exposure to holistic wellness that the body-as-a-temple approach is a relatively new one in Western consumer wellness. Having interviewed a host of people as part of various projects, I think there is still a skepticism from older generations about holistic views of the body and wellness and they generally default to the body-as-a-machine approach to care (or trying to ignore the body and its ailments as much as possible). It’s going to take awhile to help older generations embrace a more holistic approach to care, and I think the messaging of holistic and functional health for longevity from thought leaders like Peter Attia (e.g., "training for the centenarian decathlon") will help—another nice example of a hybrid-approach of machine and temple.

Mismatching Quadrants

The big disconnect is between how companies represent the body as a tradeoff between machine or temple (negative correlation of r = -.89, black dotted line), even though consumers often view the body as both a machine and a temple (positive correlation of r = .41, gray dotted line). You can see how empty the upper right quadrant is in the figure below: the marketplace lacks adequate high temple-high machine, hybrid offerings!

What This All Adds Up To

If you zoom out, the takeaway isn’t that the fitness and wellness marketplaces are wrong—it’s that they’re incomplete. This consumer moment is new, and these marketplaces are still in the early stages of figuring out how people actually want to care for their bodies. Just as consumers are still figuring out what the marketplace can offer them. It’s an awkward dance, but it takes two to tango.

Companies have done a good job carving out clear, differentiated positions: intensity or calm, performance or balance, machine or temple. That clarity makes marketing easier. But human psychology is messier than a continuum or a 2×2. Most want both, some of each, or to be able to move dynamically between these mindsets as their lives, bodies, and needs change.

Right now, consumers are doing the bulk of the work of integration themselves. With few hybrid offerings available on the national scale, they might stitch together things that work: HIIT and yoga, biometrics tracking and mindfulness, for physical and mental wellness. Companies like Equinox, Peloton, LifeTime, Oura Health, and others are realizing the value of all-in-one offerings. Equinox has partnered with Grown Alchemist for hair products, earthbar for nutrition, Clean Market IV for medical spa, Oura Ring for biometric data and tracking, and Aescape and Hyperice for recovery tools. Peloton bough Breathwrk. And yet, the marketplace still largely treats the hybrid philosophy as an exception rather than the norm.

That gap matters. It shapes who is represented by fitness brands in marketing and in classes, which products feel “for me,” and how people interpret what it even means to take care of their bodies and be healthy. When hybrid needs go unmet, consumers either over-index on performance at the expense of balance and risk burnout, or retreat into themselves without building physical resilience and strength. Neither is ideal. But this will evolve.

As younger generations continue to push for more holistic philosophies in the mainstream, older generations look for sustainable ways to stay strong and independent for longer, women build muscle and physical strength, and men tap into mobility and mental wellness, the line between temple and machine will blur further. The companies that win won’t be the ones that pick a side. They’ll be the ones that help consumers integrate the best of both.

And for you, the reader, the question to carry forward isn’t “Which one am I?”
It’s: When do I need my body to be a machine and when do I need it to be a temple?

Subscribe now

Share

In this follow-up article, I expand beyond boutique fitness classes to examine the broader fitness ecosystem, including gyms, apparel, equipment, virtual training platforms, fitness apps, and wearable technologies. I then zoom in on the fast-growing wearables category—drawing on both marketplace data and my own experience wearing a WHOOP band and an Oura ring—to show how different body philosophies are represented not just in marketing, but in product design and user experience.

Expanding the Scope of the Fitness Marketplace

One of the limitations of my original analysis is that the studio fitness class marketplace skews female (about 70%), which likely contributes to the stronger presence of body-as-a-temple representations (Temple = 5.51, Machine = 4.37; d = 0.45).

Company attributes including number of Instagram followers, physical locations, and Meta ads, as well as using a franchising model to scale, were all positively correlated with the body-as-a-machine orientation. While each factor explained less than 1% of incremental variance on its own, together they suggested a pattern: larger, national fitness companies with bigger advertising budgets tend to frame the body more as a machine than as a temple.

To test this hypothesis, I analyzed the social media of 75 additional national fitness brands across 6 other domains:

• Fitness technologies (9 companies incl. WHOOP, Oura, Garmin, Hyperice)

• Virtual workout platforms (20 incl. Peloton, Bodi, Sweat, Les Mills)

• Tracking apps (5 incl. Strava, MyFitnessPal, Training Peaks)

• Workout equipment (12 incl. Wahoo, Nordic, Liforme, Tonal, Gaiam )

• Gyms (15 incl. Planet Fitness, LA Fitness, LifeTime, Equinox, etc.)

• Apparel (14 incl. Nike, Alo, Vuori, Lululemon, Underarmour, Adidas)

Across these categories, brands scored substantially higher on body-as-a-machine than body-as-a-temple (Machine = 7 and Temple = 4.16), a large effect size (d = 1.24).

Together, these results suggest that the default orientation of the broader fitness marketplace is machine-based, particularly in domains centered on performance, optimization, and technology. Boutique fitness studios appear to be an important exception. They are at the forefront of a rising trend in the popularity of the body-as-a-temple orientation in pursuit of general wellness.

In general, technology supported fitness is more consistent with the body-as-a-machine (correlated r = .63 in the Part One analysis; machine = 8.33, temple = 3.44 in the chart above). But this raises an important question: does all fitness technology have to frame the body as a machine? To explore how companies differentiate within this category, I analyzed social media content from 24 wearable fitness brands, using the same AI-based model.

How Wearables Represent the Body

I first became interested in wearables when I interned at a small wearables startup, AMPY, in 2015 (I played a lot of ping pong but also built and tested prototypes of different wearable concepts).

Fast-forward 10 years and AI is the new but better IOT and the wearables market has more than tripled in size (~$20B to ~$70B). Over 1 in 3 US adults used a wearable fitness tracker as of 2023. With better technology and AI-aided analysis and insights, wearables can be great tools for managing the health of your body.

I analyzed the social media profiles for 24 wearable fitness tracking brands to see how my framework represents the marketplace landscape. And what I found suggests interesting trends and insights.

Four Segments of Wearable Fitness Trackers

The results reveal four distinct segments in the wearable fitness landscape, organized by how strongly brands emphasize machine and temple representations, relative to the category averages (average machine = 2.4, temple = 4.9) .

Low Machine + Low Temple (machine = 0.56, temple = 1.56)

These brands position wearables primarily as consumer electronics (Apple, Google, Samsung) or fashion accessories (Fossil), with fitness as secondary. This is the largest marketplace segment, consisting of more general consumers. I wore an Apple Watch for years because I liked how it handled notifications and reduced dependency on my phone (and also could help me locate my phone).

High Machine (machine = 7, temple = 3.33)

These companies emphasize optimization, strain, recovery, and performance output, positioning the body as a system to be trained and tuned. This is the second largest segment, led by Garmin and Fitbit, with up-and-comers like WHOOP, Ultrahuman, and Withings.

High Machine + High Temple (machine = 5.5, temple = 7.5)

Oura stands out as the only brand that strongly integrates both philosophies, positioning the body as something to both optimize and care for. As the below ad suggests: “there’s one ring that can do it all.” This is a new segment with high potential, because it combines the best facets of the body-as-a-temple and machine approaches to health. Other brands like Luna Ring are trying to catch up to Oura.

High Temple (machine = 2.5, temple = 8.5)

These brands emphasize stress regulation, sleep, subjective well-being, and hormonal health. This is a newer segment, with companies like Apollo Neuro, Evie, and Bellabeat (Ivy). This segment may have less growth potential, as my research suggests that the more body-as-a-temple oriented consumers are less interested in taking care of their bodies using physical metrics tracking and technology.

Evidence from Using WHOOP and OURA

WHOOP and Oura are often compared, but they differ in more than form factor (ring versus band). They reflect different philosophies about how people should relate to their bodies. After a year of wearing a WHOOP band, I was curious to try something new and bought an Oura ring. Wearing both devices simultaneously for 3 weeks made clear how their designs, language, and motivational strategies translate abstract body philosophies into product experiences.

First, Some Background on the Companies

Both are growth startups. WHOOP was founded in 2012 in Boston and has $404.8m in funding, with a series F round in 2021. Their celebrity spokespeople are primary top-performing athletes like Cristiano Ronaldo, Patrick Mahomes, Michael Phelps, Sha'Carri Richardson, and Rory McIlroy. WHOOP seems to be going after the Peter Attia/Andrew Huberman-type fitness and health enthusiasts who value science-backed results (one of their products is called MG for “Medical Grade”). This seems to be a popular approach for companies to move from wellness to health (Weight Watchers and Noom with GLP-1’s), but they’ve gotten in trouble for overstating claims about blood pressure insights, without an FDA clearance.

Oura Health was founded in 2013 in Finland. It’s a bit bigger at $1.5b in funding, with a series E in Oct of 2025. Their celebrity spokespeople are more wide-ranging and informal, with some athletes like Odell Beckham Jr., Lindsey Vonn, Chris Paul, as well as a host of celebrity wellness influencers: Jennifer Aniston, Kim Kardashian, Prince Harry, and Gwyneth Paltrow. Oura is going after a higher-end luxury wellness market, exhibited by nice Apple-esque packaging, and partnerships with Amex (the Platinum card provides a $200 credit per year and the ring comes in a special co-branded packaging), and Equinox (also a brand that leverages a hybrid representation of the body as both a temple = 5 and a machine = 8; and also has an Amex credit of $300/year).

*New note as of 2/12: Chase Sapphire Reserve customers can get a free year subscription of WHOOP. It’s interesting to see these companies taking sides.

Data Emphasize Different Goals

The ring versus band designs afford different usability: I can play basketball with the WHOOP but not with the Oura. Aside from that, their sensing technologies and tracking features are pretty similar. For those unfamiliar, both WHOOP and Oura provide a daily sleep score, a recovery/readiness score, and a strain/activity score. Oura also prominently displays stress and resilience ratings.

Despite similar sensing technology, there are subtle differences in how their algorithms calculate the scores. Others have tracked their scores on both, and found that while the sleep tracking is similar, the readiness vs. recovery scores differ greatly. Here’s my data for the three main scores (Oura in blue, WHOOP in orange).

Oura was much more generous with its readiness and activity scores (demanding less of me, in pursuit of optimized performance), but slightly less forgiving on sleep scores (WHOOP was ~2 points/night higher). Other values, including heart-rate variability, resting heart rate, respiratory rate, and cardiovascular/biological age (I’ll tell you mine if you tell me yours) were all pretty similar on both.

Product Experiences Emphasize Performance Vs. Balance

The differences in scores are further linked to the in-app messaging and experience.

WHOOP tries to encourage you to match your strain score (from workouts) to your recovery score (from sleep and rest) to be in an optimal band to “Build Your Best Self.” I wore the WHOOP band and journaled my health behaviors for 358 days straight and definitely saw gains. After I stopped using the device for a bit, WHOOP invited me to join a 30 day streak challenge to “finish the year strong and earn a free band.” They’re strategically leveraging how the body-as-a-machine approach corresponds with discipline, habit-building, and streaks. From the chart above, you can see how the tracking app category is also higher in the machine-orientation (machine = 7.0, temple = 3.4).

Oura on the other hand has been a less effortful in-app experience so far, with only the occasional tagging, instead of daily journal entries. I like the integrated Body Scan meditation from Headspace, linked to Stress and Resilience.

One night I got notifications to go to bed from both apps, which nicely illustrated the different language used to convey the approach to motivating better health. “Winding down your body and mind” for the purpose of “good quality sleep” (Oura: emphasis on holism, positive experience) vs. “wind down effectively” to “stay on track” (WHOOP: emphasis on results, discipline).

Some Other Features

Both automatically track and categorize activities, although WHOOP only records those that elevate your heart rate enough for more than 15 minutes and Oura seems to record over 10 minutes (I’m assuming to give you credit for lower-intensity activity).

Oura is better able to track naps. It’s nice to see your readiness and sleep scores go up after one, whereas once WHOOP sets your recovery score, it does not change until the next night’s sleep.

Oura has a designated “Symptom Radar” feature and tab to detect early signs of illness. On WHOOP you might see your recovery scores drop lower than usual without explanation, but have to interpret it for yourself.

WHOOP has a nice social feature where you can create custom groups to share and compare stats with other users (shout out to the Equinox early risers in Sauna Iguanas) or join a large fitness community (e.g., Road Cyclists, Road Runners). Oura has a similar Circles feature. I really liked WHOOP’s alarm that would wake me up when I hit 100% of my sleep goal and help me feel well-rested.

Overall Comparison

The marketplace messaging and in-app experiences of Oura and WHOOP, afforded by their different form factors, help to differentiate these wearables in the marketplace. WHOOP is geared towards high-intensity performance training and recovery (body-as-a-machine) and is quite effective at that. I saw nice gains in HRV and VO2 Max across my year of usage, but ultimately felt a bit burned out from using the device.

Oura is positioned as more of a luxury wellness and lifestyle product that can help you balance performance, stress, and rest (body-as-a-temple AND body-as-a-machine). Both motivate better care of the body, but do so through different approaches. At this point in my life, I’m drawn to the more balanced approach supported by the Oura Ring—one that emphasizes effortlessness and resilience over discipline and constant optimization. But if I’m looking to train for a marathon or triathlon I would put the WHOOP back on.

Final Thoughts

Wearable fitness trackers can help people care for their bodies by translating physiological data into motivating insights. The default approach in fitness is by framing the body-as-a-machine to be optimized for performance. But there is such a thing as too much optimization and data. Many users—including myself—eventually reach a point where constant tracking can become a health deterrent. I liked this post about how these devices help to establish better care, but we ultimately must “transcend” them in favor of our own bodily awareness (and something else yet to be innovated?)

I’m optimistic about where wearables are headed. With advances in sensing technology, AI-driven analyses, and integration with healthcare systems, wearables have the potential to move beyond gimmicky self-quantification toward genuinely supportive health infrastructure. The Oura CEO talked about how they’re trying to get into more of an enterprise, insurance-covered healthcare model. What about going beyond wearables? Both companies (Equinox too) seem to be pushing advanced lab tests. EightSleep offers an intelligent, bed-cooling system. But the most compelling future may lie in combining the best of both body philosophies: the precision of the machine with the care and balance of the temple. Oura offers one example of how this hybrid approach can work. Whether WHOOP and other performance-oriented brands follow remains to be seen (Garmin has a sleep band now).

I liked analyzing this sub-industry of wellness and highlighting a couple of interesting companies. I might do a few other analyses like this, maybe on mental health and wellness products. If there are companies or areas you’d like to have analyzed, let me know.

Leave a comment

I coded up an AI model to rate how the body is represented across ~10,000 social media images from fitness companies. I’m excited to share some of the results of the analyses, which are hopefully interesting and informative for both consumers and professionals. In this article I’ll provide more details on the methods and highlight some of the results. In future articles I’ll share more about segmentation trends and spotlight specific companies and sub-industries of fitness.

Consumer Psychology of the Body as a Temple and Machine

As a refresher, I previously defined the two primary philosophies of body that guide health and wellness consumption as:

The body as a machine (or vehicle) means the body is viewed as a mechanical system that requires maintenance and functions efficiently to perform tasks.

The body as a temple (or garden) means the body is viewed as a living entity that should be nurtured, emphasizing holistic well-being.

These philosophies reflect motivating beliefs about the body, relative to not thinking about the body, either explicitly or metaphorically. The two philosophies are not fully independent but have overlap. Many people view their bodies as a combination of both, and the two philosophies are positively correlated with psychological constructs like body awareness, self-efficacy, and overall well-being, as well as with health and wellness consumption.

That’s the consumer psychology side of the framework in a nutshell. The natural next question is whether—and how—these philosophies show up in the marketplace. Do companies strategically represent the body as one or the other (or both), and if so, which types of companies and for which target segments of customers? I’ve spent the last few months answering this question.

Marketplace Analysis Methods

In order to do a rigorous quantitative analysis of the wellness marketplace, I wanted to capture marketing media for a large and representative sample of U.S.-based companies. I had to make several methodological design decisions in order to prepare the model and analyses.

Focus on Fitness

First, I narrowed the scope of the industry because as I’ve written, and others have also reported on: the wellness marketplace is huge. I chose to focus on physical activity and fitness because this is a mainstream, well-established sub-industry of wellness that is centered around the body. It would be hard to be a fitness company and not talk about or show the body in marketing media.

To provide a clean list of fitness companies, I included companies that offered classes through ClassPass.com, a popular consumer marketplace for fitness classes with over 248 million classes booked as of October, 2024. I selected classes in the three largest U.S. cities (New York, Los Angeles, and Chicago) to get some variance in the classes offered and be more representative of national trends. Typical wellness customers are generally urban and higher income, so I felt okay leaving out smaller markets. I ended up selecting 678 companies. I coded each company for the class category, using the ClassPass.com options (e.g., Yoga, HIIT, Martial Arts, etc.) and adding a general fitness category for companies that had 3 or more workout types offered, and a wellness category for companies that offered meditation, healing, and/or other holistic services.

Social Media for Fitness Marketing

Next, I chose to capture marketing communications through social media, given that this channel is so integral for wellness (and fitness) marketing. Nearly every wellness company has an Instagram, and throughout interviews I’ve conducted, social media always comes up as a source for recommendations. I gathered Instagram homepages (with up to 10 images on each) and any active company ads from the Meta Ad library.

Analyzing the Social Media

The aggregated social media for 678 companies yielded 9,901 individual posts and ads to be analyzed. Now, that would be a problem for me to analyze by hand. But I’ve been hearing so much lately about how AI is pretty good and can do helpful stuff. So rather than attempting to rate thousands of images by hand, I built an AI model to do the heavy lifting. Specifically, I used GPT-4.1 via the OpenAI API to rate each company’s marketing media with scores of 0 to 10 for how it represented the body as a temple and as a machine. This generated overall temple and machine scores for every company. To validate the model’s outputs, I compared its ratings on a subsample of 25 companies with ratings from independent human judges recruited through an online panel. They correlated above r > .70.

Another great feature of leveraging an AI model to do the rating task is that I could prompt it to gather a ton of additional variables, which would be burdensome for human raters to code for. To give a sense of how the model was scoring the representations of the body, I also had the model code the media for the different attributes below.

No big surprises here: the correlational results showed a clear distinction between temple and machine representations of the body, which were consistent with my framework and prior research. Temple-oriented media was characterized by a stronger focus on holistic wellness, recovery, community, positive emotions, and themes of body transformation (like the CorePower ad). In contrast, machine-oriented media was higher in the muscularity and leanness of the models, intensity of the workout, use of technology, and an emphasis on performance (like the Solidcore ad). The temple-oriented media was also higher in luxury aesthetics and beginner friendliness. There’s a lot to unpack just in the themes here, but I’ll keep moving forward to the results.

Results of the Marketplace Analysis

The results of the analysis provided nice validation of my framework, demonstrating that the fitness marketplace generally represents the body as a temple or as a machine. It also yielded some new insights, suggesting opportunities for fitness companies to innovate to address unmet needs within the fitness marketplace. I’ll share a few of the main findings.

Body Represented as a Temple or a Machine

First off, as expected, most companies do indeed represent the body as a temple or as a machine. While I assumed that the machine representation would be more common, as it has generally been the default philosophy of body for much of the health and wellness industries for the last century, the temple representation was in fact more popular within the sample of companies. The bar chart below shows the counts of companies with each type of body representation.

This pattern is likely attributable to the sample, which was dominated by boutique fitness classes that disproportionately attract female consumers. My model calculated the average percentage female of the media to be 69%, right in line with online search traffic statistics, and consistent with ClassPass trends. Across converging evidence in my research, temple-oriented representations are more strongly associated with femininity. So the skew towards temple representation makes sense. However in further analyses, when I controlled for gender and other predictors of temple and machine representations, I found that the machine representation is still the default fitness marketing strategy.

Furthermore, the marketplace representations of the body as a temple and machine were very strongly negatively correlated at r = - .89. Illustrated another way, you can see heavy clustering in high temple - low machine (e.g., CorePower at 7.5 - 4) and low temple - high machine quadrants (e.g., Solidcore at 2 - 8) in the figure below. This suggests that, in practice, many companies treat these two representations as opposing branding strategies—even though they are not psychologically incompatible for consumers.

This distribution of body representation scores was a surprise, as I expected more companies would represent the body as both. In my consumer psychology studies, the body as a temple and machine were positively correlated within consumers’ beliefs between r = .25 and .50. This means that consumers often think of the body as both a temple AND a machine, and not just one or the other. In other words, many companies may be operating under a false dichotomy—missing a sizable segment of consumers who view the body as both a temple and a machine (“holistic machines” or “temples for performance”).

To reach these consumers, companies could utilize a hybrid approach on either the marketing and brand level by leveraging language of both holistic wellness and balance, and high-intensity and performance, and on the service level by offering a variety of workout and recovery options (similar to how Equinox, LifeTime, and TruFusion do).

In my framework, each of the temple and a machine representations is a lever for a company to pull for branding and marketing to strategically target different segments of consumers. Given that few companies are leveraging both, what types of companies use each representation?

Temple or a Machine Based on Type of Class

Class type was by far the largest predictor of body representation, explaining about 2/3 of the variance (other factors like gender and age added an additional ~7%). Boxing, HIIT, weight lifting, endurance, and sports class companies scored higher in machine. Yoga, wellness, dance, and barre scored higher in temple. I’ve included the average body representation scores by class type, with standard error bars and counts of companies by class type below.

Looking at the distribution of scores within two of the largest categories—HIIT and Yoga—shows that each of these are quite skewed towards either the machine or the temple representation.

Pilates showed a more balanced distribution of both temple and machine orientations, and provides a particularly clear illustration of how these two philosophies can coexist within a single fitness category.

Pilates was invented 90+ years ago as a mind-body exercise (traditionally temple-oriented). It has continued to be one of the most popular and fastest growing fitness categories. With increasing demand, companies have differentiated themselves by ramping up the intensity with newer offerings like Solidcore and the Lagree style, which align with the machine philosophy of body. This is reflected in its growing appeal to men (see the two men in the earlier Solidcore ad), including even NFL players. My model showed regional differences in Pilates as well, with more LA-based studios offering the higher-intensity Lagree style. As a result, LA companies rated higher in machine and lower in temple representations than either NYC or Chicago.

In summary, while within the existing marketplace dynamics temple and machine representations are often heavily tied to the type of fitness offering, they don’t inherently need to be, and there is demand for more hybrid offerings. This is especially true as categories become crowded and companies seek to differentiate themselves and attract specific segments. Within Pilates for example, companies are offering hybrid classes on each end of the spectrum: from HiiP (high-intensity interval pilates) to “metaphysical” pilates. Relatedly, CorePower has added a higher intensity Yoga Sculpt class to complement their core Yoga classes. I expect this trend of crossovers and hybrid body representations to continue as the fitness industry grows.

Final Thoughts

I covered a fair amount of ground here, particularly on the methods, to set up a series of deeper dives in future posts. If I glanced over anything too quickly, or you found something worthy of more in-depth discussion I’d love to hear it.

The main takeaway from this article is that by using an AI model to rate a large sample of fitness class companies, I found that companies represent the body as a temple and as a machine. More specifically, fitness companies typically leverage one or the other marketing strategy, which often aligns with the type of class (temple for yoga/wellness vs. machine for HIIT/strength/endurance training).

This is a simplistic way of representing the body though, given that most consumers actually view their bodies as a combination of a temple and a machine, to varying degrees. Companies can leverage the temple and machine framework to differentiate their brand and find new consumers. As competition intensifies and consumer needs evolve, I expect more fitness and wellness brands to experiment with representations of the body that integrate both temple and machine philosophies—better reflecting how high-value wellness consumers actually experience and care for their bodies.

I didn’t have enough space to talk about the makeup of the different segments of consumers and drivers of temple and machine orientations beyond class type and briefly mentioning city and gender. In the next few articles I’ll go deeper into how different demographic and company attributes predict body representation marketing strategies. I’ll also share analyses that go beyond just fitness classes and look at other product offerings like workout apps and wearables, and highlight interesting companies identified by my model (e.g., Equinox, Oura Health, and more).

William James spent more time than most, even as far as academics go, observing his own thoughts. In 1892 he coined the phrase, “wonderful stream of our consciousness” and went on to write that the “inner experience” of the human mind is:

“Like a bird’s life, it seems to be an alternation of flights and perchings… The resting-places are usually occupied by sensorial imaginations of some sort, whose peculiarity is that they can be held before the mind for an indefinite time, and contemplated without changing; the places of flight are filled with thoughts of relations, static or dynamic, that for the most part obtain between the matters contemplated in the periods of comparative rest.”

James was a pretty cool, philosophical guy and went on to become the father of American psychology. He wrote about spiritual, mystical, or "religious" experiences, another topic I’m fascinated with. Unsurprisingly, his dad was a theologian. His younger brother was the famous author, Henry James, who wrote one of the few classic works of literature that I enjoyed reading in middle school, and his godfather was Ralph Waldo Emerson.

As an off-and-on-again meditator interested in wellness, and a PhD student with a lot of time on my hands to think about psychology, James’ writings inspired me early in my program to engage in a little introspective phenomenology myself. Biking down the West Side highway path, walking through Central Park, or not paying attention during a seminar, I observed how my mind wandered. Stress made my mind return to the same thought, caffeine made my thoughts race even faster, a marijuana gummy made them slow down, and a mushroom chocolate made them go to only god knows where (sometimes to the realm of self-transcendent experiences). No one in consumer research was talking about mind-wandering, despite research showing that the mind wanders almost 50% of our waking lives!

Mind-wandering is well researched within psychology though. Since the days of James, modern neuroimaging techniques have revolutionized the study of mind-wandering. Since an accidental discovery by Raichle and colleagues in 2001, mind-wandering has become associated with the “default mode network” (DMN), a pretty incredible network of brain regions that are active when our attention is not engaged in a demanding task and when the mind is wandering. I first heard about the DMN in 2019 when I was a TPM at pre-IPO Lucid Motors and read Robert Wright’s Why Buddhism is True while on a quest to figure out why I was so anxious all the time and what I could do about it. It turned out to be quite a helpful piece of information that set me on a path forward.

Mind-wandering is our default brain setting, and is therefore the foundation of our inner experience. It provides shape to our consciousness by contributing to many of the contemplative states (perches or resting-places of imaginations), and the flights between. But mind-wandering often gets a bad rap. The 2010 Killingsworth and Gilbert article in Science found a correlation between mind-wandering and unhappiness. Mind-wandering is often linked to rumination, a common feature of anxiety or depression. And mind-wandering is often incorrectly categorized as the opposite of “mindfulness.” While research has shown that meditation can lower activity in the DMN (as well as reduce anxiety), it often operates through mindful awareness (Jon Kabat-Zinn says present moment AND NONJUDGMENTAL → this second part so often gets missed!) of one’s own mind-wandering. It’s not mind-wandering’s fault! Mind-wandering is simply there: a misunderstood scapegoat, inseparable from the human experience. It’s how we react to the mind that is wandering that makes all the difference.

I like to think of mind-wandering as the default algorithm that helps us make sense of the world and motivate action. It’s likely the case that few other species mind-wander, with limited evidence that some animals may be able to mind-wander and self-reflect. As far as we know, AI does not mind-wander. And it may never do so in the way humans do. When the mind wanders, it is most often to consider autobiographical plans in the future. We mind-wander to things we have yet to do, as a helpful motivator to get stuff done. If you ever notice your mind going back to the same far away fantasy, it’s probably meaningful to listen to and consider why. I won’t go as far as saying Freud was totally right, but mind-wandering does give us limited access to our subconscious, spontaneous, and more implicit processes. We can receive that information, and then decide what to do with it = Mindful awareness of our mind-wandering. Wright made a nice point that specific aspects of mind-wandering may not be evolutionarily adaptive given the demands of modern life and influence of addictive technologies. But only by better understanding our mind-wandering, can we move forward in adopting helpful tools and practices to thrive.

So that was a quick overview of some of the psychology of mind-wandering. In terms of the consumer psychology, here’s ChatGPT-5.2’s summary of the article we wrote:

Most of us assume our consumer choices come from deliberate thinking—comparing options, weighing prices, deciding what we “should” do. But this article argues that a huge share of buying, browsing, and brand influence actually happens while our minds are elsewhere. Drawing on psychology research showing we spend roughly half of our waking life mind-wandering, Russman and Schmitt propose a simple idea: those off-task thoughts—daydreams, mental “what if” simulations, and little moments of self-reflection—aren’t mental noise; they’re a major engine of consumption. In everyday life, things like boring routines, emotional states (stress, guilt, excitement), salient cues (an ad, a smell, a storefront), and unresolved “current concerns” (health goals, status worries, vacation fantasies) open the door for wandering thoughts that can nudge us toward three kinds of outcomes: aspirational consumption (buying toward an ideal self or future), generative consumption (spontaneous, creative, sometimes impulsive choices), and contemplative consumption (using products and experiences—like art, wellness, travel, or mindfulness—to reflect on values and identity). The takeaway for non-academics: if you want to understand why you suddenly “need” something, why certain brands stick, or why scrolling can turn into spending, pay attention to what your mind is doing in the in-between moments.

If you want to read the full article you can download it here. Ultimately, we make the point in the article that mind-wandering is an important phenomenon and understudied in consumer behavior and marketing. It is relevant in the context of consumption and can act as the mechanism or catalyst for purchasing (e.g., impulse scrolling and online shopping), or we purchase with the intent of altering our states of mind-wandering (e.g., coffee to speed up, or meditation or fitness activity to slow down). It is innately tied to wellness products and services. And it’s just super interesting. Hopefully you think so too. Feel free to message me or reach out with questions or follow-up topics.

Early in my PhD program I sought out to answer questions like, why do people meditate? What are the top reasons people say for why they go to a therapist? To the gym? Aside from the obvious answers e.g., to chill out, talk through problems, get huge muscles, I wanted to understand the underlying motivations that might predict why people do one thing over another. Do I even know why I do any of those things? Yes, it’s to feel better. But at a deeper level it seems to somehow fit with my own idea of what a healthy self and body entails, and the different preferences and parameters within my control.

Despite health and wellness interest and consumption continuing to grow, with the global marketplace as large as $6.3 Trillion by some estimates, it turns out that there’s not a lot of empirical research on the underlying motivations and drivers for different types of health and wellness products and services. For example, what do protein shakes, fresh juices, sound healing, fitness competitions, and yoga retreats have in common? Well, probably not that much, beyond the fact that they might all be more or less “healthy” things people do to feel better. (ChatGPT says of them: “they are modern rituals of self-optimization tied to personal well-being and identity expression”)

Metaphors for the Body as a Temple or Machine

So as a way to understand what models of healthy bodies different people follow (and by extension how people consume to take care of them), the Body as a Temple and Machine was born. Abdi’s posters got my advisor, Bernd, and I thinking. Akin to seeing your body as a home to live in, you might also hear people talk about the body as a temple, machine, vehicle for pleasure, garden, and so on. I talked about this a little in my first Substack post. These metaphorical phrases get tossed around in modern culture, media, and literature, but no one (including within the psychology or consumer behavior disciplines) has set out to clearly conceptualize what they mean and how they impact health and wellness behavior. For my dissertation, we’re calling the body as a machine versus temple the Philosophies of Body, a play on philosophy of mind that puts the emphasis on the body.

Take a moment and consider, how would you complete the following sentence: “My body is a ___.” How does that answer change by the situation or over time? How does that answer relate to how you take care of your body and which products and services you might purchase to do so?

In general, metaphors help us make sense of complex concepts by linking them to easier-to-understand concepts, guiding perception and action. Each of these body metaphors offers a way of simplifying the complexity of the human body and distilling it into the most important rules for care. As background reading for this project, I picked up Bill Bryson’s The Body: A Guide for Occupants, and quickly got overwhelmed with all of the parts and systems involved in the human body. I’m not a medical doctor or in the biological sciences. Thinking about my body as a temple or machine is a lot easier for me.

Participants Provide Body Metaphors

In an early study we asked a couple hundred people that we recruited from an online sample of participants around the country to choose and describe metaphors for their bodies. “Temple” was the singular most mentioned metaphor (34 times). “Machine” and related metaphors (including “tool”, “instrument”, “factory”) were also commonly mentioned (45 times), as well as vehicle or car and related (including car parts – 21 times), and metaphors alluding to an organic system (e.g., organism, plant, flower, forest, planet - 10 times). Some other metaphors mentioned were a “house” or “happy place” (7 times). While people may use many different body metaphors, we find evidence across studies that the primary metaphors of the machine and temple capture a great deal of the variance in explaining relevant beliefs about the body.

Several people mentioned explicitly negative metaphors such as “bag of flour,” “roller-coast,” “trainwreck,“ or “meat sack.” I feel for those people, and some days I can relate. I would say that these people do not identify through a body metaphor and would be part of a “No metaphor” third group.

Two people mentioned the temple body metaphor, but when they described why, they said it was the only one they could think of. The body as a temple has become more popular, and used somewhat ironically at times (like by Homer Simpson or Jeffrey Winger), as new age offerings incorporate more spirituality into the mainstream. And as the machine metaphor may at times lack a nuanced understanding to be successful in modern life.

Definitions Generated from Participants

In a later follow-up survey we asked a couple hundred participants to describe their bodies through metaphor and used text analysis methods and ChatGPT to generate definitions. We paired similar metaphors (machine and vehicle, temple and garden) to increase the identification with the category of body metaphors. Especially because some people may have more of a holistic view of the body than the machine metaphor entails, but not feel spiritual enough for the temple: enter the garden. Here’s what we came up with:

Body as a machine or vehicle: the body is viewed as a mechanical system that requires maintenance and functions efficiently to perform tasks. In the literal sense a machine or vehicle needs energy dense fuel and regular maintenance and downtime to operate efficiently and productively. Which might translate to a person prioritizing macro nutrients, exercise, sleep, recovery, to perform at a peak physical level at the office, in the gym, or on the field.

The body as a temple or garden means the body is viewed as a living entity that should be nurtured, emphasizing holistic well-being. A temple or garden needs devoted care or nourishment to keep clean, holy, or weed-free in order to cultivate a community of worshippers or harvest of flowers, fruits, and vegetables. This might be reflected in a person practicing self-compassion or mindfulness, eating pure or natural ingredients, in order to care for the mind, body, and spirit.

Maybe body metaphors will help attendees at the 2025 Faces of Fitness Festival in Chicago select which station of the “Wellness Village” to go to.

Next Steps and Reflections

We have plenty of work to do next to understand at a deeper level how body metaphors connect with care. And how people come to identify with a given body metaphor, or may identify with multiple. A theme that’s been common in interviews and is apparent for both Abdi and me, is that our upbringing, including members of our community’s beliefs and behaviors around health have a lot to do with our own. My mom was a big fan of homeopathic remedies (I used a lot of echinacea and calendula until I discovered the magic of Tylenol and Ibuprofen as an adult), alternative milks (circa 2000, before they were cool), and all kinds of exercise classes. Growing up in Ann Arbor, that was pretty normal. Although I think the machine body metaphor served me well playing sports and competing in the classroom growing up, these days I’m looking for more than just physical performance, but also holistic well-being. It's interesting to think about how these body metaphors might change across contexts or eras of our lives, with evolving goals and priorities.

It also makes me wonder, is there a “healthiest” body metaphor? How do body metaphors fit into actual health and well-being (e.g., weight, injury prevention, strength, endurance, mental health, social community, etc.)? Despite the growth in health and wellness marketplace offerings, why does it seem like people are only getting more unhealthy and unhappy? With obesity and loneliness rates on the rise, where is the disconnect? Can body metaphors help people to recenter and identify what matters most for them, so they can better navigate through the chaos that has become the wellness world?

Thanks for reading and I’m curious to hear what thoughts or questions come to your mind.

As a Marketing PhD candidate at Columbia Business School studying Health and Wellness consumer behavior, and a regular Well section reader, I was lucky that the event was happening around the corner from me. I’ve also been meaning to try my hand at longform, less academic writing to explore the intersection of my research, personal interests, and broader consumer and cultural trends. The launch of the Well Festival felt like a good moment to launch this Substack.

Here, I’ll be sharing reflections and research on health and wellness consumer behavior through a multidisciplinary lens—marketing, psychology, sociology, philosophy, and a little existential inquiry. My hope is to offer thoughtful perspectives for consumers, managers, and policymakers navigating this rapidly evolving space. In this first article I’ll describe some of my observations from the festival that relate to some of my ongoing research endeavors and random musings (actual title of a pinned note on my notes app).

The Well Festival

So without further ado…

For those unfamiliar, the Well is the New York Times’ lifestyle section focused on personal health and wellness—distinct from its Health desk, which covers medical news and public health. The reputation is built around providing, “Evidence-based guidance to help you live your healthiest life.” The entryway to the festival featured the tagline of the day: “The facts of feeling good.” The phrase previewed both the editorial approach—journalist interviewed wellness figures, as opposed to marketing booths and group immersive experiences—and the breadth of topics covered: Happiness, Nutrition, Relationships, Fitness, Longevity, and Sleep. In her opening remarks, Lori Liebovitch, the editor of the Well, summed up the section’s mission

“Many of us are asking the question, how can I live better? That question is at the heart of what we do every day at well. Our journalism is designed to help readers live smarter and healthier lives. We demystify health trends. We parse the latest research and we offer evidence-based guidance on your mental and physical health.”

For those interested, recordings of most sessions—excluding those hosted by Estee Lauder and Gilead Sciences— and summaries are available on the NYTimes website.

Attendees

Most of the roughly 400 attendees appeared to be stylishly dressed women in their 30s to 60s—a demographic that mirrors broader trends in wellness consumption. I spoke with attendees from New York as well as nearby cities like Washington, D.C., many of whom had professional ties to the health and wellness space. Their backgrounds ranged from plastic surgery, nursing, and therapy to journalism, health coaching, beauty, and executive coaching. Others were curious wellness enthusiasts or guests of specially invited attendees who had received complimentary passes.

Sponsors

The festival’s main sponsors were Estée Lauder and Gilead Sciences, each hosting their own dedicated panel session. While Gilead and its HIV prevention drug might not be the first thing that comes to mind when you think of wellness, their presence lent a sense of scientific legitimacy to the event. Estée Lauder similarly highlighted its research credentials with taglines like “40+ years of night science research” and “The night skin science experts.” Other sponsors included Capital One Venture X—a reminder that you most often need money to pursue wellness—and Danone, the French multinational food-products company.

Positioning Among Other Wellness Events

As someone who's followed the wellness space closely—both through my research and own consumer behaviors over the last 15 years—I’ve observed a steadily growing interest in wellness. We make this point using google trends search data in a work in progress paper (here’s the relative interest in the term over the last 20 years).

In recent years, large-scale, wellness-oriented events have also begun to emerge more visibly. The Global Wellness Institute has hosted industry conferences since 2007, though these are primarily geared toward professionals, not consumers. More recently, Faces of Fitness has hosted annual festivals in Chicago since 2023, branding itself as a kind of Lollapalooza for fitness—a key facet of the broader wellness landscape. Last summer, while conducting field research in Bali and Southeast Asia, I was surprised to come across a multi-week wellness festival in Singapore, staged in front of the Marina Bay Sands. And in 2024, the Eudemonia Summit launched as a more consumer-facing experience, appealing to the masses but also marked by a more skeptical stance toward mainstream medicine (more on that later). I expect we’ll see more of these kinds of events in the years ahead—particularly as social connection, community, and relationships continue to be recognized as core components of happiness and overall well-being.

Terminology of Wellness

Anyone who’s worked in marketing, watched Mad Men, or posted on social media knows that the language we use—especially around identity and aspiration—is powerful. One of my ongoing research projects focuses building a conceptual framework to better understand what “wellness” means to consumers, and the implications of using the term in marketing and consumer contexts. While wellness can refer to something deeper: “the active pursuit of activities, choices and lifestyles that lead to a state of holistic health”—as the Global Wellness Institute defines it, I am intentionally referring to it through the lens of the modern consumer and cultural trend.

At the festival, one attendee told me he wasn’t sure what “wellness” really meant—only that it seemed to apply to everything. Still, he added “wellness” to the name of his real estate business because it sounded appealing. My physical therapist did the same when launching her practice. These examples aren’t unusual. But ideally, after engaging with our research, people might think twice before casually adopting “wellness” as a catch-all label. To draw clearer distinctions, we use “health” as a point of comparison. While “health and wellness” are frequently bundled together in both conversation and branding, we find that consumers perceive them as related but meaningfully different.

To demonstrate how these terms showed up in the festival, I downloaded the full transcript of the Well Festival from YouTube and ran a word count analysis. The terms “health,” “wellness,” “health and wellness,” and “well-being” were used throughout—sometimes interchangeably, but often with subtle differences. Well appeared 130 times, wellness 23, well-being 14, health 52, healthy 16, and healthier 7. In her opening remarks, Lori Leibovich referred to “in-person conversations with some of the most thoughtful minds in health, science, and wellness”—grouping the terms broadly. But in the sessions themselves, they were often used with more specificity: wellness appeared in phrases like “multiple dimensions of wellness” and “mental wellness,” well-being in references to a “cornerstone of well-being,” and health—such as “physical and mental health”—when seeking “expert advice.”

Based on my research, I would define wellness as broad in scope (e.g., encompassing mind, body, and spirit), lower in stakes (focused on small, incremental changes), and often tied to enjoyable or emotionally resonant experiences. Health, in contrast, is more narrowly focused on physical conditions—often illness or disease—and lacks the same intrinsic pleasure to consume. Health is more likely to appear in clinical or scientific contexts. Well-being is a broader, more neutral term that refers to how someone is doing in general or within a specific domain—emotional, physical, social, financial—and is less clearly tied to consumerism.

My best guess is that the Well brand intentionally avoids the more consumerist, buzzy term wellness in favor of the more objective and end-state–oriented well-being. That would help explain the festival’s focus on journalism and information (rather than products and marketing), and the inclusion of medical themes—like HIV prevention from Gilead Sciences or Peter Attia’s research on longevity—alongside conversations about lifestyle, community, and happiness. To more clearly apply our framework to the festival, I’ll go through examples of each of the dimensions.

The Festival Covered a Breadth of Topics

The wide range of topics discussed at the festival maps closely onto the first dimension our framework for understanding wellness. Unsurprisingly, mental well-being emerged as a recurring theme across sessions. Academic researchers like Dacher Keltner and Robert Waldinger spoke about happiness and subjective well-being, while media figures and celebrities—including Dan Harris, Sara Bareilles, Charlamagne Tha God, and Dwyane Wade—shared personal experiences with anxiety, depression, or panic attacks. Therapists such as Terry Real and Orna Guralnik focused on relationships, while fitness instructors like Robin Arzón, and Kelly McGonigal emphasized exercise as a key driver of mental resilience. Other panels tackled beauty, aging, and longevity (Paulina Porizkova, Jameela Jamil, Peter Attia), and, of course, food was on the table as well—most notably in discussion with Samin Nosrat.

It was a lot to digest in just eight hours (that pun was also intended). To give a sense of thematic emphasis, here’s a snapshot of the most frequently used words across the transcript: Life (177), feel(s) (134), therapy (70), anxiety (51), relationships(48), social (41), food (41), meditation (29), exercise (26), mental (37), feeling(s)(40), physical (15), sleep (22), beauty (19), diet (19), and nutrition (8). The emotionally and psychologically rooted vocabulary underscores how wellness is increasingly framed as satisfying holistic needs.

People Often Referred to Making Small Changes

Supporting our framework’s second dimension—low stakes—many speakers framed wellness as being about the little things: small shifts, daily rituals, or fleeting moments of self-awareness. Samin Nosrat offered a thoughtful response when asked how to support people with fewer resources or less access to wellness knowledge. She drew a distinction between large-scale public health solutions and individual wellness behaviors:

“To me, that's a social safety net question. Those are long-term, big-picture answers to how to help people have better lives… things way beyond what I can offer. But also, I don't think that the only way to have a food ritual is a beautiful weekly meal with your friends in their gorgeous backyard… I think it's just something as small as time and memory and food together.”

Jameela Jamil struck a more critical but grounded note:

“Wellness… was supposed to be a bit of self-care and a bit of self-pampering… I think small touches, you know, of something that makes you feel good—but not something that is being turned into an imperative that you're supposed to do.”

Both perspectives push back on the idea that wellness needs to be about making big changes – that can be left for health interventions. It can be simple, even mundane. And importantly, it shouldn't overpromise. Sometimes, wellness is just a moment of calm—or a warm meal with a friend. And sometimes, yes, it veers toward the fluffy and self-indulgent. Which leads to the final dimension in the framework.

Joy and Pleasure Were Discussed A Lot

Regarding to our third dimension—positive experience— wellness isn’t always about satisfying high level needs, but sometimes can entail feeling good in the moment. The emphasis on positive emotion during the festival reflects ties to the field of positive psychology. “Joy” was mentioned 53 times, “happiness” 37 times, and “pleasure” 19 times. Tracee Ellis Ross captured this nicely as, "Joy can come from so many different places... I think the wholeness of my life is where I find joy... my own self-care, noticing the flowers, noticing the trees." As did Kelly McGonigal: “The type of movement that has the most meaning and joy is dance... movement as something that can help us access the joy of life... using your body to engage with life." Together, these comments underscore another differentiator: wellness as a consumer endeavor isn't about putting in the disciplined work, but more about following intrinsic pleasures and passions.

Some Other Wellness Topics

Sharing Wellness Wisdom and Word-of-Mouth

In keeping with the idea that wellness often centers on small, incremental changes, practical suggestions and wisdom were featured throughout the festival. Two wall installations curated by the Well team offered tips pulled from the section’s archives. Attendees could select postcards with messages that resonated with them and were encouraged to mail them to friends.

Every speaker touched on a personal experience with poor health—most often mental health challenges like anxiety, depression, or panic attacks—and expressed a desire to share what had helped them in hopes of helping others. This resonates with what I’ve observed in both wellness professionals and consumers (myself included): it’s profoundly meaningful to share the tools that have shaped your own healing. Charlamagne Tha God encouraged people to literally “touch grass” and reflected on how his father’s silence about his own mental health had delayed his own process. Dan Harris and others repeatedly returned to the value of therapy and meditation. Robin Arzon, Kelly McGonigal, and Jameela Jamil emphasized the mental well-being benefits of exercise. Suleika Jaouad spoke about creativity through journaling, and Tracee Ellis Ross reflected on self-care rituals.

This spirit of sharing is closely tied to another area of ongoing research: the power of word-of-mouth in wellness. Recommendations from peers often carry more weight than those from typical marketing channels, especially in a category where trust and personal experience are central. Companies like AG1 and BetterHelp have effectively used podcast advertising to simulate the feel of a trusted referral. Against that backdrop, it makes perfect sense that Well would seek to deepen its brand community through live, experience-based events. I’m curious to see how this festival evolves in future years—and whether more media and wellness brands follow suit.

Philosophies of the Body

Is the body a temple? Machine? Vehicle for pleasure? My dissertation explores how people use metaphors to describe their bodies—and how those metaphors shape the ways they care for themselves. These embodied philosophies reflect deeper psychological beliefs and show up in consumer behavior across health, wellness, and fitness. They also connect closely to body image, a recurring topic at the festival—especially given wellness’s (often problematic) entanglement with diet culture.

Many speakers touched on the importance of maintaining a healthy body image, with conversations frequently emphasizing function over appearance. Research supports this reframe: individuals who adopt a more functional body image—focusing on what their bodies do rather than how they look—often report greater well-being. Especially for women. I was struck by how often this theme came up across sessions. Tracee Ellis Ross offered one of the more practical, grounded takes: “My body's point is to support my dreams and my life—and carrying groceries.”

Others leaned into more hedonic, pleasure-centered views. Kelly McGonigal spoke of movement as a source of joy: “The type of movement that has the most meaning and joy is dance… using your body to engage with life.”

Jameela Jamil agreed emphatically: “Like, is your body a vehicle for pleasure? Oh yes.”

Ross, returning to the topic later, spoke about her body as a site of comfort and presence—especially with age: “I feel the most in my body I've ever been. One of the things I discovered for myself is that my favorite place to be is present in my body. My least favorite place is not in my body—because I can't be in the experience of what’s happening in my life.”

Psychological research has found similar patterns: as people age, they tend to become more attuned to their bodies and more present in their physical experience—coinciding with lower rates of mind-wandering. Given my interest in mindfulness and wellness, it probably comes as no surprise that I find mind-wandering fascinating. (I’ll likely write more about that in a future post.)

Of course, not all views of the body were about joy or presence. Some were unapologetically mechanical. Peter Attia, for example—known for his work on longevity and health optimization—embodies the machine metaphor. During the festival, he described a two-day battery of tests his patients undergo:

“We do an assessment on patients that takes about two days. It’s a pretty robust physical assessment… We test everything we want to understand about them—markers of aerobic fitness, peak aerobic fitness, base aerobic fitness… Based on your assessment and your goals, we identify where the gaps are. And then you train for those gaps.”

There’s still more work to be done to understand how these contrasting metaphors influence how people pursue wellness. But I suspect they’re not mutually exclusive and many of us might shift fluidly between them across contexts and eras of our lives.

Where are all the men?

Gender came up frequently throughout the festival. The word “women” appeared 54 times in the transcript, while “men” showed up 37 times. But the more striking gender gap was visible in the crowd itself: wellness remains a largely female-dominated space. I was surprised by how few other men were in attendance—a reminder of wellness’s persistent gender skew. At one point, Dwyane Wade asked the men in the audience to raise their hands if they’d ever been to a wellness retreat. Only a few hands went up. He and Charlamagne Tha God spoke candidly about the importance of men talking to other men and seeking out mental health support. Notably, the Well editorial team also appeared to be all (or nearly all) women.

There’s plenty of research showing that men are less likely than women to seek out mental health care, despite clear need. Barriers include stigma and the perception that mental health products and practices—like therapy, journaling, or meditation—are “feminine.” Data from a popular meditation app company I’m collaborating with confirms this: male engagement consistently lags behind. What I hadn’t fully appreciated until this event is that this gap extends beyond mental health and into wellness more broadly.

Of course, the consequences of this are well documented. Men, on average, die earlier than women—often due to preventable causes. A recent SNL skit nailed this, poking fun at how men need to be tricked into going to the doctor—by disguising the visit as an appearance on a podcast like Huberman Lab or Joe Rogan. There’s a massive untapped wellness market among men—and you can bet that marketing teams are actively trying to tap into it. Philosophies of body could be one approach, with evidence that the body as a temple tends to carry feminine associations and take better care of the body, while the machine metaphor skews more masculine and consumes less for wellness.

Science and Spirituality

One of the Well’s most distinct points of differentiation is its emphasis on factual, evidence-based reporting. This was made explicit in signage at the festival entrance:“At a time of rampant misinformation about health, the Well desk strives to give readers something different: Practical, useful coverage. Rigorously fact-checked. Built on trust, clarity, context, and independence.” That editorial stance shaped the tone of the event. The Well consistently pushes for scientific accountability, particularly around pseudoscientific or scam health claims. One memorable moment came during a session where moderator Kate Lowenstein repeatedly challenged Peter Attia on the evidence behind protein recommendations. His work embodied a thoroughly scientific, mechanical view of the body. There’s little space for mysticism or the beyond in a worldview built around diagnostics and training gaps. This is in stark contrast to some of the wellness content we often see online, where the line between backed-by-science and belief-based can be blurry. Compared to other festivals like Eudemonia, which seemed to lean more spiritual or speculative, the Well clearly stakes out a more journalistic, evidence-first position. Spirituality is a contributor to viewing the body as a temple versus a machine.

But science and spirituality aren’t necessarily at odds. One of the most compelling speakers at the festival, Robert Waldinger, blends both. A Zen priest and lead researcher of the Harvard Study of Adult Development, Waldinger’s philosophical message tied in closely to his research on happiness and relationships:

“One of the things that Zen teaches—and many spiritual traditions remind us of—is that this is all pretty fleeting. What really matters is what we do in the world and our connections to each other.”

Politics of Wellness

I had an interesting conversation with a politics reporter (while he was waiting to meet his hometown hero, Dwyane Wade), giving me new perspective about political ideology and wellness. I had assumed wellness was more of a liberal concept. The original wellness movement of the 60s and 70s appealed more to counter-cultural leftists looking for an alternative to mainstream medical care. However, there is a new wave of wellness seekers from the establishment-distrusting, RFK supporting conservative right, with new interest in wellness (think Make America Healthy Again).

This is suggestive of the idea of a political horseshoe, instead of liberal-conservative continuum. These extreme left and right groups are likely united by a desire for body purity, interest in “naturalness,” and distrust of mainstream medicine. There may be underlying differences in each of these groups’ wellness approaches that I’d be curious to better understand. I can’t see conservative wellness consumers identity signaling about psychotherapy on a dating app. But maybe these differences exist more towards the center, and not so much at the extreme ideologies.

The Suspect Side of Wellness

While the tone of the festival—among both speakers and attendees—was overwhelmingly positive, it’s important to acknowledge that the wellness consumer trend isn’t all loving-kindness, rainbows, and skin-boosting supplements.

In her opening remarks, Joy Robins, the New York Times’ Chief Advertising Officer, acknowledged the pressures of modern life and declared, “We need more wellness.” That sentiment is hard to argue with. Most of us could probably live healthier, more fulfilling lives. But as a consumer industry, wellness isn’t always benign—and it isn’t for everyone. It can be expensive, exclusionary, and aspirational to the point of being alienating. As the space continues to grow, there’s important work to be done around making health and well-being more accessible, inclusive, and educational—not just marketable.

That tension was addressed head-on by Jameela Jamil, who offered one of the more critical views presented at the festival:

“We've taken wellness, which was supposed to be a bit of self-care and a bit of self-pampering, and turned it into another way for women to judge themselves and judge each other… I just think wellness needs to stay away from diet culture, which it is creeping endlessly into. And wellness needs to actually be about longevity—not just a panic toward the standard that we're destroying our skin barriers for. There's too many products now. There's too much going on.”

Her concerns echo a growing fatigue in the wellness space. The market is oversaturated, and the sheer volume of tips, tools, and products—each promising transformation—can feel more overwhelming than empowering. Perhaps sensing this, Dan Harris offered a grounded reminder: sometimes the best approach isn’t to follow every recommendation, but to “see for yourself.” It’s a good note to end on. Because as the wellness industry continues to evolve, what matters most might not be the next product, panel, or protocol—but the practices that feel real, meaningful, and sustainable in your own life.

In Conclusion

There’s a lot happening in the wellness space right now—and if you’ve made it to the end of this first piece, chances are you’re just as intrigued as I am. I see the Well Festival as one of many emerging efforts to build community and foster experiential engagement from major players in the industry. Don’t be surprised if we soon see similar events from brands like Peloton, Oura Ring, Lululemon, Nike, or Equinox.

Looking ahead, I’m especially interested in a few trends: the integration of GLP-1 medications into wellness lifestyles, the rise of body metrics tracking as wearable tech becomes more advanced – like the new Whoop MG (Medical Grade), and the role AI might play in reshaping therapy and mental health care (it was just announced as the top use case for LLMs).

I’d love to hear what’s capturing your attention in this space and I’m always happy to connect and continue the conversation.