Everything at the Touch of a Button
Imagine a subject in a room with food, water, and two mysterious buttons. The subject sticks only to food and water at first but eventually starts pressing the buttons to see what happens. The first button allows the subject to drink unknown liquid #1 from a fountain, while the second button injects unknown liquid #2 through a tube in the subject’s side. Over time, the subject is pushing the buttons habitually.
“Beyond the study walls, we now have buttons to access food from our phones at all times, whether restaurant ordering or grocery shopping, and we’re in crisis mode with human metabolic health. While we can engineer healthier food, that is all for naught without helping humans make better decisions. It’s time to think about how we fix the button.”
What if I told you that unknown liquid #1 is sugar solution, and unknown liquid #2 is cocaine solution? Which button do you think is pressed more?
As it turns out, the subject presses the sugar button far more.
This is a real study on lab rats, demonstrating that addiction to sugar in this environment far outweighs cocaine. Sure, the rats got addicted to both, but there was a consistent and large preference for sugar over cocaine. Remember - this wasn’t for lack of food - it was simply more addictive. 
Beyond the study walls, we now have buttons to access food from our phones at all times, whether restaurant ordering or grocery shopping, and we’re in crisis mode with human metabolic health. While we can engineer healthier food, that is all for naught without helping humans make better decisions. It’s time to think about how we fix the button.
Scarcity to Abundance
When do industries move from scarcity to abundance? When production becomes easy. Any one of the technological capital-R “Revolutions” you learn about in school is a movement from scarcity to abundance: Agricultural Revolution, Industrial Revolution, etc. You could say we’re going through a lowercase-r revolution in generative artificial intelligence now (and maybe it will prove to be capital-R down the line). The point is that a breakthrough occurs in production power and something that was once hard to produce becomes relatively easy (food in the Agricultural Revolution, iron during the Industrial Revolution, digital images during the generative AI revolution).
During the capital-R “Revolutions”, the business models that emerged on top of newfound production power capitalized on the massive demand for goods that were once hard to supply. In recent history, such businesses emerged not only from producing new goods but unlocking existing supply - for example, creating value from existing apartments and cars when not in use. Moving from scarcity to abundance generally leads to commoditization to some degree, but some industries are more resistant than others.
Two definitions are needed to clarify this stage:
Commoditized supply exists when you don’t care if the good you’re purchasing was made by company A or company B, because the good is the same - like corn or gold.
Differentiated supply exists when you do care which supplier made the good you’re purchasing. You can have differentiated supply because the nature of the industry itself means different suppliers produce different outcomes - like music or restaurants. Or because there is a perceived brand difference despite fundamentally identical goods - like pharmaceuticals or laundry detergent.
In today’s internet business landscape, category-defining companies such as Airbnb and Uber emerged as marketplaces connecting buyers and sellers. Marketplaces make sense for commoditized supply: you do not monitor your top hosts for new listings or wait for your favorite driver to be available, but you do tune in for the new Taylor Swift album at midnight.
Airbnb host differentiation only matters in shared spaces and experiences, when the host themself is involved. Entire space rentals, which turned out to be the true success driver, are indeed commoditized - housing location and quality matter, but not who supplies it. Commoditization accelerated as the supply side professionalized - most “hosts” shouldn’t be called hosts at all - they aren’t sharing their homes but buying property for the explicit purpose of renting. So Airbnb increasingly commoditized the supply they originally unlocked by favoring what the customer actually valued in their marketplace.
For Uber, their vision and commoditization are more straightforward: getting from point A to point B safely and on time is all that matters, irrespective of who is supplying it to you. Uber had it right from the start in understanding what their value was, while Lyft tried to focus on differentiated supply (their early app encouraged you to sit in the front seat and make conversation) before realizing that’s not really why people were paying them.
Marketplaces made for durable businesses in these examples - so when do industries evolve past marketplace abundance? When supply is differentiated and distribution becomes easy.
Abundance to Curation
Hal Varian, chief economist at Google, said that the best way to forecast the future is to look at what rich people do and scale it. 
Rich people use marketplaces at a similar rate to the average consumer when supply is commoditized. When they need to get from point A to point B, they order an Uber. Why wouldn’t they? Some still have personal drivers, but a lot less do than 10 years ago.
But I suspect that rich people use Doordash at a lower rate than the average consumer (I don’t have data on this, but let me know if you do). They tend to value health and food quality enough to hire a personal chef who knows their tastes and needs. Or maybe they pay for Doordash, but it’s their assistant browsing and placing orders for the right meals. This is common at large corporations, where executives are on a meal plan separate from the rest of the employees (even if it’s good food - it might not be right), and someone serves as their surrogate in curating the optimal daily food. Anna Wintour’s go-to work lunch at One World Trade cost $77.33 - but it was keto! 
When supply is commoditized, a marketplace is optimal. When supply is differentiated, moving beyond abundance to curation is valuable and people will spend money to do it.
So why haven’t food businesses moved into curation mode yet? Because distribution is still hard, as the final product is a physical good.
What does it look like when distribution is easy?
The best examples for easy distribution come from the late stages of another capital-R: the Digital Revolution, when anything that could be digitized, was digitized. Starting with business applications but eventually swallowing massive consumer industries like music and film, generation-defining companies like Spotify and Netflix emerged.
Step beyond why streaming players beat the legacy options (CDs and DVDs), which is quite straightforward: scarcity to abundance. What differentiated them from other digital counterparts? Why did Spotify beat iTunes, and why did Netflix beat cable “Movies on Demand”? Abundance to curation. Improved user interfaces, surely, but that’s part of a highly curated experience - as everything is right where you need it to be.
With differentiated supply and endless optionality, it became important to understand the end consumer and curate to their needs. Distributing music and movies had no marginal cost, so the mavericks understood that they would not win just by serving up everything (which anyone with sufficient pockets could do), but by serving up the right goods at the right time. For both Spotify and Netflix, it was an increasingly surreal experience, where you were guided to relevant undiscovered songs and shows based on what you’d enjoyed in the past.
Part of the magic was the arrangement, where you and your friend had different playlists and track orders and autoplay despite the same catalog. But the catalog itself also changed over time to better reflect evolutions in taste. Pop songs are shorter and hooks come sooner than ever before. Netflix spun up an entire production studio to further capitalize on understanding what the consumer wants in films and shows.
There is still a cost to producing music and film, but it is fixed, not variable. You write and produce and record and edit, and then it’s done. Once the underlying good has been digitized and obtained, the cost of distribution is negligible, so Spotify and Netflix derive their primary value from curation.
The business model is beautiful because the goods are digital, not physical. So let’s dig into some messier problems, where distribution is not easy.
Atoms to Bits and Solving Hard Distribution
All businesses fall along a spectrum of atoms to bits, where they capture value between purely physical goods and services (“atoms”) and software (“bits”). Going from “atoms to bits” is an efficient move that gives a business more leverage. While Spotify and Netflix (“bits businesses”) can digitize their core offering entirely and make distribution easy, food can never be a purely bits business, nor can healthcare, gyms, or plumbing (“atoms businesses”).
Bits businesses with differentiated supply naturally evolved from marketplaces to curation. Spotify and Netflix started with pure digitization (easy distribution) and evolved into curation engines by understanding what people like and surfacing it.
Food marketplaces put every option in front of every person and put the burden of sifting through abundance on the user. The result is incredible inefficiency, with massive food waste, chronic poor health outcomes, and restaurant failures - among many other problems*. The paradox is that a natural evolution to curation hasn’t occurred because distribution is hard, but that building curation could solve hard distribution. By understanding what people want and need and surfacing the right options, order predictability increases and distribution eases.
In fact, there is more impact in curating atoms than bits: the error of getting the wrong song is irrelevant, but the error of getting the wrong meal is frustrating, and the error of getting the wrong meal daily shortens your life.
What would it look like to prioritize curation in food?
Start by looking at the existing options.
Legacy food suppliers (Compass Group, Aramark, Sodexo - serving universities, hospitals, etc.) push out eight billion meals per year, essentially twisting the same low-cost and low-variability ingredient base into a variety of options.
Newer food marketplaces (Doordash, Uber Eats - serving consumers directly) inevitably spotlight chains that can spend more on ads (just open the apps to verify). Most users settle on five or so restaurants that they cycle through , so the first options seen often become their lifetime selections, thereby negating the marketplace’s potential for individualized health discovery. In effect, food options are commoditized to a low-cost and low-variability ingredient base of chain offerings, despite a long tail of suppliers with true differentiation.
None of these businesses help consumers capitalize on their personalized nutrition needs, despite three out of four consumers seeking such a service.  Current food access options have disastrous effects: 75% of US healthcare costs (a whopping $3.2 trillion) result from food-responsive metabolic conditions , and a staggering 88% of Americans are metabolically unhealthy.  The list goes on.
Curating food at scale requires shifting atoms to bits by digitizing the consumer journey. The experience of food culminates with what you put in your body, but it is a journey of health, culture, seasonality, locality, and so much more. Creating user interfaces and data structures that respect a person’s unique metabolic needs and a supplier’s differentiated capabilities is the foundation for guiding people down a healthful and delightful path. With that foundation in place, distribution eases: users’ likes and needs can be propagated back to differentiated suppliers as predictive analytics on what orders can be expected when and where. The effect continues upstream to increase predictability throughout the food supply chain, from ingredient transportation to the seeds planted in the ground.
While I’m focused on food, many areas are primed for long-term curation building, like healthcare, child education, and corporate advancement.
We’re not rats in a cage, but it sure feels like it sometimes. The next generation-defining businesses won’t win by getting people to press buttons on momentary urges, but by serving as trusted partners down a meaningful path of individualized discovery.
Interested in partnering, providing feedback, or learning more? Email email@example.com
*I’d be remiss not to mention that food abundance is not universal. Many people do not have adequate access to food (a crisis that must be solved), but the increasing population that has abundant access faces the challenges we’re addressing here.