There is a version of “start small” that is timid — a retreat from ambition dressed up as prudence. And there is a version that is strategic — a deliberate compression of scope designed to generate the learning that makes large-scale success possible. Oak Street Health practiced the second version with remarkable discipline, and the result was a $10.6 billion acquisition by CVS Health in 2023.
The lesson embedded in their story is not about healthcare. It is about how innovative companies grow.
The Problem They Set Out to Solve
Medicare patients in underserved communities have historically received fragmented, reactive primary care. Fee-for-service medicine creates a structural incentive to treat illness rather than prevent it; providers are paid per encounter, not per outcome. The sicker the patient, the more appointments, the more revenue. The system is not malicious; it is misaligned.
Oak Street Health, founded in 2012 by Mike Pykosz and a small team in Chicago, decided to build a different model from first principles. They would focus exclusively on Medicare patients, operate under a value-based care arrangement (meaning they were paid to keep patients healthy, not to bill for visits), and use data and technology to personalize care at scale. The economics only worked if patients stayed healthy; the model was structurally aligned with the outcome.
None of that was new intellectually. Value-based care had been discussed for years. What was new was the systematic, operationally rigorous way Oak Street intended to execute it.
One Clinic
They started with a single clinic on the north side of Chicago.
This is worth dwelling on. Not a platform. Not a network. Not a national rollout. One clinic; one patient population; one operating team learning how to make the model work in the real world, with real patients who had real complexity.
The temptation in any disruptive innovation is to scale before the model is proven. Capital is available, the idea feels right, and urgency seems to demand speed. Oak Street resisted. They used that first clinic as a living laboratory: every patient interaction, every care protocol, every operational decision generated data. Which patients were at highest risk for hospitalization? Which interventions reduced emergency room visits? Which staffing models produced the best patient satisfaction? Which appointment cadences correlated with better chronic disease management?
The first clinic was not just a revenue source; it was an instrument.
The Adjacent Possible Flywheel
Biologist Stuart Kauffman originally used the term “adjacent possible” to describe how biological systems evolve: at any given moment, the next step in development is constrained by and enabled by the current state. You can only reach what is one step away from where you already are; but each step you take opens new adjacencies that were previously unreachable.
Applied to business innovation, the Adjacent Possible Flywheel describes a specific growth dynamic: each new capability you build makes the next capability achievable. Each new market you enter gives you data that improves your model for the next market. Each new customer funds the infrastructure that makes the next customer cheaper to serve.
Oak Street Health ran this flywheel with unusual clarity. The sequence looked something like this:
Clinic 1 generates operating data. Patient outcomes, cost structures, staffing ratios, technology performance, care protocol effectiveness — all of it becomes institutional knowledge. The model improves.
Improved model reduces risk for Clinic 2. The second clinic opens with better protocols, better technology, and better trained staff than the first. The adjacency has been expanded; the unknown territory is smaller.
Clinics 2 and 3 generate more data at higher volume. New patient populations, new geographies, new operational challenges — all producing new learning that feeds back into the model.
Scale attracts better payer contracts. As Oak Street demonstrated outcomes — lower hospitalization rates, better chronic disease management, lower total cost of care — they gained negotiating leverage with Medicare Advantage plans. Better contracts improved unit economics, which funded more clinics.
More clinics accelerate technology development. More data means more accurate risk stratification. More accurate risk stratification means better-targeted interventions. Better interventions mean better outcomes. Better outcomes mean better contracts. The loop compounds.
Most companies think about iteration on a fixed landscape: faster cycles, tighter feedback loops, reduced waste. Oak Street was doing something categorically different: each clinic didn’t just improve execution, it expanded the landscape of what was possible next. Their horizon itself was growing.
By the time Oak Street was acquired by CVS, they had over 160 clinics in more than 20 states and were serving more than 200,000 patients. They did not get there by executing a national rollout from day one; they got there by running the flywheel, one turn at a time.
Technology as the Engine, Not the Product
One of the most instructive aspects of Oak Street’s model is how they treated technology. They were not a technology company that happened to deliver healthcare; they were a healthcare company that used technology as an operational lever.
Their proprietary platform, Canopy, aggregated patient data from multiple sources and translated it into actionable care protocols. Which patients needed outreach today? Which were at elevated risk for a hospitalization in the next 30 days? Which had gaps in preventive care that correlated with downstream cost? The technology answered those questions at scale, allowing a relatively small care team to manage a large, complex patient population proactively rather than reactively.
This distinction matters enormously for any mid-market company evaluating its own AI and technology investments. The question is not “what technology should we deploy?” The question is “what operational bottleneck, if removed, would unlock the next stage of our flywheel?” For Oak Street, that bottleneck was the ability to identify high-risk patients before they became high-cost patients. Canopy was built to solve that specific problem; everything else followed from it.
Repeatable Innovation as a System
The word “repeatable” is doing significant work in this story. Anyone can open a second clinic. The difficult thing is ensuring that each new clinic benefits from everything the previous clinics learned, while also generating new learning that feeds back into the network.
Oak Street built that feedback mechanism deliberately. New clinic openings followed a standardized playbook: market selection criteria, community outreach protocols, hiring profiles, technology deployment, payer contracting timelines, and performance benchmarks for the first 18 months of operation. But the playbook was not static; it was updated continuously as new data from operating clinics surfaced patterns.
This is the difference between replication and systematic innovation. Replication copies what worked before. Systematic innovation copies what worked before, measures what didn’t, and modifies the next iteration accordingly. The result is a learning system that gets better at execution with each new deployment.
For a mid-market company, this is the most transferable insight. Whether you are opening new territories, launching new product lines, or expanding to new customer segments, the infrastructure for capturing and applying what you learn is at least as important as the quality of your initial idea. The companies that build learning loops into their growth architecture tend to compound; the companies that treat each expansion as a standalone initiative tend to plateau.
Why CVS Paid $10.6 Billion
The acquisition price is not incidental to the story; it is the story’s punctuation mark.
CVS did not pay $10.6 billion for 160 clinics. They paid for a proven, data-driven, scalable model for delivering value-based primary care to Medicare patients; a proprietary technology platform with years of outcome data embedded in it; and demonstrated unit economics that CVS could leverage across their existing retail footprint.
In other words, they paid for the flywheel. The individual clinics were simply evidence that the flywheel was real.
This is the monetization thesis that mid-market companies often underweight. The question is not just “how much revenue does this line of business generate?” The question is “what does this line of business make us capable of that we were not capable of before?” Oak Street’s clinics were revenue-generating; they were also data-generating, capability-building, model-validating assets. Each one increased the value of the network.
The acquirer saw the whole system, not the individual components.
The Lesson for Companies That Are Not in Healthcare
Oak Street Health’s model was specific to Medicare primary care. The flywheel is not.
Start with a bounded experiment. One market, one customer segment, one product variant. Small enough to learn from; large enough to generate meaningful data. The goal of the first iteration is not success at scale; it is learning at low cost.
Build the feedback infrastructure before you need it. The companies that scale effectively are the ones that capture what they learn during early experiments and build it into the model for the next iteration. This requires deliberate data architecture; not just operational reporting, but the systematic question: “what would change how we run the next deployment?”
Identify the technology that unlocks the next flywheel turn. Not technology for its own sake; technology targeted at the specific bottleneck that limits the model’s performance. For Oak Street, it was risk stratification. For your business, it is something different; but the discipline of identifying it specifically is the same.
Let outcomes create leverage. Oak Street’s demonstrated results gave them negotiating leverage with payers. Demonstrated results in your domain — customer retention, margin improvement, time-to-value, defect reduction — create leverage with prospects, partners, and capital sources. Build the measurement infrastructure that makes your results credible and visible.
Value the flywheel, not just the revenue. Each stage of Oak Street’s growth made the subsequent stages more valuable, more defensible, and more attractive to acquirers. The same logic applies to any business building a systematic, data-reinforced growth model. The compound effect is only visible if you are tracking the right things.
A Final Note on Patience
There is a temptation to read the Oak Street story as a validation of “move fast.” It is not. Between the first Chicago clinic in 2012 and the CVS acquisition in 2023 was eleven years of disciplined, iterative execution. Eleven years of building the model, testing the model, improving the model, and expanding the model one turn of the flywheel at a time.
The starting point was small. The ambition was never small. That distinction is the whole lesson.
Mid-market companies often operate as though they must choose between disciplined execution and transformative outcomes. Oak Street Health is evidence that disciplined execution, applied systematically over time, is precisely how transformative outcomes get built. The horizon was always $10 billion; they just chose to walk toward it one clinic at a time.
