America spent $4.8 trillion on healthcare In 2023. If that were a nation’s GDP, it would be the third largest on Earth. No wonder everyone—and so many heathtech startups—aims to fix it.

But since my company Zocdoc launched in 2007, I have watched waves of self-appointed “disruptors” enter the healthcare arena like lions only to retreat like lambs. In fact, 90% of healthtech startups have gone bust.

It’s not just the upstarts. Healthcare has also burned Amazon, Apple, Google, Microsoft, and Walmart. What chance does disruption have in the healthtech sector that defeats David and disgraces Goliath?

But America’s healthcare system needs disruption as costs rise and outcomes decline. Seventy-three percent of American adults say the system is failing them in some way. Without intervention, this will break the bank, our health, or both.

Given the stakes, we must remain bullish on technology’s ability to improve healthcare cost, quality, and access. But instead of blindly adopting “disruptor” playbooks that worked in other consumer sectors, healthtech entrants must study the industry’s past failures. While each has its nuances, there are instructive commonalities and landmines. I have distilled them down to these seven.

Fragmentation

In the U.S., 340,000 physician practices run on hundreds of electronic health record systems, and 1,100 payors operate thousands of plans. Complicating matters, processes for scheduling, insurance eligibility, pricing, revenue cycle management, authorization, referrals, and hundreds of other workflows are spread across disparate systems or locked up in walled gardens.

Many disruptors who enter this environment think of healthcare as a greenfield when it is the brownest of fields. They ignore the trillions of dollars of entrenched yet disconnected infrastructure and believe that everyone will magically switch to their built-from-whole-cloth interfaces, systems, services, or processes that do not connect to the world as it is.

Adoption

If you build it, and they like it, it might take them five years to buy it. And even longer to adopt it broadly.

Payors, providers, and patients are all stuck in their current systems, and inertia and sunk costs are powerful countervailing forces. Motivating change at scale requires time, resources, multi-stakeholder process management, a strong value proposition, compelling evidence, and exceptionally long sales cycles. These are investment-intensive efforts for lean startups and large companies alike, and driving adoption at scale takes time.

Expectation

The healthcare industry measures successful outcomes over decades—or in some cases, patients’ lifespans—and not in financial quarters. However, venture investors and shareholders like to see near-term returns on their investments. The mismatch in time horizons often pressures healthtech startups to cut critical corners, unsustainably buy their way to growth, or run before they have product-market fit.

Intermediation

Healthtech companies inevitably encounter the principal-agent problem: Patients, providers, payors, and employers all participate in care decisions but they have differing roles and incentives.

Patients are the eventual healthcare consumers, making choices with little insight into quality or costs. Providers are negotiating prices with payors. Payors are negotiating to acquire patients from employers. And employers are stuck trying to stem the bleeding of rising costs. Everyone is on the hook for someone else’s decisions.

Disaster strikes when disruptors don’t respect or account for the reality of these disparate stakeholders and incentives. At best, they uncover misalignment far too late, leading to hard pivots or speculative quests for viable customers or payment models. At worst, they have hidden forces working against them when incentives are at odds.

Regulation

Moving fast and breaking things does not work in healthcare because people’s lives and taxpayer dollars are at stake. Healthtech tartups that break, ignore, or misinterpret state or federal laws risk hefty fines and criminal investigations.

Others struggle to retrofit viable business models into outmoded regulations. This nearly killed Zocdoc. Due to the ambiguity in a federal statute that was signed into law before the advent of the internet, we were unclear on whether charging providers for new patient bookings was permissible. And so we initially charged every provider the same flat subscription fee to be listed on our marketplace.

But there is a good reason no other consumer marketplace—like Priceline for travel, OpenTable for restaurants, or Airbnb for homestays—charges its supply-side customers the same flat fee: The price is almost always wrong. As a result, doctors were leaving our platform faster than we could sign new ones up.

This made it impossible for Zocdoc to grow. By 2017, we understood that the only way to save Zocdoc was to charge providers per transaction like every other marketplace. But due to the law’s ambiguity, we spent years securing permission from regulators to do so. From there, we embarked on a multiyear effort to perform open heart surgery on Zocdoc and transition to our current bookings-based business model. We barely survived this harrowing journey, but we emerged stronger on the other side.

Productization

Disruptors prone to wielding technological hammers tend to see every healthcare problem as a nail. So tech-focused founders aim to productize healthcare delivery in an attempt to make it scale.

This is the correct playbook in industries where the customer can learn your product, like converting travel agents into apps or making simple tax filing self-service. This is not the case in healthcare. The complexity, uniqueness, and skilled judgment required for each clinical encounter, plus the situational circumstances—whether the patient is healthy, sick, or unconscious—make it untenable for patients or products to assume the role of a physician.

As a former physician who now works in healthtech, I believe that the industry is a long way from replacing doctors with software. As for AI, for now, it is better suited to drive efficiencies surrounding the clinical experience than to replace or “productize” clinicians themselves.

Localization

Healthcare will always be a hands-on industry. While COVID-19 was a watershed moment for telehealth, patients still overwhelmingly prefer in-person care. Excluding mental health, 92% of Zocdoc’s bookings in 2023 were for in-person appointments. This means companies delivering healthcare must think physically and locally. They must be ready to offer ample in-network services within mere miles of patients’ offices or homes.

Establishing a local presence is possible. It is also costly and risky. The physical assets of hospitals across the U.S. are collectively worth about $670 billion. Yet even Walmart, which benefited from an extensive brick-and-mortar footprint, closed its health clinics. Establishing a fraction of Walmart’s physical footprint requires significant capital, sending disruptors back to the “expectation” landmine.

Less disruption, more pragmatism

I have never liked the term “disruption” in healthcare. Not only has it been soiled by self-described disruptors’ lack of traction, but it is also decisively ineffective. If a single success story in healthcare has come from a disruptive “move fast and break things” approach, I must have missed it.

Healthcare needs less outside-in disruption and more inside-out pragmatism. Healthtech startups vying to change healthcare must accept the system as it exists today and work to connect it and improve it, bit by bit. There is no fast pass. There are no silver bullets. There is slow, painstaking work to deeply understand healthcare—its incumbents, incentives, irrationalities, and idiosyncrasies—to connect all the disparate dots and make it better one step at a time.

Every aspect of America’s healthcare system can be improved by technology. But to realize this potential, those of us building in the healthtech space must first accept and account for its seven deadly landmines.