Did Amazon sink Haven?

Amazon's diffuse management strategy and policy of burning money in competition is a major root cause of Haven's downfall

In 2018, Amazon, Berkshire Hathaway, and JPMorgan Chase launched a secretive, splashy new venture to make over health care. Making it even buzzier, the three companies hired Dr. Atul Gawande, an endocrine surgeon at Brigham and Women’s Hospital in Boston and a writer—perhaps the most famous modern physician-writer—known for taking on big health care topics. They called the venture Haven.

This week, eight months after Gawande announced that he was leaving the enterprise, Haven collapsed.

Reactions from health care friends

What happened? There have already been several well-reported autopsies, including one from Chrissy Farr and another from Erin Brodwin and Casey Ross of STAT News, both of which I’m drawing from for this newsletter.

In short: Berkshire Hathaway and JPMorgan Chase seem guilty of benign neglect, and certainly of hubris in thinking they could bring smart employees together and automatically solve Health Care™️. But the part that seems like an overwhelming cause of death to me is Amazon’s approach.

Amazon’s strategy—a combination of its method of attaching employees to diffuse projects that don’t communicate with each other, and its willingness to burn money to win a race, including over other entities it’s funding—is what really destroyed Haven.

Haven’s failure points

In their reporting, Farr, Brodwin, and Ross outline a few theses for the collapse:

  • A lack of planning and coordination between the three entities

  • Competition with Amazon’s simultaneous internal efforts in health care

  • Gawande’s lack of experience as a manager

  • High employee turnover

In my opinion, these failure points are closely linked to Amazon’s business strategy.

A lack of planning and coordination

It appears that the three founding organizations—Amazon, Berkshire Hathaway, and JPMorgan Chase—never really formed a clear goal for Haven in the first place. This is a pretty common mistake, especially when successful businessmen without health care expertise decide to “fix” health care. (See below, when I talk more about Healtheon, the Haven of the 90s.)

But by not having a single, agreed-upon leverage point, Haven opened itself up to competition…from Amazon itself.

Competition with Amazon’s individual, simultaneous efforts in health care

While Haven was struggling to remake health care in some vague way, Amazon had several health care-focused teams working on internal projects. At least two of those internal projects have been made public.

First, in September 2019, Amazon Care launched for Seattle-based Amazon employees. Amazon Care is an employee health clinic with a heavy telehealth component. Its website implies it will eventually be extended to non-employees.

The corporation was simultaneously building out its pharmacy program; in November, it launched Amazon Pharmacy, two years after acquiring PillPack in 2018.

Given how confused Haven employees seemed to be around some of Amazon’s announcements (in this STAT article, one person notes that “there have been instances in which Amazon’s plans were not shared with the Haven group”), it appears there was little or no communication between Amazon’s internal health care team and the Haven team.

In fact, it appears that this was almost deliberate competition. A cynical interpretation is that Amazon’s head strategists decided to invest in a few different health care ventures and then burn money as they let those ventures battle for the top spot. Amazon’s internal efforts won (and notably, it didn’t share these efforts with Haven).

This strategy, if true, left Berkshire Hathaway and JPMorgan Chase in the dust. As Dr. Howard Forman, a professor of health care management at Yale, noted in a postmortem, “Two of the three companies only had a stake in this from the point of view of their employees. The third clearly had an eye on this being a service line. That was an issue from the beginning that you had to wonder about” (italics mine).

In fact, my most cynical interpretation is that Amazon jointly funded Haven with Berkshire Hathaway and JPMorgan Chase, but it never planned to see anything materialize. Instead, it used Haven almost as an R&D lab, and then internally developed the ideas with the most promise. (I’m holding out hope for a Gawande New Yorker article explaining everything.)

Gawande’s lack of experience as a manager

This, and the point below, both seem like symptoms of Amazon’s approach, indications that Haven was failing rather than root causes of failure. Nonetheless, Gawande’s appointment didn’t help. As Kimberly MacPherson, the executive director of health management at the Berkeley Haas School of Business, said, “Atul is a visionary. What he’s not is the CEO of an operating company that can do something different in health care.”

High employee turnover

Although I think this is also a symptom of Haven’s lack of overall strategy and impossible competition with Amazon, I would love to understand what facet of the internal atmosphere was causing high turnover, particularly on a project that’s stacked with luminaries and (presumably) so well-funded.

The only thing I can find that would contribute to this turnover, besides the ennui of working at an enterprise with no clear strategy and suffering from competition with Amazon, is that Haven was a not-for-profit. One anonymous employee told STAT that the not-for-profit structure made it difficult to retain high-level talent. But I suspect it was more the ennui.

The challenges of disrupting health care

Haven’s failure to achieve much of anything is particularly notable in light of the enormous success that many health tech firms (both startup and establishment) have achieved during the COVID-19 pandemic.

My working thesis on this is that there are two kinds of disruption that new health care ventures attempt (thank you to Marshall Kosloff for helping refine this):

  1. Overwhelming, large-scale ventures attempting to take on an entire industry

  2. Smaller ventures that focus on just one lever—which they use as a beachhead for eventually scaling up and disrupting the entire industry

The first type of venture fails much more often, particularly if the people involved have limited knowledge of or experience in the industry of focus.

For example, Facebook’s attempt to disrupt the education system fell totally flat. The venture didn’t seem to have much beyond its broad, general goal of “disrupting education,” and beyond being a little weird and potentially creepy, all they seem to have accomplished is making laptops a bigger part of the curriculum.

However, ventures that focus on solving just one piece of a larger problem can find success more easily.

For example, Airbnb remade the hotel industry, but it didn’t start that way. It started as a platform for people to find quick, cheap overnight stays without resorting to hotels. Once people adopted that part, they also began using Airbnb for weekend getaways and experiences. In the end, Airbnb became the most valuable overnight stay company on the market.

This is an especially important distinction in health care, which is—as most people know—almost a fifth of the U.S. GDP and perpetually growing. There are so many players, all so complicated and with so many misaligned incentives stacked atop one another, that large, very well-funded enterprises are drawn to the idea of taking the whole thing on.

Without a single leverage point in mind—by trying to remake the entire suite of non-emergency services that an employee could use—Haven failed to move the needle on anything and fell prey to competition from Amazon.

The Haven of the 90s: an illustrative example

Major efforts (and subsequent failures) like Haven have happened before.

In the mid 1990s, the founder of Netscape, Jim Clark, decided that because he had permanently changed the landscape of technology, he could do the same for health care.

He hired a lot of extraordinarily smart engineers, put them in a room, and named the venture Healtheon. Healtheon got funding from some of the top VCs in Silicon Valley, including Kleiner Perkins, and at first it seemed like they might have a shot.

But the engineers weren’t prepared for how complicated health care is and how inadequate tech band-aids could be. The team had limited health care expertise and barely produced anything. Healtheon never had a chance in the broader U.S. health care market.

By 2000, the company had merged with WebMD—a well-known company for sure, but not exactly a disruptor.

More the bane of every oncologist’s existence than actual disruption.


I’ve written before about how I think Apple, Google, and Amazon’s ventures into the health care space are something to be concerned about, and very smart health care thinkers have pushed back, saying that the Big Tech entities just aren’t prepared to make major, ground-shaking change in health care.

Haven’s collapse is forcing me to agree (unless TikTok completely smashes the game on AI drug discovery and makes us all think twice). Although the downfall of Haven seems more about Amazon’s diffuseness and willingness to crush competitors, it doesn’t look like Amazon’s going to be able to remake all of health care any time soon either. (It remains to be seen, however, if Amazon is able to roll up the retail pharmacy space, which would have a long-term negative effect on patients by decreasing access and raising costs.)

I have a lot more hope for smaller health tech ventures, and ventures that are focused on pulling just one lever of patient care. It appears that, unless the U.S. government decides to take on health care, full-scale disruption will remain a distant goal.