Not Too Big to Fail

20 Jan 2021

The demise of Haven Health illustrates that for a data-driven enterprise, success depends on more than having the deepest pockets or the strongest industry connections.

An important recent news story that resonated very little was the closure of Haven Health by its owners, Amazon, Berkshire Hathaway and JP Morgan Chase. When Haven Health was started amid much fanfare just under three years ago, it was heralded as a focused effort by three of the most powerful companies in the United States (and their three celebrity CEOs) to “to figure out how to reduce health-care costs for their hundreds of thousands of U.S. employees.” The prominent author-physician, Atul Gawande, was appointed as Haven’s leader. The alliance of the three companies was viewed as so imposing that the share prices of healthcare companies such as Aetna and CVS fell by 2-4% the day after the announcement.

However, according to the WSJ’s recent post-mortem, even the effort’s first step, collecting data about each company’s current healthcare costs proved too difficult:

Initially, the leaders of the joint venture imagined that if they could see what the three companies were spending on health care and why, the data would show them what to fix, those people said.

Getting a hold of those figures proved difficult. Haven employees built a platform to allow them to compile cost and claims data from all three companies, but the companies were unhappy with how it worked, people familiar with the matter said. Due to those concerns, Haven had to rebuild aspects of the system, further delaying the goal of understanding, analyzing and reducing costs, the people said.

Haven also had difficulty gaining access to information that health insurers try to keep secret: specific details about pricing as specified in agreements with hospitals and other medical providers.

Another obstacle to the healthcare venture’s success was the varying structures of the three corporations’ workforces. For example, Amazon has an HQ contingent of technologists, business/finance teams and other white collar specialists; and an even more numerous mass of warehouse workers. Berkshire Hathaway is a federation of individual companies, each with its own healthcare approach.

Not surprisingly, Haven had significant turnover in key positions:the handwriting was on the wall in May 2020 when Gawande resigned as CEO.

Here are the key lessons that business and technology leaders can draw from the Haven Health experience:

• generous backing from deep-pocketed backers is no guarantee of success;
• when trying to solve a big problem, sometimes it’s more effective to start small and to iterate like start-ups do (read up on “lean startup” methodology pioneered by Eric Ries);
• if a business experiment isn’t working, then better to cut losses sooner rather than later.

On this last item, at least, the three companies got it right: on the scale of three massive corporations with total annual net profits over $50 billion, two years and $100 million spent is literally a rounding error.