Amazon offering health insurance: a potential mistake


35%+ of consumers who purchased on Amazon over a 30-day period say they would be “open” to health insurance created by Amazon according to a new survey (see Becker’s headline and the LendEdu study description). Not surprisingly, interest varied depending on degree of commitment the consumer had to Amazon: 62% of the sample were Amazon Prime customers, and, of them, 42% were open to the idea. That implies that 26% of non-Prime customers were open.

This data point is a fine example of the kind of freebie thinking Amazon can stimulate just by creating uncertainty about their intentions (although I doubt the specific concept is new to Amazon).

In this particular case, however, I think offering individual insurance (presumably the model consumers were considering) would be the wrong move for Amazon now (and likely ever) for a variety of reasons:

1. New entrants in health insurance have struggled. Oscar Health and Clover Health attracted a lot of venture money and seemed full of disruptive promise but have fallen short of expectations. Even ventures backed by major insurance incumbents (Harken Health launched with UnitedHealth funding) have failed. The trouble is that scale is essential to win at being a health plan but market share growth is best done in increments. If you grow too fast, you find out you’ve attracted a lot of bad risk and may spend years digging your way out.

2. The most financially attractive parts of health insurance are highly regulated by either the states (commercial group) or CMS (Medicare Advantage). Regulation – in part intentionally and often wisely – slows down iterative innovation processes and make sources of advantage (e.g., clever benefit design) transparent and therefore replicable. Slow innovation does not play to Amazon’s strengths.

3. Health plans must often say no. While clinical pathways are making progress driving towards more consistent EBM, prior authorization and utilization management remain powerful tools for managing the cost component of value. Same goes for tightening network designs. Does Amazon want to be in the business of saying “no” to its customers?

4. We are a long way from establishing what is the right model for health care provision and coverage. There are plenty of technology and operating model ideas with great promise and a lot of experimentation is underway but the right combinations continue to elude us. Low value care remains commonplace along with patient frustration and widely varying outcomes. I’d argue that neither Amazon nor anyone else has the right answer and, even an organization as capable and resourced as Amazon, will not figure it out on its own. We need to stimulate a more liquid market to allow that right answers to emerge. As I’ve argued elsewhere, creating the platform for that more liquid market would be a play more worthy of Amazon’s patterns and capabilities.

Therefore: despite the data point, the more likely scenario is for Amazon to enter healthcare as a shopping facilitator rather than the insurance or service provision. This view does not exclude Amazon experimenting on its own employees and those of Berkshire Hathaway and JPMorgan but that isn’t the same as selling Amazon insurance to Amazon customers. It also doesn’t preclude Amazon getting into some service provision eventually once there is more clarity on where the market might shift. Take the deal to sell OTC (white-labeled from Perrigo) as an example; buying into a telemedicine provider for minor acute care would be the corollary if there is a working Amazon shopping/scheduling platform.