The 8 steps to Amazon entering the health care market

This is a thought experiment:  What if Amazon really wanted to go all in in the health care sector?  What might that look like?  What would their strategy be?  Where could they deliver value?

Amazon looks for industries that are not sensitive to the customer, that have profits or premium pricing based on barriers to entry (often capital related), and looks to exploit those opportunities. It’s pretty straight forward.  And, whether that industry is cloud storage space or groceries or “last mile” distribution networks, Amazon is thinking about it.

They didn’t go from 32,000 employees 5 years ago to almost 500,000 today by sticking with books.

This thought exercise is based on off the record conversations with a number of folks: brokers, health plan executives, tech company leaders, and a few folks at Amazon.  I didn’t want to put anyone on the record.  That’s not important for the purposes of this exercise.  All of this is produced below from open source content.

Here goes.


Amazon decides to enter the health care business.  How do they bring value?

First, they start with insurance.

Amazon already provides health insurance benefits to over 400,000 employees:  390,000 or so at Amazon, and 90,000 or so at Whole Foods.  They also cover some number of employee dependents.  I have no idea what that is, but given the relatively rich benefits provided to employees, let’s imagine it’s a modest 1:1 ratio – or one employee dependent for every employee.

(I recognize that’s not how most brokers or HR folks think about a self-funded account.  Generally speaking, the focus is on the employee and doesn’t include employee dependents when thinking about risk.  But, roll with me: this point is modest in the grand scheme.)

If we just count that total number of lives, they immediately become one of the largest health plans on the west coast.  With somewhere around 1 million covered employees and dependents, that number would put them about 6th in Washington State by size.

(I understand that their employees are beyond Washington State. Counting covered lives in ERISA plans is tricky and often multi-state, however.)

Even if they just count the lives in Washington State, an employee count I can’t identify with certainty, I’d still put them 6th or 7th in Washington State after Premera, Kaiser, Regence, United and Molina.  Aetna may be in there as well.

In any case, they’re big.  They have a big risk pool already in place.

Second, they bring pricing transparency to their members.

Amazon has over 1.5 billion SKUs.  They sell the same product, or similar products, across multiple countries.

They might have a provider network of 1000-2000 providers in a dense urban community.  Each of those providers might have 3-10 plan contracts. And, Amazon might have significant operations in 90 communities.  Each provider might have about 100 CPT or DRG codes that constitute about 90% of their billed charges.

I wasn’t a math major, but that tells me the maximum number of pricing items, assuming every plan has a different contract with every provider (we know that’s not the case), is equal to the number of providers times the number of contracts times the number of communities that might have a different provider network times the number of codes that might be billed.  Or, all together, at a maximum, this is the equivalent of 180m SKUs – or about 16% of what they currently manage.

That sounds pretty doable.

And, realistically, plan contracts aren’t generally that different from one another. Reimbursement may vary, but often it doesn’t.  While hospital or pharmaceutical contracts might vary, the vast majority of contracts a plan has – in terms of quantity of contracts, not total reimbursement – is with small provider groups, often solo practitioners. These providers don’t have much, if any, pricing power.

That means if they want to be “in-network” they have to take what the plan is giving. That means a significantly reduced level of variation among SKUs’, meaning a smaller number of SKUs to manage.

It’s just not a lot to manage within the context of what Amazon already manages.

Let’s put this in context.

Netflix is video/TV company. It’s entire budget is focused on video content.  That’s what the company does.

In 2017, Business Insider estimates Netflix will spend $6 bn.  Amazon is estimated to spend 75% of that at $4.5 bn just on video content.

COTD_4.10 amazon vs netflix spend

This means that Amazon’s scale can bring considerable resources to a problem, which compared to a company that is all-in on a given line of business, might still be hugely competitive.  All Netflix does is video, and Amazon practically matches them in spend.

All insurance companies do is cover lives.

Amazon already does that for 500,000 to 1m beneficiaries.  This isn’t a new line of business for them.  They just haven’t engaged the broader market yet.  Compared to any local Blue plan, Amazon will potentially bring a scale that no regional plan can match. And, I’m guessing they will find efficiencies in contracting – meaning SKU generation – that a health plan might not.

All this to say if 180m is the maximum number of SKUs in health care, it’ll probably be much, much – much – lower than that.

And, once they have those SKUs, why wouldn’t they share those prices.  It’s a central part of what Amazon does.

Let’s imagine that they start with their employees.  I can imagine the Amazon HR folks saying the following:

Here is your defined benefit.  We have pricing for every provider in our network.  You’re welcome to go anywhere you’d like.

We’ll cover this amount of the service you need.  That means at this list of providers, we cover it all. At this list of providers we cover a range of amounts based on what we’ve negotiated with them. Some providers aren’t price competitive, and you’ll see that in the pricing.  But you can choose where you’d like to go.

This isn’t terribly innovative. Castlight does this for its clients now.  It’s already happening at places like Providence Health and Services.

But, at Amazon, that’s just step one.

Third, they bring transparency to quality.

Amazon has been covering lives and has its own data on provider performance going back a number of years.  It can pull that data and share it with its members.  Any plan that might manage the Amazon account might want to keep that data private, and for good reason.  It’s generally considered proprietary in the existing market thinking.

But, would they say no to Amazon if the company demanded it?  If Amazon threatened to sue?  If Amazon got aggressive about what they might call “their data about their health plan covering their employees?”

That’s a tough spot, and one likely to result in Amazon getting what it wants.

But, let’s imagine it doesn’t get any of that data.  How do they report on quality amongst its current array of products?

Customer reviews. Just like covering lives, Amazon already does that, too.

Turn on the review machine and let customers tell each other about what it thinks about specific physicians, hospitals or providers.  Let them talk about the efficacy of a drug regime or what it felt like to be in an MRI machine.

Transparency vendors will tell you that folks don’t generally trust reviews from outside their tribe or community.  But they trust reviews 1) when they come from within their community (like that which would be covered by Amazon, likely a similar demographic), and 2) when there are enough reviews to show a preponderance of opinion.

That’s something Amazon can turn on overnight.

And, in a defined benefit model where Amazon beneficiaries are empowered with tools to select between providers based on prices and quality feedback, that will be empowering.  Now, hold this thought for a second.

Fourth, they lower prices.

That’s what they did with Whole Foods.  And, that didn’t seem like it was necessary.  Whole Foods shoppers like to spend a lot of money. The idea of buying organic, healthy, locally sourced food must come with high prices, right?

Maybe, but Amazon’s consumer centricity augured against that.  Prices came down because that’s good for consumers.  Amazon’s scale and purchasing power allowed this to happen.

Health insurance is different though, right?

Well, sure.  But, their risk pool is probably a relatively healthy one.  Their tech workforce is relatively young, well educated, and well paid.  Their warehouse and customer services workforce, I’m guessing through anecdotal conversations, is relatively young.  And, probably healthy as a result.

In other words, they probably have a relatively healthy risk pool into which new beneficiaries would enter, thus allowing Amazon to start with a price competitive premium for the outside market.

And they are relatively similar, thus reinforcing the power of their reviews of providers from their peers.

Fifth, they go to work on pharmacy prices and delivery.

Medicare can’t negotiate pharma prices, but I don’t think Amazon listens to the same advocates.

Put differently, pharmacy pricing is a perfect opportunity. Whether you think about the mom-and-pop corner pharmacy (which sounds like the brick-and-mortar bookstore they displaced), or the pharmacy benefit manager (which sounds like the middleman they’ve displaced among department stores), or about the pricing power that scale affords (which Medicare can’t bring), it’s relatively clear to see where Amazon can enter develop pricing strategies to lower costs.

But what about when you need a pharmaceutical that day? How would Amazon address that?

Currently, if your physician writes you a script, you have to go to the corner pharmacy.  You might drive 5-10 minutes, wait 10-20 minutes, drive home 5-10 minutes.  Whether that’s 20 minutes of travel or 40 minutes of travel, that isn’t terribly exciting when you or your kids are in pain or discomfort.

Why don’t you just go home after the doctor visit…  Because when your doctor enters the script, Amazon gets to work on the delivery immediately.

They already deliver Prime products within two hours.  They are already working on drone delivery in urban markets.  It’s entirely reasonable to see that they could deliver pharmacy products within the same time frame – and maybe even faster.

Sixth, they build their beneficiary membership.

Do you know that you can currently buy professional services from Amazon?  This week, I get my iPhone screen replaced. I purchased the service on Amazon. They come to my office, at a time that works for me. I get a call to talk through what I need and to confirm timing.  It costs me $120.

It costs me $100 to do it at the Apple store. There, I need to return to the Apple store, at a time that works for them.  It might be 3-4 days out, or if I have a reasonably busy schedule (don’t we all?), it might be a week out.

With Amazon, I get the professional service I want, at a time that works for me, with good customer service, and the premium I pay is $20.


If I’m a small business, and the second or third highest line item in my annual budget (after labor) is my health care benefit, I like what Amazon offers.

They offer lower prices, pricing transparency among providers, quality reviews of providers, and timely delivery of prescriptions. Why wouldn’t I move there?

I have to guess they offer significant analytics to my business or my HR team, as well.  They offer tools related to wellness (already do that),  and will have data about utilization (see their dashboard for publishers) that traditional health plans don’t offer (for a host of reasonable reasons that don’t apply to Amazon).

Moreover, they have a proven, deeply ingrained culture of customer centricity that is comparatively unique among all modern businesses.  It would be considered unique among health care organizations, too.

So, if I get improved pricing, analytics, tools, and customer service, that is a very compelling offer to the group marketplace.

And, this is probably a service I can access directly – meaning I don’t have to pay the 10%-15% markup for brokers in the small group space.  That is a savings Amazon will also pass through to consumers.

Seventh, they address the unit-cost of care delivery.

Amazon has already saved costs for consumers by offering a lower premium (via a healthier risk pool), with lower pharmacy prices (though pricing power and delivery services), and by eliminating brokerage fees (by going direct to consumer).

This means it’s time for Amazon to start pushing for better rates from providers.  Now, one might argue that plans already do this – and they do.

But, given what I’ve said above, if Amazon has pricing for every provider in a given region it serves – not every region in a state, mind you, they aren’t likely to engage markets outside of the 90 regions where their employees already live – why not make that information public?

If you’re a plan, you call that information proprietary.  There are a host of good reasons to keep that private, including provider relations and competitive considerations.

If you’re Amazon, you call that giving consumers the tools they need to make good decisions.

And, once those quality and cost benchmarks become public knowledge, providers will be forced to respond to consumer feedback loops that they simply don’t now.  Under this model, they will still be able to charge whatever they want, Amazon will simply publish their pricing.  That is if Amazon chooses to contract with the provider.

Amazon views all contracts like this as vendor relationships.  They beat up on everyone they contract with in order to get low prices for their customers.  That culture would likely be employed among its provider markets.

Eighth, Amazon gets in the provider space.

Most of Amazon’s beneficiaries are likely healthier than the other small group markets and non-group market risk pools that other non-tech plans have to manage.  These are employed, well educated, richly rewarded folks – even among their blue collar employees – at Amazon.  This is not a Medicaid or Medicare population.  This is not the individual exchange.

These are Amazon employees and group employers.  Of the group employers, I’d imagine a significant portion are in the small group, a market that is generally healthier than the large group market.  They are tech folks, other companies with an affinity for what Amazon might offer.

Typically, these folks need good primary care, and in many cases need a higher rate of behavioral health care services, like ABA therapy for their kids, than the population at large.  Moreover, these employees may not be able to access those services easily – between work, managing kids, running errands, etc.

Luckily – again – Amazon already services this population with other non-health care services.  This happens via video chat support for products that may not have been delivered in time, or telephone support for customer reviews, or email correspondence about complaints.

In those markets, or among those services where it can’t get strong pricing leverage (see the seventh step), it might think about entering the provider space itself.  If it will enter the bricks-and-mortar world of groceries, it’s not a leap to enter the higher margin world of health care services.

That’s where tele-health comes in very easily.

The leap from their current customer service support to a model of tele-medicine is not that far.  Providing tele-medicine and tele-health services, including behavioral health, ABA, etc, is a modest step from where Amazon currently is.  And, there are a number of firms that Amazon could offer to consumers via its platform.

For tertiary care or emergency services, Amazon could still rely on existing vendors or preferred providers.

But, if Amazon can buy Whole Foods, it can buy a network of urgent care centers, or primary care providers.  There are a number of them out there that would jump at that chance.

So, why not?

Here’s the important thing:  Amazon is already paying for these services for its members.  This is a new market for them, but it would be on similar platforms and Amazon already procures these services. Why not try to manage that spend more effectively, and share that with its broader non-Amazon membership in the small group market?

Right? What do you think?

None of these steps are revolutionary for Amazon.  Amazon already does most of this work, and that which it doesn’t do, it pays for – at least for its members.

Seems like a strategy someone smarter than me has already thought of.