This health investor shadowed Jeff Bezos and worked for Bill Gates — here's what he learned

Jeff Bezos's ambitions in health care are no longer a secret.

Between the Haven partnership with J.P. Morgan Chase and Berkshire Hathaway, his company's stealthy work in cancer research, investments in health clinics for employees and last year's $753 million purchase of online drug seller PillPack, the Amazon CEO is openly taking on the $3.5 trillion health-care industry.

While most of those developments are very recent, Bezos has had his eyes on this market for over two decades. Nobody knows that better than Peter Neupert, who became CEO of online pharmacy Drugstore.com in 1998, after Bezos recruited him for the job in tandem with Amazon's investment in the company.

It was the beginning of a long relationship between the two and their effort to apply emerging technologies to life sciences. Neupert went on to Microsoft, his second stint at the company, where he led a new health solutions group. He left in 2012 and has since been a sought-after advisor for hospitals, biotech companies and start-ups.

"People are often approaching me for solutions," Neupert, 63, told CNBC in a recent interview. "What I tell them is I don't always have the right answers, but I do have a lot of scars and I'm happy to share those."

Bezos stepped off Drugstore's board in 2001, and Neupert left the company three years later. But they're both engaged, more than ever before, in advancing health care.

Amazon is hiring aggressively at PillPack to become a major player in prescription delivery and is exploring a host of efforts to bring down the cost of care. Neupert is on the board of medical lab company LabCorp, as well as Adaptive Biotechnologies, which recently went public, and a couple early-stage ventures. He's also on the board of trustees at the Fred Hutchinson Cancer Research Center.

Meanwhile, Amazon is still a big part of Neupert's life in other ways — he has two kids and a son-in-law who work at the company's headquarters in Seattle. That's where it all started for him 21 years ago.

Jumping to a dot-com

Neupert was working for Bill Gates at Microsoft, where he'd launched the MSNBC network as a joint venture with NBC. He was approached by John Doerr of venture capital firm Kleiner Perkins about a new start-up that aimed to sell prescription medicines, beauty products and over-the-counter drugs online. Doerr, an early Amazon investor and board member at the time, brought Bezos along for the recruiting meeting.

"They were looking for someone who could take a 10-page business plan and turn it into a real business at a time when everyone was starting companies," Neupert recalled.

Leaving was a tough call. Microsoft was on a tear and would soon surpass General Electric as the world's most valuable company, while Drugstore was an idea on a whiteboard backed by a cash-burning dot-com darling.

So Neupert asked if he could shadow Bezos for a few days, following him around for conversations with board members, staff meetings and conference calls. Bezos agreed, and Neupert took the job. From a short time observing Bezos, he learned something that was fundamentally different at Amazon than at Microsoft.

"I learned quickly from those days that opinions don't matter," Neupert said. "Data matters."

Bezos talked about A/B testing (comparing two versions of a design or project), rapid customer feedback and the importance of experimenting and failing.

In April 1999, two months after Drugstore announced the investment from Amazon and Kleiner Perkins, Bezos and Neupert appeared together on Charlie Rose. The host asked Neupert why a consumer would rather buy from Drugstore than their local pharmacy.

'Nobody likes to browse the Preparation H aisle'

"They don't like waiting in line for the pharmacist up behind the glass counter," Neupert said. "They don't like shopping in public for very private things."

Bezos chimed in to say, "browsing for books is fun but nobody likes to browse the Preparation H aisle." That was followed by a momentary pause and then Bezos's signature — some say maniacal — laugh.

Drugstore survived the dot-com bust but struggled to grow in an extremely fragmented business. Neupert left in 2004, five years before Walgreens bought the company for over $400 million and ultimately shut it down.

Along the way, Amazon learned some key lessons that would benefit the company as it pursued a deal with PillPack many years later. One big realization was that the established pharmacy benefit managers (PBMs), the industry middlemen, would go to great lengths to protect the status quo.

Amazon founder and CEO Jeff Bezos in 1997

Paul Souders | Hulton Archive | Getty Images

Drugstore was shut out of the pharmacy business by Medco Health Solutions, one of the largest PBMs at the time, which meant it had no real way of selling prescription drugs. PillPack more recently experienced similar resistance from Express Scripts.

Over time, Neupert saw glimmers of Bezos's ruthless style, which investors and analysts have said is a primary driver to his success. During the dot-com crash, for example, Amazon stopped giving away its email marketing access list to Drugstore and start charging for it.

"He did what was the best thing for Amazon," Neupert said. "I didn't like it at the time, but I ultimately respected it."

Second act at Microsoft

After leaving Drugstore in 2004, Neupert wasn't quite ready to give up on health care.

Over the following months, he spent some time in Washington, D.C., paying close attention to regulators and pharmacy lobbyists. He observed how the major players had their own entrenched interests, which created a big problem for medical software.

Record-keeping inside large hospitals and doctor's offices was still run on a combination of paper and homegrown systems. Electronic medical records systems were emerging but only a few large health systems, like Kaiser Permanente, were starting to shift over to vendors such as Epic Systems. So much critical data was siloed.

"It struck me that a few thousand people die every year because of adverse drug events, and a lot of that was down to a failure of systems engineering and software design," Neupert said.

That thinking brought Neupert back to Microsoft, where he saw an opportunity to build a business to help the health-care system with that problem. He kick-started a health unit, launching a product called HealthVault for consumers and health providers to aggregate medical information. One former Microsoft Health colleague, Sean Nolan, said the mission was ambitious and ahead of its time.

"Peter was always drawn to big ideas," Nolan told CNBC. "And he was never scared of challenging conventional wisdom."

While it's still working on that mission today under corporate vice president Peter Lee, the consumer part of the equation failed to get much traction. Microsoft announced in April that it will shut HealthVault later this year. Lee said there are 72 projects he inherited from Neupert.

"It's remarkable how much energy he injected into the health-care space that really never left," Lee said, in an interview. "It ranges from synthetic biology to radiology imaging. It was comprehensive. It's surprising to me how much stuff Peter got started."

Since leaving Microsoft in 2012, Neupert has advised numerous other health-tech companies. And he's watching from the outside as Amazon battles the industry incumbents.

"Bezos deeply understands pharmacy and all its complexities," Neupert said.

When he's not sitting on boards, cycling or spending time with family, Neupert is often sharing advice with entrepreneurs.

One thing he learned from Bezos is to write down your plan or approach to the market in a document and not in a bullet-pointed presentation, because complete sentences and paragraphs don't leave much room for assumptions and interpretations.

"In rapidly growing companies, this is essential for folks to stay on the same page," he said.

'Engage the regulators early'

He encourages entrepreneurs to be bolder and take more risks. In health care, that means not just developing software but also doing the hard work of providing insurance, improving outcomes and cutting down on costs.

"Over the years, I've concluded that it's really hard to make money by just selling software to the health-care system in the United States," Neupert said. "You have to be in the delivery side of the business, too."

A final lesson he shared is one that, as a technologist, seems counterintuitive and took him a long time to learn. And that's to reach out to regulators early on.

"Engage the regulators early if you are doing something they haven't seen before," he said. "It will save you grief later."

Follow @CNBCtech on Twitter for the latest tech industry news.

Read More