Are we in the midst of a Digital Health evolution or revolution? Venture Capitalists Bob Kocher and
For three years running, Venrock partners Bob Kocher and Bryan Roberts have put their digital health market predictions in writing. They recently sat down with Clifton Leaf, editor-in-chief of Fortune Magazine, to go over their 2018 market insights. Bottom line: it’s all about scale, profound engagement and provable ROI. Also, never lie during diligence.
Clifton Leaf: Let’s jump into one of your predictions. You said that this is the year where you’re going to get some healthcare IT exits. Some venture exits. Now, is this wishful thinking?
Bryan Roberts: Yes.
Leaf: That really describes a level of maturity in the space, that you think that these funders are ready to get some money out and start investing it elsewhere.
Roberts: I think it connotes two things. One is a level of maturity in the space for developed products with real ROI and customers. It also speaks to the desire for many of the other large participants in the ecosystem to reinvent themselves. Every payor that’s not United wants to create its own Optum. United wants to become partially a provider. Every provider wants to become partially a payor with IT systems. Nobody in healthcare, leave aside the pharma folks, makes enough money to be actually happy with their business, right? They’re all around break even, on a good day. On huge revenue. So,it’s not like they need scale, but they need different lines of business and they’re all searching. And that pull from the acquirer side, in my mind, is what is driving a bunch of that prediction.
Leaf: Well, it’s not just about size, it’s about ecosystems. It’s about trying to get more than the horizontal integration. You need to move across and make sure that all the pieces fit together.
Bob Kocher: There are things that are big enough now, that are worth buying if you’re a larger company. Because they’ve matured to have enough revenue that when you put it into a large organization you can distribute it, you can scale it, and you can make it actually be lucrative. Also, with stocks at highs, and markets that are really favorable, currency’s cheap for acquirers to go out and buy things.
Leaf: But the acquisition targets are getting pricey too, though, in some cases.
Roberts: God, I hope so.
Leaf: When you’re thinking about the next acquisition targets, what do you look for beyond the fact that there seems to be a commercial market for the product?
Kocher: First, really, really, unequivocal ROI, for whomever’s buying the product. There’s a lot of things in healthcare that will make us healthier and perhaps happy and have better experiences, but I look for things that really have tangible, unequivocal ROI in that first year, for whomever’s paying for the service. The second thing I look for, are things that a lot of people need to use or are using. So, high levels of engagement. Because healthcare is full of distractions — it’s hard to engage. You need things that have profound engagement so they’re likely to scale and grow to be big enough.
Roberts: We invest in early stage formative businesses, right? So actually if the world likes the business when we first get involved, that’s a bit of a problem. Because it’s not a mature enough product, or team, or product market fit, to actually service anybody. So, my hope is that we find things at the start that have the characteristics Bob described to you, but other people don’t see yet. It’s a product, it’s a need, that that whole, either customer, or M&A base, doesn’t fully understand and want yet. Because that gives us two, three, four years to actually do something and become mature enough, that the market doesn’t run away from us.
Leaf: That brings up a good question. There are a lot of startup founders who say, “Hey, I know this is going to be valuable. I know this is a service that people want and need, and that there’s a market for it. And I’m getting in now while there’s no competition. How do you find that sweet spot where it’s still not discovered, but it’s there?
Kocher: It starts with finding some early customers that you think you’ll learn a lot from, who are willing to give you the feedback, the help, the access, who will take a chance and let it get iterated through them. A lot of what we look for is an idea that hopefully isn’t too obvious, and hasn’t already been discovered. And then there’s somebody that we believe will be a good early partner to them. And then after a few years of iterating the thing, it ends up being good. Like in the case of Castlight Health. Safeway, way before there was a product, said, “We want to try this.” And that experience allowed them to build a product that then could be sold and scaled.
Leaf: Bryan, talk a little bit about your time with Illumina. I think it’s one of the best stories.
Roberts: I had an early investment a long time ago in Illumina (it is now a very mature company).
Back in 1999, 1998, Affymax had spun out Affymetrix. Affymetrix was the big gene expression, gene array business, and it was using semiconductor processes to make chips to investigate DNA. Illumina and they were sort of the big dogs in the space with a lot of IP around the feature size to make chips. They’d kicked a bunch of people out. Illumina’s initial focus there was to say, we think we can take a very different approach to gene expression reading. We’re going to make random arrays, which gets around all the IP of trying to have specific feature sizes, and then use bioinformatics to decode that random array for every specific experiment. Illumina’s co-founder, John Stuelpnagel, realized that were Illumina to become successful, they would need more than all of the world’s supply of oligonucleotides. They found a one man show, a guy out of his garage, who had taken an orthogonal approach to making oligos. They brought it in-house, made it industrial, and not only used it for their products, but licensed it out to the people who were making oligos and selling them commercially and brought down the price of oligos by something like 90%. It actually unlocked for them, and even more so once they acquired Celexa, the ability to do next gen sequencing for the whole oligo market.
Leaf: So this is someone who had the vision to see what they would need, if they were successful, and have the courage and convictions to go and do that and build a market.
One of your predictions for 2018 is that there will be another Theranos. So there was a Chicago-based company that seems to have fallen on hard times. Do you have any other thoughts on where you think this could happen?
Kocher: One of the things that’s surprising to me was when I worked at the White House, I would sit in my office everyday and have a bunch of meetings, always with interesting people who lied to me about things that they wanted the government to do for them to make their businesses better. From their perspective, make America better. And you get good at, sort of, smiling, saying thank you, and then going to the next meeting. I’ve been surprised as a venture capitalist that I get lied to, I think, more now than I did at the White House, by people making claims that aren’t true. In diligence, when you go check things that you hear from entrepreneurs, such as the customers they have or the revenue that they have, or investors that they have, or people that they fired, an amazing amount of the time it’s not true. Like, literally, things that are easily checkable. There’s a bunch of businesses over the past few years where we’ve done diligence and a lot of the things that were said weren’t true. Yet, they’ve gone on to raise sometimes prestigious amounts of money.
Leaf: Yeah, the surplus of capital is the problem.
Roberts: We’re in year 6 or 7 of there being way too much growth capital, right? Rates of return everywhere else in the universe are single digits and this growth tech innovation sector continues to provide interest and rates of return, so capital has flooded to the space. Healthcare IT with, you know, the wiring up of docs, and a variety of the regulatory reforms, has attracted a whole slew of new investors to the space. So, you have a surplus of capital and new people entering the space. It’s a situation ripe for people investing in things that look interesting, but below the tablecloth it doesn’t pan out.
Kocher: One thing that is interesting is that when companies have raised a reasonable amount of money, many investors assume that the investors before them did a lot of work to make sure that it all made sense, and nobody’s doing the work in some of these cases. Now there’s a series of companies where you go beneath the tablecloth, there’s not a lot there. And there will be a bunch of people who will be spectacular surprised.
Leaf: That happens in a lot of these excited periods of investment. We’re probably going to see that in some cryptocurrency investments. One of the areas where there is a lot of excitement, some would say hype, is in AI and machine learning. And this is an area where you were biting your tongue before now, this year, you’ve said we’re going to start to see some significant investments in this in a way that we’ll see some payoffs, too.
Roberts: Let’s use machine learning into AI as sort of a continuous spectrum of things. You have sets of systems that continuously improve and increase the performance in barriers to entry over time. We, for some set of years, had seen a bunch of the healthcare equivalent of identifying cats on the internet, right? Image recognition, x-rays and so on. None of those felt super compelling to us from a business model prospective at scale.
More recently, we’ve begun seeing a bunch of really good work that, for us, is more broadly applicable across healthcare. So, one of the issues in investing in these sorts of businesses is this. Every pitch that people ever come to us with starts out with, healthcare is a 3.4 trillion dollar industry and there’s this many hundreds of billions of dollars of waste. And along some few sets of dimensions, it is an enormous industry. And across lots of dimensions it is like 10,000 little industries. Right? Especially from a disease focus. If you’re going to go do AI for x-rays, you need a whole lot of x-rays of ankles and that doesn’t help you at all with knee prognosis or shoulder. You’ve got to do it over and over and over again. And so, it’s that set of applications that are useful across a lot of healthcare that have gotten our interest.
Leaf: You’ve written in the past, both of you, that people are underestimating the sheer complexity of human biology. And that it’s not going to be solved by even a few clever algorithms and all the data in the world.
Roberts: I think the application that we’ve spent a bunch of time on lately sits at the intersection of machine learning, AI, and voice. It’s a different input set. It’s an input set that can be passive. You’re not asking docs to do something else. And improving both the efficiency and quality of data capture for docs could be awesome.
Kocher: We don’t understand human biology as well as we think. So I think that the AI applications might be actually less clinical and more about everything else we do in healthcare, because more than half the labor is doing coding, collections, revenue cycle, credentialing, network management, provider directory. There’s all that other stuff. But we suck at it too. And so, AI can help us in all of those dimensions. And then we don’t have to worry about the FDA, reimbursement, harming somebody. You can work on all that other back-office stuff, and I think in those settings, I mean, credentialing, every doctor for every hospital, like, sends a copy of their diploma in. There’s a lot of ways you could apply technology to that.
Audience Question: How do you think the markets going to adjust with the first tranche of millennial parents who are going to be forced to deal with a healthcare environment that’s not ready for them?
Roberts: Every part of the ecosystem is trying to form more of a relationship with the patient.
Right? The payors are trying to reach through the employers into the patients. Maybe, for the first time, sometime soon, a doctor or physician group will think about churn, and lifetime value of customers, and stuff like that. So honestly my hope is that the set of folks who currently continue to think about themselves as invincible and having no healthcare problems at all, sort of cresting into the healthcare system, will change some of the dynamics of interaction. Telemedicine is something that everybody in this room should do as their entry into the healthcare system, right? It’s more convenient, it could be cheaper, it’ll send you to the right place or time and still, I guarantee you, not all of you use it. So, hopefully, it will push that forward.
Leaf: Will we see rule changes? Do you think a relaxing of interstate rules on telemedicine?
Kocher: Many states this year have joined the interstate compact to license doctors across state lines. It’s easy as the doctor gets multiple licenses, so you can serve patients in multiple states. I think you’re going to see more relaxation of regulatory boundaries.
Leaf: What about the idea that when we start to put software and sophisticated algorithms and other computer technology in devices that they become medical devices in and of themselves. Is your Fitbit a medical device if it can diagnose atrial fibrillation? Is your Apple watch a medical device if it warns you of a heart attack?
Kocher: It’s really hard to draw clear lines. It’s not going to be a hundred percent clear. The FDA is trying to make it easier to understand the lines, and things like Apple watches, I think, are going to define what that line is, because they’re getting much more sensitive. For a long time the false positive on the sensors were such that you didn’t want to use them as medical devices, because it would lead to lots of unnecessary treatment. But they’re getting better. And so, I think we’re going to see the FDA get a lot more permissive on what you can do with them and take an approach that allows you to update the software much more often. The real issues been less that the FDA has been hard to predict, but more that once they said what it was, if it was a medical device, you couldn’t upgrade it at the same rate that you upgraded everything else.
Leaf: What do you think about the role Amazon is going to play in healthcare over the next phase.
Kocher: I do think that they can do what they do really well, which is deliver things to healthcare organizations. I think they’re going to take on the distributors first. Because there’s no reason why Prime can’t deliver bandages, and sutures, and everything else to the hospital before they deliver the meds. The worst thing to deliver is the medicines, because that has all the regulation, and refrigeration. Like, there’s a trillion dollars of other stuff to sell first. And so, I think they’ll [medications] later.
Audience Question: There’s a proliferation of services. Someone’s got to pay for them. Under what circumstances does a direct-to-consumer model work?
Roberts: Today, in the seven deadly sins.
Kocher: That’s the actual answer.
Roberts: Vanity, envy, sloth, greed, you know, that sort of stuff. Seems like there may be a business today in trying to help collect out of pocket payments that have been going up, because of high deductible health plans and, sort of, bad debt has soared. So, that would be the countervailing notion, to people being willing to pay for anything. We’ve been, or I, have been unsuccessful at figuring out a direct-to-consumer business model that pulls additional dollars out of patience pockets.
Audience Question: How can digital health entrepreneurs overcome death by a thousand pilots, where they go from pilot to pilot, but never get deployed? And do you guys have any comments on the CVS Aetna deal?
Kocher: CVS Aetna is intriguing to me, in that large health plans struggle to figure out how to differentiate themselves vis a vis other health plans to large employers. And, many of you have probably had some combination in your life of United, Aetna, Cigna, Anthem, and it’s difficult to remember the differences. They have slightly different fax numbers for when you submit your reimbursement. And none of them have websites that work very well, but beyond that, it’s difficult to figure out which one’s which. And that’s terrible if you are a health plan trying to sell to a large employer or any market in particular. Aetna is a plan that had low market share, in all the markets that they’re in, so they’re always going to a hospital saying, “Hi, I’m Aetna. I want a contract.” And you have less market share than the Blue or United, usually. And so you’re taking a higher price. So, your network’s not cheaper. Your admin costs are, sort of, the same, and so it’s really hard to compete. And they make no money. CVS has something interesting which was a big PBM that makes a lot of money in a lot of stores, where you can pick up medicines and MinuteClinics, that have a lot of excess capacity. And so, I think the idea is, you put them together and you, hopefully, have a health plan that allows people to go to MinuteClinic to start care and pick up meds in the drug store instead of over mail order. That overcomes, with some margin over there, the network disadvantage that they have that might look a little different to employers. I think that’s the premise. The hard part I think is when you combine two enormous things with hundreds of thousands of people between them. With net promoter scores each that are like, five. How you make that net promoter score not negative five. The idea that you can, you know, when I go into a pharmacy to pick up my meds, I’m never feeling like they’re treating me any nicer than the hospital, that’s not going to be very nice. And, the idea that you’re going to make that into some sort of awesome member experience that’s seamlessly integrated and technologically connected to your healthplan, that feels hard to me. But I think that’s the bet that they’re trying to make and that will be interesting to see how it plays out. I think it will get approved. Competitively, they are not overlapping.
Roberts: I think if you are at risk of suffering death by 1000 pilots you deserve it. Every large organization has an innovation team and an innovation budget and you can get 50,000 bucks out of just about anybody. And, if you start taking that 50,000 bucks, it doesn’t do you any good at all. There’s no tolerance for products that don’t have very short term, very hard ROIs. And if you have a short term ROI, you’re not spending your time in the innovation group. Like I just stay out of the innovation group. You want to get into the actual operating business. And, if you have something that really makes their lives easier, and, let’s be clear, the first 3 things on their top 10 list are makes me more money, and makes me more money, and makes me more money.
Leaf: One of your predictions is that, believe it or not, they are going to repair what they can on the ACA, and sort of leave things alone eventually.
Kocher: The President thinks that he has repealed the ACA. And so, by taking away the individual mandates and leaving the subsidies, you actually leaving intact all the ACA, for the most part, because all the insurance market regulations are there, the subsidies are there, the exchanges are there, the ACA is there. But without a mandate, if you do nothing, premiums are going to go up a bunch. And that will be upsetting to whoever’s in charge. And so, I believe that, while they will, out of one corner of their mouths, say that they have gotten rid of the ACA, they will probably do a bunch of things to stabilize the markets, whether it’s reinsurance, or risk quarter payments, or funding the CSR’s that they kept withholding. They’ll want to do things to buffer that increase, because actually, the government’s the one who pays the price. As premiums go up since most everybody’s getting subsidies, the subsidies are actually an extra income. And so, if premiums go way up, the government just pays more back in subsidies.
Leaf: How would you pitch [a room of entrepreneurs] on what you’re looking for and the next great investment at Venrock?
Roberts: We tend to be very time frame agnostic investors. There are a couple of boards that I’ve sat on for the last fifteen years. Most of the time when you talk to VCs people are like, “How quickly do you want an exit, and what’s your pressure to get out?” We are in the fortunate position of not having a lot of that pressure. Which puts pressure on another part of the equation for us, which is on the scale of the business. We tend to get involved in businesses that we think have some chance of success, but are solving a problem that we see as large and very needy. We’re trying to build, over probably a decade, a product of real interest. Trying to build companies that will be very large. Because that’s the only way you can justify spending a decade or 15 years involved in something. Right? So, one is scale and neediness of problem. And the other is long-term company-focused people. Oftentimes, we get involved in things where we have those two things, but aren’t quite sure how we’re going to go about accessing the market. So, there’s a bunch of to-ing and fro-ing and pivoting that goes on, on the exact thing, but generally, we’re right on the product need. We don’t tend to invest in incremental things. We don’t tend to invest in product features, and stuff where there are huge dependencies on other people. Because it’s hard for us.
Leaf: Last question. Think about a timeline for where we are here. Is this an evolution or a revolution?
Kocher: We are really early. And so I’m excited because I think for the rest of our lives we’re going to continue to see rapid adoption, improvement, disruption and change. And it’s going to keep going for 15 more years, because there’s 10,000 boomers a day aging, for 15 more years. And so, we’re just getting to the beginning of the healthcare economic crisis that we’ll have, of the demand that we’re going to see for healthcare. And things like AI, and machine learning, and all forms of computative technology are going to allow us to, hopefully, solve problems faster than we get destroyed by the cost.