Earlier this month, my friend and industry colleague Rob Coppedge wrote a thoughtful post on the death of digital health, a space associated with too much hype and too little substance.

We agree that health care is really broken. It bankrupts our families, exhausts government discretionary spending, and leaves American industry at a global disadvantage.

However, here are a few divergent reasons for digital health’s (seeming) underperformance, which explains why early-stage investors aren’t getting the kind of returns they’re looking for:

• Incrementalism: Hundreds of startups produce enterprise dashboards, online symptom checkers, and evidence-light apps. Many of these are flippant and those that are worthy often offer only incremental progress. By nature of their modest impact and the barriers to scale, many expectations are not being met.
• Invisible unicorns: An aside, but there are many quiet successes in digital health that didn’t produce PR-driven Silicon Valley exits. How about CoverMyMeds? This team in Ohio used tech, network effects, and empathy to solve a real problem, reducing dispensing time for medications requiring payer prior authorization from an average of 14 days to 30 minutes. How? By embedding tech in the pharmacy, doctor’s office and payer — all while getting pharma to foot the bill. The result: a $1.3 billion exit to McKesson. Start-up land barely noticed.
• Late bloomers: There may be many sleeping giants, as solving certain problems with scale sometimes takes longer in health. Just look at Pear Therapeutics, which this week received FDA clearance after a years-long process for its app to help treat substance abuse. Investors will need to think very hard to discern what can scale quickly versus what will need to bake slowly for longer.
• Poor execution: The underlying issue with many failed or struggling digital health companies is poor execution. The excitement of the mission seems to slow down the need for timely product feedback. Most CEOs suffer from wishful thinking on the imminence of product-market fit, and defensible distribution has too often been an afterthought.
• The “O” word: Oligopoly. As providers seek local market monopoly in the name of “scale and efficiency” or to pay for expensive medical record systems, costs soar and innovation falters. Likewise, the pharma supply chain makes money because of a lack of competition and transparent data, and payers are swimming in record profits. A special team of founders should not turn back form these challenges, but they should know the treacherous waters into which they sail.
• Lack of ambition: Where is the Elon Musk of care provision? Why hasn’t anyone actually taken on the hospital? Hundreds of medication adherence apps entered the market in recent years, but only one start-up I can find is directly taking on the pharmacy benefits manager. In other verticals, we are seeing massive incumbents falling to data-driven, disruptive companies. We need some bigger bets in health care.

I try to imagine how investors and founders felt in 2004. The bubble had burst, forcing the posers to depart. Meanwhile, a much smaller group of entrepreneurs kept building things, believing full well that the promise of the Internet from 1997 was not a false mirage but would take two or three rapid ecosystem iterations to actualize. And I believe some VC vintages in the following years ripened quite nicely.

This is where we are with digital health. We still need better real-time informatics that surround all that touch the patient. We still need to move data, remove silos, reinvent primary care and elder care, and bring genomics to the clinic at scale. We need to do this while taking on new models for risk and improving outcomes for populations.

First, so many of the best ideas from the past 10 years were just too early. But with dependencies removed, lessons learned and chess boards realigned, it’s time for the bigger second acts. Napster and MySpace and AOL paved the way for some much bigger platforms.

Secondly, the quality of some entrepreneurs starting ambitious companies in health and data is breathtaking. The advances in health research, design thinking and technology stack capabilities are helping a new class of health-technology founding teams. The talent and the tools are improving.

And then there’s the data. Deep computation is transforming how we live and work, and how we make human progress in areas like space, agriculture, manufacturing, transportation and retail. This deep compute has yet to truly make its mark on how we provide health care, and one or two large organizations dressing up buzzwords as marketing doesn’t count.

Make no mistake, the advances in computational power and data science are rapidly arriving in health provision. In the next ten years, we are going to create massively-valuable companies using deep compute to improve and QA decision-making, better enable empathetic health providers, and impact systems of care to create more personalized experiences for patients at every stage of their journey.

In fact, at some point it will no longer be called “digital health” because all aspects of health care provision will be built upon and infused with data and intelligence, from the optimized patient to a “learning” system of care.

In addition to jumping through the aforementioned hurdles, we also need to address the assertion that there’s no clear path to make money in digital health.

There are many models, but here are three that are proven::

1. 50X better: Yes, you can sell into the belly of the beast of our health system, but you do so with a product that solves a very specific problem and is 10 or even 50 times better than what was offered before. You still face speed bumps in validation, regulation and adoption, but you persevere.
2. Full-stack: Rather than convince a huge number of stakeholders to adopt your technology, you identify the one that has to say “yes”, like a large self-insured employer. And you then go “full stack” as a tech and health business. Take Virta Health, a start-up that aims to reverse Type 2 diabetes. Rather than convincing thousands of providers and other groups to adopt new software and guidelines, Virta sells to the employer to become the “full stack” for the patient and builds a holistic provider-led clinic around this patient.
3. Network effects: Creating a virtuous cycle where your product gets more valuable to users as more people use it is incontrovertibly powerful. And yet, there have been very few network effect platforms in U.S. health care. Network effects don’t solve every problem, but in solving problems like moving data, delivering new networks of caregiver support, and deep areas of research, we need to see more of these.

In sum, we need to be more ambitious, invest more heavily in our best entrepreneurs, and rebuild every use-case with our best data and computation. The first chapter is behind us, and now is the time to be bullish.

Scott Barclay is an entrepreneur and operator who has worked in health care for 20 years and is now a Partner at Data Collective, an early-stage investment fund focused on deep tech and compute.