Compete or Connect: Scalable Business Models in Bio

The bio space is absolutely humming right now with new ideas, new companies, new regulatory opportunities, and new go-to-market strategies. The COVID pandemic has served as a siren call to innovators across biotech and digital health, beckoning talented people from many fields to work at breakneck speed, and asking how they might contribute to the mission of improving human health.

This accelerated pace of progress is a wonderful development, but it does beg the question: Out of thousands of startups, which bio companies are likely to have the greatest impact and scale? And for all bio companies, which business models will endure the test of time? My answer, in a nutshell, is to “Compete or Connect” to win in bio. And perhaps because I’ve never been able to fully shake the management consultant in me, I’ll share with you a greatly oversimplified but nonetheless useful framework for how I think about business models in bio. I hope it may help guide startup founders as well as strategy and business development leaders at established bio companies.

Navigating the world of bio can seem complicated: rapidly evolving science, many different stakeholders, intricate regulation, various enterprises both directly and indirectly engaged with patients (B2C; B2B2C; B2C2B; others), and many instances of end users who aren’t actually customers. So let’s start by simplifying the space. There are really just four major categories of constituents in bio:

  • life sciences — those making drugs and diagnostics
  • providers — those delivering healthcare services
  • payors — those paying for both life sciences products and healthcare services
  • patients 

Yes, there are subsets within each category, and flavors of each with different bells and whistles, but these are the major entities through whom dollars flow in bio. As a result, most new bio companies will either resemble existing players in these categories, or require members of one or more of these categories to be their customers; for example, “a pharma-facing SaaS company”, “a next-generation care delivery start-up”, “a biotech newco developing therapies”, “a marketplace for payors to find digital health solutions”, and so on.

I am particularly intrigued by companies that pick business models which enable them to either “Compete or Connect.” What does this mean? By “Compete” I refer to companies that envision becoming full-stack competitors to existing players in any of the three bio enterprise categories: life sciences, provider, or payor. And by “Connect” I refer to companies that simultaneously engage two or more bio verticals as customers of their products or data, thereby creating mutually reinforcing connectivity between multiple verticals. There are myriad successful business models in bio (all of which do not fit this framework!), but “Compete or Connect” models are especially exciting for reasons I’ll describe below.

Will you “Compete” or will you “Connect”?


Let’s discuss “Compete” first. A full-stack “Compete” company in bio decides that, in order to recognize the full value of the core technology that they are building, they need to own a significant portion of the value chain themselves. In other words, they use their technology to bet on themselves and compete with (rather than serve) incumbents in their vertical. 

In life sciences, this means the company decides not to become a software vendor or a parts-supplier to Amgen, but rather to become a next-generation Amgen themselves; they commit to investing in an internally owned drug or diagnostics pipeline, supported by their own innovation. In the world of providers, this means the company decides against providing infrastructure to existing care delivery organizations, and instead becomes a better, more efficient provider themselves (albeit starting with narrow scope) — by arming their own care team with streamlined software. And among payors, going full-stack means the startup company decides not to sell services into the world’s largest insurance companies (as many great companies do!), but rather uses its technology to better manage risk across a population on its own. These companies compete head-on with the universe of incumbent payors by providing a better member experience and leveraging smarter risk analytics. 

Decisions to “Compete” are gutsy choices. It’s hard to compete with incumbents who have war chests of capital, and it can be hard to turn down the flow of services revenue which might flow from those same war chests. But the advantages of going full-stack are real: a more direct line to one’s customers and users; more control over the end-to-end product development cycle; a chance to grab and grow a much bigger size of the total pie. Given the outsized value that bio enterprises can recognize at the finish line — when a drug is approved, when a provider organization succeeds in value-based care, or when total cost of care is reduced across a payor’s members — it behooves bio startups to ask whether there is a way for them to seize more of that value and “Compete.”

An illustrative — but by no means complete — selection of companies choosing to “Compete”. Portfolio companies described or referred to in this presentation are not representative of all investments in vehicles managed by a16z and there can be no assurance that the investments described are or will be profitable or that other investments made in the future will have similar character or results. A list of investments in a16z funds is available at


Now let’s turn to “Connect”, an orthogonal go-to-market strategy that can also be powerful in bio. Companies that “Connect” position themselves as the necessary intermediaries between the major bio constituents (life sciences, provider, payor, and patient). “Connectors” sit at the intersection between (at least) two different bio verticals, and create mutually reinforcing linkages that didn’t exist before — or at least didn’t exist efficiently before. Often, in bio, the basis for these connections is a new ability to share valuable data that would previously have never crossed the walls between two otherwise highly regulated and siloed verticals. Data can function as a layer of connecting pipes, routing around or punching through those walls in ways the entities themselves could not do alone. 

Becoming a “Connector” is challenging, too. Not only are some of the competitors already sitting at these interfaces mature players (akin to incumbents in the “Compete” model), but being a “Connector” requires pursuing a two-sided business model and keeping both sides of your equation sticky. In some sense, it might feel like building two businesses simultaneously rather than just one — this could mean everything from two product stacks to two separate customer success organizations (e.g., to service both providers and life sciences companies, or to engage both payors and patients as distinct groups of users).

Sometimes, “Connector” companies evolve into this role after first developing a product focused on one vertical, and then later realizing there exists a market at the interface between their existing base of users and those in another vertical. In other cases, “Connector” companies start with a two-sided business model from day one: Marketplace businesses are a great example of companies that navigate the “cold-start” problem early in their journey by building an initial cohort of evangelical users on both sides of their marketplace.

An illustrative — but by no means complete — selection of companies choosing to “Connect”. Portfolio companies described or referred to in this presentation are not representative of all investments in vehicles managed by a16z and there can be no assurance that the investments described are or will be profitable or that other investments made in the future will have similar character or results. A list of investments in a16z funds is available at

Since “Connect” can mean so many different things, let me share two specific examples of very different “Connector” businesses; one more established, and one newer: Flatiron and Headway.   

Flatiron Health, where I was previously part of the product management team, had the mission to organize the world’s cancer data, and use it to improve patient care. They initially developed customer relationships with providers (community and academic oncology clinics), but ultimately realized they would always be leaving value on the table if they functioned only as software vendors to an already margin-constrained group of providers. The team then observed that the same datasets being generated every day by oncologists as part of their routine care for cancer patients could also be incredibly useful to a second bio vertical: the life sciences industry. So Flatiron, using venture funding, acquired one of the most widely used, cloud-based oncology electronic medical record (EMR) companies (Altos Solutions, the developer of OncoEMR). They then placed this provider-facing business alongside a second, distinct pharma-facing business unit within the company that became focused on cleaning, annotating, and deriving insights from EMR data for the life sciences industry. 

Two product stacks, two sales teams, and two customer success organizations. A hard feat to pull off, but today, most pharmaceutical companies with an oncology pipeline access Flatiron data products to help them make real-world data-driven drug development and commercial decisions. And as a testament to the strength of this “Connector” business model, Roche acquired Flatiron in 2018.

Headway (full disclosure: an a16z portfolio company) is a startup building a marketplace in the behavioral health space. Today, the most commonly cited obstacle in access to care is that only ~30% of mental health professionals accept insurance (payor) reimbursement, making it unaffordable for most patients. While many successful behavioral health companies have chosen to “Compete” in the provider category by offering lower-cost, direct-to-patient care, Headway looked at this payment/access problem, and instead saw an opportunity to become a scaled-up “Connector” between multiple bio verticals: between providers and payors, and also between patients and providers. 

The company decided to build a software platform to help providers, such as therapists, seamlessly submit insurance claims to payors. It seems simple, but the administrative overhead required to do this — and the risk of claim rejection — has been the primary reason so many mental health providers have stayed off of insurance reimbursement rails. Headway’s ability to submit insurance claims on providers’ behalf naturally also supported the development of a marketplace. Here, therapists can list their availability, and patients can schedule appointments online with individual therapists who accept their insurance and fit their needs (location, video vs. in-person visits, clinical specialization). Over time, Headway’s data connectivity could create network effects: with more clinical and financial data from more participants, one could imagine increasingly precise provider referral pathways, as well as value-based and quality-based reimbursement rates. These advances would ultimately benefit all of the bio verticals with which Headway interfaces (providers, payors, and patients), in turn driving further marketplace expansion.

Higher risk, higher reward

A common theme among companies who “Compete” and those who “Connect” is a willingness to bet on oneself, and an early desire to start taking on risk in the bio ecosystem. Sometimes this proclivity toward risk indicates that a company with a different business model is actually ready to “Compete” full-stack, or “Connect” in a rewarding way. 

An instructive example here is the evolution of Schrodinger’s business model. Schrodinger, today a leader in chemical simulation software, started out over 30 years ago selling physics-based computational tools to help drug discovery teams at pharmaceutical companies design and optimize small molecules. This model worked for many years, attracting all of the top 20 pharma companies in the world as customers, as well as a long tail of smaller customers with high retention and a gradually expanding annual contract value (ACV). 

But eventually, Schrodinger realized that if their software was so impactful in helping customers decrease the risk that a drug program would fail, then why not take on some of this de-risked drug development directly? In 2009, Schrodinger co-founded a joint venture called Nimbus Discovery, and licensed its software to this entity for exclusive use against a set of initial targets. This was a step toward going full-stack, and accessing more of the software-enabled upside of a drug’s ultimate success. By the time of their public filing in 2020, Schrodinger disclosed more than 25 collaborative drug discovery programs with biopharma companies (each associated with upside development milestones, option fees, and royalties), as well as five internal, wholly-owned drug programs. And thus, a company that started as a software vendor morphed into a company that is now ardently “Competing”: gradually taking on more and more risk, and ultimately building a full-stack biotech team to support their ambitions. 

At the provider-payor “Connector” interface, meanwhile, a similar evolution in risk-sharing is playing out. As payors roll out waves of different value-based reimbursement models, many SaaS companies emerged to help providers participate effectively, and succeed, in these programs. An initial business model was to engage only the provider as a customer, and sell software (licensed at a fixed monthly price) to power the additional data collection and quality measurement required by payors. 

Over the past decade, a spectrum of new business models have emerged in which a software company accesses a portion of the provider’s upside savings, and possibly also protects them from some of the downside risk (examples include Aledade and Agilon Health). Some companies have gone more full-stack, even acquiring chains of medical practices that become power users of their technology and negotiate directly as a large group (e.g., an ACO) with payors. And still other companies are becoming true “Connector” businesses, which are neither provider nor payor themselves, but that facilitate and adjudicate new value-based contracts between entities that would otherwise have been unlikely to transact. 

Note that there will always be a spectrum of risk-sharing, and not all successful examples will fit the mold of the “Compete or Connect” framework! See my team’s recent piece on horizontal biotech platforms as a case in point.

*   *   *

Software is eating the world. And yes, software is certainly eating bio, too — but I see this ringing especially true when companies pursue business models that position them to “Compete or Connect.”

The bio world is all about taming risks: the risk that a drug fails in clinical trials, the risk that a patient’s condition deteriorates without adequate intervention, the risk that healthcare costs increase untethered to value, the risk that a patient has a negative experience. If your software helps you mitigate any of these risks, you are to be commended — you have already achieved an incredible feat! But to get full credit for this feat — with more potential impact and durability — consider finding a way to “Compete or Connect.”

Biology is eating the world

Sign up for our bio newsletter to get the a16z take on the future of biology and engineering.