A venture capital approach to science grantmaking

One of the major challenges facing the US science ecosystem today is the shortage of funding for novel and risky scientific ideas.

In talking with a number of scientists, the core of the problem seems to lie with the design of the current funding model.

The traditional peer review-like approach to grantmaking, limited funding resources, and political pressure, leave a suboptimal environment for surfacing, cultivating, and funding the types of risk needed for breakthrough science.

Making progress against this reality requires fresh thinking that looks outside-of-the-box for solutions.

One way to start the discussion for change, is to look to other risk-seeking funding models for inspiration. One place where this could be warranted isventure capital. As the most mature funding model that constantly transacts in risk, lessons from how VCs fund startups may be instructive for science.

A comparison of the two provides a foundation for debate around what I call“a venture capital approach to science grantmaking” .

Comparing venture capital investments vs science grantmaking
The VC ecosystem today is highly differentiated and works as a funnel. The top of the funnel is loaded with a large number of risky startup ideas that all receive small amounts of funding.

As ideas become validated with real outcomes data, only the best move onto the next stage of financing. Each stage of financing brings its own types of investors specializing in certain risk-return profiles.

Current VC Funnel For Investments

While the venture capital “funnel” shares some common characteristics with science — mainly, in its segmentation of risk into tiers, there are some notable differences. The main one being: the science funnel (today) is inverted with no robust support (in funders and funding) for “pre-seed” stage research ideas.

Note: Pre-seed stage research ideas are scientific hypotheses that just need a little bit of funding to get some preliminary data — helping to see if the line of research is worth further exploring.

Current Science Funnel For Grantmaking

Many traditional science funders have noted that there are practical constraints to vetting and managing a portfolio of research grants fitting the pre-seed description.

Therefore, changing the risk paradigm in science starts with addressing this crucial gap in the pre-seed space. We must not be afraid to experiment with new and oftentimes unconventional approaches. By addressing the challenges at the top-of-the-funnel, we should see a positive trickle-down effect through the rest of the ecosystem.

Fixing the top-of-funnel by going smaller
Funnels in venture capital work because they help to segment risk, standardize expectations, and most importantly, create right-sized investment opportunities at each risk tier. At the pre-seed stage, small investments are critical for uncovering just enough field data (per idea) to help get to a go/no go point on more significant investment.

Also, by going small, investors can diversify over a larger number of investments, while at the same time providing more “at-bats” for entrepreneurs.

Creating a similar dynamic for science practically means more experimentation with grant structures that are smaller in size, more concise in scope, shorter in duration, and less of an administrative burdento secure.

Those (funders) willing to take up the challenge and opportunity of funding pre-seed research, will need to re-think grantmaking conventions on a number of dimensions:
  • Grant size: What is the appropriate “small” grant size for a pre-seed and seed stage research idea? Does it differ by field? How can we make $5,000, $25,000, $100,000 work consistently?
  • Peer-review: How should we engage “experts” to evaluate high-risk ideas in a constructive and cost-effective manner? Should we engage non-experts? Can crowdsourced review, cross-disciplinary review, and endorsement-based review fit in?
  • Due diligence: How can we design a process where “vetting effort” is inversely proportional to the risk, and proportional to the size of the grant? This means considering short proposals, with fast decision-turnaround times, etc.
  • Overhead: What is the appropriate level of upfront overhead for small grants so as not to stifle innovation? Can overhead be “fairly” recovered from larger follow-on grants if the preliminary ideas are successful?
  • Follow-on funding: How much funding should be made available if the preliminary idea is successful–either by the original funder or other funders within the ecosystem? How can we create more co-ordinated funding partnerships across the ecosystem?

New funding mechanisms
In addition to technical changes and experiments that need to be run when funding smaller, the mechanisms that surface ideas, allocate the funding, and manage the portfolio of research projects will also need to evolve accordingly.

In venture capital, the mechanisms that have been effective in servicing the earliest parts of the funding funnel are accelerators and crowdfundingplatforms. These are distributed community-centric funding mechanisms that involve diverse groups of stakeholders — including domain, and sometimes non-domain experts [1].

Adapting these mechanisms for science may be a good starting point for operationalizing the funding of higher risk early-stage research.

Accelerators: a local focus
Popular startup accelerators (Y Combinator, Tech Stars, etc) bring together startups in batches, in one physical location, and for a short period of time. They surround them with the tools/guidance to develop their ideas (usually a product), and seed them with a small amount of funding to get started.

The environment is collegial, and fosters a strong sense of community amongst the participating entrepreneurs that helps bring up the quality of the companies/ideas.

The general partners running the accelerator are typically long-time startup veterans who source and vet applications for the accelerator, help entrepreneurs think through their challenges, and act as a sounding board to help them see the forest from the trees. They are not there to prescribe step-by-step guides.

A similar setup can be employed for science — building on the types of research collaboration that already happen within universities, colleges and public lab spaces.

Science funders (be it philanthropy or government) can explore partnerships with groups of local institutions to design accelerator programs targeting early-stage research ideas.

Like a tech accelerator, small funds (grants) would be dispersed to a number of scientists (which can be from all experience levels). Professor emeritus/retired scientists associated with these local institutions would be a possible choice as the general partners running the programs, and helping with the process for determining who’s included in each funding batch.

Institutions benefit from this type of partnership, as it strengthens their own early-stage research pipeline, and provides a possible avenue for winning larger grants later on. Funders benefit, as this is a source of local “deal flow” that targets different grant risk-return profiles that they wouldn’t normally have access to.

Crowdfunding platforms: a global focus
AngelList (equity), Kickstarter (donations), Indiegogo (donations), and other platforms have forever changed the landscape for funding startups — or products that later become startups. Crowdfunding platforms rely on the crowd to vet and fund early stage high-risk ideas.

Many of the projects/deals on these platforms are for preliminary (sometimes just “back of the napkin”) ideas that are inherently risky. In the case of the donation-based platforms, some plain-English text and a video explaining the idea is all that’s provided for funders to make decisions.

This is something that is in its nascent stages for science but could have significant implications as platforms scale up.

For example, crowdfunding platforms, opposite of accelerators, are built to be geography agnostic. They allow ideas and funders from around the globe to participate in proposing and funding ideas.

Furthermore, crowdfunding that employs an all-or-nothing (AON) funding approach double as a marketplace that can vet/vote on high-risk ideas at a low cost (addressing some of the above questions around the technical elements of small grants).

Larger science funders could leverage crowdfunding by engaging with platforms to use their community or “crowd” to complement their top-down decision-making (balancing biases), and as a source of global “deal flow”.

This practically would mean experimenting with matched funding, or setting aside portions of their funding for the crowd to allocate on their behalf. The diversity of the community forming around each project would be assets for supporting the scientist, critiquing the science itself, and as an audience for science communication efforts.

In closing
Changing the risk paradigm in science is complicated, and will involve experimentation and new thinking.

Looking to venture capital gives us one model that can be used as the basis for things that could be tried. A venture capital approach to science grantmaking points to the need for new grant structures that target bigger risks in smaller more and incremental ways, and supported by funding mechanisms that are fit-for-purpose.

Though we have a ways to go, the silver lining in all of is that we have the tools, talent, and desire to make this change. We just need to take that first step, and take some risks.

[1] Any solution for science funding must also consider one important caveat differing from the investment world: it should try to engage a broader audience in the funding decision-making process.

As an example, the philanthropy of billionaires that directs what research gets funded and what doesn’t is a growing concern from a funding “systems” point-of-view.

Thanks to Trevor for reading versions of this.