Last week, the Federal Government’s Office of Science and Technology Policy (OSTP) released a Request for Information (RFI) on how the government can best incentivize and speed up the creation of “high impact” learning technologies. Although this is a noble goal, the details of how it will get there merit deeper consideration. Luckily, it’s taking input from experts in the field from now through March 7.

OSTP is approaching the Byzantine edtech market and paltry supply of education R&D with a particular economic principle in mind: pull mechanisms. Push and pull mechanisms are two distinct approaches to supporting new market activity. In OSTP’s words, “push programs pay for research inputs; pull mechanisms pay for research outcomes.” Push mechanisms typically aim to reduce the cost of R&D by directly funding research, regardless of what comes out of that work. Pull mechanisms, on the other hand, are structured to incentivize private sector engagement and competition by creating viable market demand for specific products. The X-Prize is a good example of a pull mechanism, as are government Challenge Grants and social impact bonds.

Pull mechanisms rely on establishing a clear end goal and setting innovators loose to get there. As OSTP considers these parameters, the government’s end goal ought to aim to foster two distinct forms of innovation. A healthy marketplace needs both sustaining and disruptive innovations—that is, innovations that improve on the existing education system, and innovations that offer something cheaper and more accessible outside of the performance metrics of the traditional system. It’s easy to envision funding the proliferation of sustaining innovations: they are simply interventions that we can overlay on our existing system that can be shown to be more effective and efficient at driving outcomes. Funding disruptive innovations, however, will likely involve different goals and parameters that mark a departure from how the government typically gauges impact and success. Therefore, the guidelines should create separate funds with separate criteria for each type of innovation, to make sure that disruptive and sustaining innovations don’t compete for the same pool of funds. Here are a few areas where I’d hope to push the Administration’s thinking, specifically to cultivate disruptive innovations in the edtech marketplace:

Are new technologies targeting nonconsumers or serving existing customers? A disruptive innovation typically gets it start in areas of nonconsumption by offering something cheaper and more accessible. In any given industry, disruptive technology is actually not as good as what companies historically have been selling to existing consumers, but instead offers something that nonconsumers can afford. Once it gains a foothold in that market, the technology then improves over time to meet the needs of more and more demanding customers. To this end, OSTP should be clear about for whom the new technologies should aim to reach and at what relative price point. The RFI provides a list of examples of areas where technology might have transformative impact: for example, pre-K vocabulary gaps, STEM education, and postsecondary skills certification. Many of these areas actually represent sizeable pockets of nonconsumption where students and their families lack access to educational resources and formal training. Because more demanding customers command higher profits and information on nonconsumers is scarce, however, existing companies and startups alike will be far more likely to target these less demanding customers if the Administration makes that an explicit prerequisite of at least some of the grants and prizes.

Is the end goal personalized learning? The powerful implication of disruptive innovation in education is that these not-good-enough technologies can actually move us away from the factory-based model that our traditional education system was built around. Instead, technologies such as online learning can start to personalize education to each student’s interests and pace of learning. OSTP’s RFI makes some tentative references to personalized learning, which suggests that OSTP would reward disruptive innovations that break the mold of our factory-based school system. But as I summarized above, this would require very specific parameters to guide grantees in terms of whom they should serve and at what cost. Simply alluding to personalized learning without redefining the underlying structure of the education system risks diluting the very concept of personalization.

Can pull mechanisms be patient for growth, but impatient for profit? Competing against nonconsumption and moving disruptively upmarket are critical elements of a successful new-growth strategy—and yet by definition, these disruptive markets are going to be small at the outset. One reason it can be difficult for firms pursuing a disruptive trajectory to attract investors is that they start off with this small market share. Thus, their effectiveness should be measured in their ability to make a profit (i.e., to create a viable, cost-effective product within this albeit small market) rather than to grow quickly right off the bat. If OSTP aims to fund disruptive innovations, then the terms of these prizes or commitments should control for expectations around scale and instead focus on the product’s viability within a more narrow market.

Are we really shedding light on “what works”? OSTP clearly wants to catalyze the creation of learning technologies that can improve student outcomes. The nuance, however, that often escapes the Administration’s efforts to codify “what works” is our ability to investigate “what works, for whom, and in what circumstances.” OSTP’s framework risks grasping for an elusive edtech silver bullet that can work for all students in all circumstances. A more reasonable agenda would incentivize competitors and providers to demonstrate and explain the circumstances where their products succeed and fail. This valuable data would go a long way in building an understanding of personalized education as a portfolio of approaches rather than a single intervention or product.