This article originally appeared in the Behavioral Scientist online magazine. [editor’s note]

On Thursday afternoons throughout the fall, Brittany and John meet to check in. Brittany is a senior in high school and John is her counselor from Bottom Line, a college success organization. They talk about which colleges are a great fit for her interests and affordable for her family, work on her applications, and complete financial aid forms so Brittany will receive grant money for college. Numerous times throughout senior year Brittany and hundreds of other students will meet with their Bottom Line counselor one-on-one, and they’ll continue to be supported by Bottom Line coaches throughout college.

Meanwhile, over 100,000 high school seniors across the country are receiving automated text messages from Michelle Obama’s Up Next campaign, which I helped develop, reminding them about key steps in the college and financial aid application process. “You took responsibility to apply to college. Now take control of college costs by applying for financial aid. Ready to start FAFSA this week?” Students across the country will receive biweekly messages throughout the fall to make sure they have the information they need to get to college.

Bottom Line and Up Next are two examples of a multitude of initiatives across the country that share a common goal: reducing inequalities in college completion. More than half of high-income youth will earn a bachelor’s degree by age 25, compared with less than one in 10 of their low-income peers. And the financial consequences of earning a college degree have never been more profound: Adults with a college degree earn almost $20,000 more each year than adults with some college but no degree.

The financial benefits of earning a degree notwithstanding, low-income students face a variety of hurdles that can prevent even hard-working, academically-talented youth from earning their diploma: an inability to afford college even after financial aid has been applied; a lack of access to advising to navigate the intricate array of major and course requirements to complete a program of study; competing work and family demands that limit how much time students can invest in their education.

As Bottom Line and Up Next illustrate, we don’t lack evidence-based strategies that substantially improve college persistence for low-income students. Bottom Line, for instance, generates substantial improvements both in the quality of institution that students attend and in whether they persist in college. By providing students with a combination of financial supports, intrusive advising, and structured guidance on what courses to take, The City University of New York Accelerated Studies in Associates Program (CUNY ASAP) has doubled associate degree completion rates. Even low-cost, highly-scalable interventions can generate meaningful improvements in postsecondary outcomes. Sending community college freshmen text message reminders to renew their financial aid increases the share of students who persist through the second year of college by almost 25 percent.

Especially with proposed cuts to federal funding for education, the lower-cost interventions, like Up Next, have garnered substantial interest among policymakers. After all, these strategies check off several important boxes: They’re low cost, they’re scalable, and they improve student outcomes. What’s more, they offer creative, clever designs.

But there’s much we still don’t know about how these lower-cost alternatives compare to more resource-intensive interventions. Do they produce lasting benefits for students? A few high-impact instances notwithstanding, are low-cost interventions just tinkering at the margins of reducing stubborn and persistent disparities in higher education?

High-impact interventions like CUNY ASAP have game-changing effects on completion, but they come with a price tag that can run into five figures. Are they viable without extensive and sustained philanthropic support? Do we sufficiently understand the “secret sauce” underlying their success, in order to replicate these models in new settings?

The path to helping more students complete college requires innovative solutions that combine the best of what low- and high-cost interventions have to offer.

On one hand, we should pursue opportunities to dial up the impact and resonance of low-cost interventions. One approach is to use modern interactive technologies not just as channels through which to convey information to students but also as “just in time, just in need” chances to connect with a college adviser. For instance, Google has begun incorporating data from the United States Department of Education College Scorecard into the facts it provides when people search for information about colleges. A major advance would be to identify users whose search profile suggests they are from disadvantaged backgrounds and, when they search for colleges, make live coaching available from partnering college advising organizations. This kind of “just-in-time” advising—particularly at the moment that students are immersed in complicated application processes and critical postsecondary decisions—could help students identify colleges to apply to where they have a good chance of getting in and graduating, at a price point that is affordable for their families.

We can also strengthen low-cost interventions by following the lead of the Amazons and Netflixes of the world. In the same way that Amazon uses data science to recommend purchases, we can leverage advanced data analytic techniques and partnerships with educational agencies. Based on trajectories of similar students that came before, we could provide students with highly personalized information about academic pathways they pursue or resources they access. These personalized academic pathways could help students identify the courses they need to finish their program of study, which would reduce the share of students who have earned substantial credits but who drop out before completing their degree.

We also need to pursue creative and novel approaches to scaling high-impact but resource-intensive interventions. In particular, we should develop new models to untether costly interventions from their reliance on philanthropic support. We can draw on modern models for funding public- and social-sector programs, such as pay for success and performance based funding. Some states, like Tennessee, have adopted performance-based funding models to finance higher education. Especially in these states, we should be able to develop financial models where, for a sufficient increase in college persistence or completion, institutions realize a sufficient performance funding premium to cover the costs of repaying a philanthropy or investor that covers the up-front costs of a high-impact intervention.

Our policy imperative is clear. To support people to lead productive, fulfilling lives and to satisfy 21st century workforce needs, we have to substantially increase the share of Americans with a college credential or degree. We need creative, disruptive solutions that make a real and lasting dent on inequalities in college completion, are financially sustainable, and have a clear path to scale.