Charity effectiveness evaluator GiveWell has won a lot of fans for its data-driven recommendations about where people should donate to have their money used most effectively. The organization usually recommends nonprofit groups that offer fairly inexpensive and direct remedies to solve global problems on a major scale. That’s because the group adheres to a philanthropic philosophy called “effective altruism,” which means it encourages donors to make investments that do the most quantifiable good per dollar invested.

While such arithmetic has endeared the group to Silicon Valley givers, who appreciate the rational approach, it has also affected the GiveWell’s ability to nurture another Silicon Valley darling: promising philanthropic startups, which may not have the track record to generate a positive rating, but do have a lot of potential.

Since the group launched in 2007, each solution they endorse has had to be measurable in either lives saved or a substantial standard of living increase. That means it must be proven in randomly controlled trials, and in a way that shows that the group delivering the solution has the capacity to replicate it consistently and effectively. “We had been recommending insecticide-treated nets to prevent malaria deworming treatments and direct cash transfers,” says Catherine Hollander, a research analyst with the group. “And so we wondered whether there might be something we could do to create or find additional top charities.”

The result is an emerging program called GiveWell Incubation Grants. The group awarded three in 2014, three in 2015 before ramping up to give eight more last year. In total, the group had awarded more than 11.5 million by the end of 2016, with a couple more groups earning endowments this year. All of it is backed by GoodVentures.

“The goal the program is to support the development of potential future top charities and we see this happening in a few different ways,” says Hollander. “One is by supporting research that our charity recommendations rely on such as randomized controlled trials of programs that we’re interested in. And another is by providing startup funding to a group that’s getting off the ground and implementing a program that’s evidence backed.”