Community buy-in is more important than efficiency in projects like this.

"The process started with the community-based participatory planning," WFP program policy officer Bakalilou Diaby shares in the video. "By the end of this process, it was agreed that one of the major action is the land reclamation or land recovery project."

At first, it took some time to convince the community that the regreening of the degraded landscape was even possible, but after learning about how to improve the land, "the people believe and they are convinced, and they are also committed," says Diaby.

Sure, these projects could be done more efficiently by one construction company with heavy machinery, but that takes dollars these dirt poor areas don't have (and I think Trump's bullshit arbitrary defunding of USAID is an object lesson for developing countries not to rely on foreign charity for anything important, among many, many other reasons why foreign charity is bad for development). So local people need to know how to do it with local resources in order to expand the projects to new areas.

Just as important, local communities need to support and maintain these projects in the long run, and sweat equity is a great way to build commitment.

Something along the lines of "hiring heavy equipment, then buying a ticket to the gym, everytime you need this done because you never learned anything else". My neighbours tell me I need heavy equipment to fell my forest, absolutely not, I'll go on foot, respect and fell every single tree I plan to fell, leave no tracks, become stronger and wiser, all for free.

Heavy equipment is not cheap, too, especially if you need an operator with skill comparable with attention to small details that manual laborers have naturally. And it does not spread virally like skills. And it burns carbon and leaves tracks. It has its uses at scale, but not in pilots. Same thing everywhere: I have a pick-and-place robot to assemble my electronics, not even turning it on before I have at least 100-ish boards to assemble; it's not expensive, it's easier than manual labor - but you've got to know the system you build before giving it away to a machine, or you'll have a long and expensive debug session ahead of you, even if you are certain know everything about it, which is totally not the case here.