Four Years In

What I've Learned in the Trenches

Someone reached out recently asking for advice on running a grants program. We ended up talking for an hour. Afterward, I realized I had a lot more to say than I could fit in a single conversation.

So here it is. After four years of running programs, advising protocols, managing grantees, and watching this space evolve as a grantee, a grant manager, a tooling provider, and an advisor across some of the major programs in this space, I have some thoughts.

Some of this might sound harsh. Some of it contradicts the narratives we’ve been telling ourselves as an industry. But I’ve seen what works and what doesn’t.

Full Circle

When I first got into crypto grants, the buzzword was “novel funding mechanisms.” Quadratic funding. Retroactive public goods funding. Conviction voting. Streaming payments. The hope was that we could unlock entirely new ways to allocate capital. That we’d improve on the grant-giving systems that have existed for centuries.

I’ve run a bunch of QF rounds. I’ve participated in retro funding. I’ve seen these mechanisms at play across multiple ecosystems. And here’s what I’ve learned: the most effective way to run a grant program is milestone-based direct grants.

That’s it. We tried all the fancy stuff and came back to what nonprofits have been doing forever.

There’s a time and place for these other mechanisms. Quadratic funding is a great marketing opportunity. The individual grant amounts aren’t huge, but QF creates buzz and excitement and gives an early signal about which grantees can mobilize community support. Retro funding can work for ecosystems with long histories of contribution, where you want to reward past impact. But retro funding can also become a popularity contest. There are tradeoffs to all of these things.

But here’s where we’ve over corrected. Because these mechanisms didn’t live up to the hype, there’s now a bias against them. That’s not right either. I’m still excited about what’s possible. Prediction markets, results-based funding, and mechanisms we haven’t fully explored yet. The mistake wasn’t experimentation. The mistake was expecting any single mechanism to be a panacea. The opportunity now is matching the proper mechanism to the right problem.

So, where are novel funding mechanisms best positioned?

Well, it depends on what you’re trying to achieve and for most grant programs with specific initiatives and targets, the answer is boring: direct grants, milestone-based funding, and knowing what outcomes you want before you start allocating capital.

The Bifurcation of Grants and Incentives

One of the shifts I’ve seen is the separation of grants from incentives. Programs would give out grants to protocols, which would then use those funds to drive user incentives, including liquidity mining, yield farming, and the rest.

Arbitrum’s journey through STIP, LTIPP, and what they called the “Incentives Detox“ is the clearest example of this reckoning.

What happened was a recognition that they’d poured significant treasury resources into programs, but the results were mixed at best. Traditional grant disbursement methods like upfront payments, committee-based selection, and milestone tracking don’t always align with the goals of targeted incentives. While grants excel at funding potential and promise, what Arbitrum ultimately wanted was something more precise: measurable outcomes tied directly to network growth metrics.

So they did something unusual. They stopped most programs for three months. They held biweekly community calls to dissect data from previous programs. I was lurking and listening to those calls. They analyzed what actually drives sustainable adoption versus what just creates temporary spikes from mercenary capital chasing rewards.

The result? A fundamental shift in thinking. Growth teams now have incentive budgets. Grant teams have grant budgets. Different functions, different goals, different mechanisms.

If you’re trying to drive DeFi metrics, that’s an incentives problem. If you’re trying to fund development of specific tooling or infrastructure, that’s a grants problem. We spent a few years conflating these, and it didn’t serve anyone well.

The Busy Person Problem

Some of the best advice I ever got came from running a food bank, not from crypto.

As I described oh so long ago in Sov’s Song I was the board president of a small food bank that was 90 days from going under when I took over. We had $15,000 in the bank. One of our mentors, a trucking and warehouse magnate, told me something I’ve never forgotten: “If you want to get something done, ask a busy person.”

In crypto, we’ve done the opposite.

There’s this positive spirit around engaging our community to help make decisions and provide feedback. We host community calls. We ask for opinions. We create governance forums.

And what happens? The people who show up are the ones with nothing but time on their hands.

I led a governance working group for a major L2 for a while. When I joined, we had weekly calls where people would just vibe with each other. Community members, very junior, lots of time. And I had to ask myself: Is this the best use of our time and resources? Paying people like me to show up and steward random community members through half-baked contributions that I then had to course correct?

I cut the community calls. No value.

Here’s what I’d recommend: seek out subject matter experts and domain experts in specific areas. If you want to drive a particular outcome, find someone who’s doing that well today and pay them to come in and do their thing. That’s worth more than a hundred open community calls where you’re getting watered-down responses from people who are there because they have time, not expertise.

The people who produce real value are busy. They have business models, as individual contributors or companies, to come in and do specific things around impact measurement or business development. When you start being intentional about finding those teams and striking agreements with them, that works way better than hosting calls and hoping good people show up.

Theater

There’s a difference between systems and protocols that are decentralized and decentralizing every decision out to a community. I haven’t seen many places where broad community decision-making has been effective.

I’ve been present in many of the DAOs that have tried. Look at some of the biggest grant-giving DAOs of the last couple of years. Which tokens had the most downside? Often, the same ones that gave out the most money with the least clarity on what it was actually for.

Now you see them pulling back and centralizing around operational structures with more accountability. There’s still a spirit of involving qualified people. We want people who are qualified in specific ways. That’s different from “let’s get community feedback on everything.”

Creating work for decentralization’s sake isn’t decentralization. It’s theater. You end up with someone hosting calls, paying for their time to steward people toward contributions, only for those contributions to be lackluster, and then someone has to course-correct. You’re just creating work.

People, Process, Technology

In the traditional tech space I came from, there’s a concept of people, processes, and technology. The order matters. In crypto, we over-optimize for technology. We have a lot of builders who like to build things. But that’s only one part of the equation.

You can have the best or worst tool in the world, but if you have a terrible or great process, it makes up for it. And vice versa.

I’ve used all the tools. QF platforms. Milestone tracking systems. Forum-based approaches. Spreadsheets. Purpose-built grants management tools. Homegrown solutions. They all have tradeoffs.

The question isn’t “what’s the best grants platform?” The question is “what does your grantee lifecycle look like, and what tools support that?”

I think about this in phases: Discover (how do grantees find you, how do you find them), Source (how do you qualify applicants), Launch (onboarding after approval), Deliver (milestone tracking and accountability), and Expand (retention and follow-on work).

Map your internal processes to these phases. Then figure out what tools support each. Some tools do one thing well. Some try to be full-stack. What matters is that your process is clear. The technology follows.

Treat Grantees Like Customers

This might be the most practical thing I can share: think about your grantee lifecycle the same way a B2B company thinks about customer lifecycle.

The easiest deal to close is the next deal with a current customer. Consulting firms know this. You’re always selling your next engagement while delivering your current one.

The same principle applies to grantees. Not all are created equal. Some you work with and think, “I don’t know if we should have given that grant.” Others you want to work with forever.

During the delivery phase, make that determination. Which grantees do you really want to continue working with? Then be proactive about finding more work for them.

At one program I run, we have grantees we’ve worked with for years. Any time we can slot work for them, we do. We’re not going to put out an RFP and hope a good builder applies when we already know who the good builders are.

This sounds obvious, but I see programs doing the opposite all the time. They finish a grant, wave goodbye, and then wonder why they can’t retain their best contributors. Or worse, they make proven grantees go through the same arduous application process as everyone else because “we want to be decentralized.”

Find your good grantees. Give them more work. Ask them what they want to do next that you haven’t asked them to do yet. That’s how you retain talent.

The (Ongoing) PGF Problem

This is the one that keeps me up at night.

I believe crypto is part of the natural progression of open source. And when you look at how much of crypto is built on open source, we’re standing on the shoulders of giants.

For a while, there was a swell of support for these teams. QF rounds dedicated to open source. Collectives fund core contributors. Corporate programs chipping in. But as grant programs have questioned their value or shut down entirely, support for foundational open source has diminished.

We’ve recreated the same problem we said we wanted to solve.

In Web2, the internet was built on open source protocols. Who maintains those protocols? Mostly the tech giants. There’s value capture happening there. They fund open source tools partly because it lets them influence the roadmap.

We’ve done the same thing in crypto, except we’re not even supporting the teams we should be supporting. We’re not funding the open source foundations of what we claim to be building.

That’s what’s missing. And I don’t have a clean answer for how to fix it.

Where We Go From Here

So that’s the state of grants as I see it.

We’ve come full circle on mechanisms. We’re separating grants from incentives. We’re slowly learning that community engagement without intentionality is just noise. We’re realizing that decentralization isn’t always the answer. We’re starting to think about process before technology. Some of us are treating grantees like customers.

And we still haven’t figured out how to sustainably fund the public goods we all depend on.

We should not solve any of this by launching more programs or building more tools. We solve it by being honest about what’s worked, what hasn’t, and what we actually want to achieve.

Grant programs that started as marketing exercises, to show we’re funding builders and compete on dollar amounts, are getting shut down or restructured. The ones that survive will be the ones that got specific about outcomes before they started writing checks.

Something I’m trying to do more of: just reaching out to other program managers and comparing notes. Learning from each other. It made my day when that conversation happened that sparked this piece. Someone running a program is asking other programs what they’re doing. I don’t see enough of that. If enough of those conversations happen, they may evolve into something greater.

If you’re running a grants program, my advice is simple: Be clear about what you want to achieve. Fund people who can actually deliver. Don’t create work for the sake of decentralization. Treat your grantees like valuable customers. And remember that all the novel mechanisms in the world won’t save you from unclear goals.

We’re still figuring this out. But at least now we’re asking better questions.

Peace,
Sov