What We Learned from 40 Program Experiments

Before Pivot opened enrollment, we ran an exercise that a lot of nonprofits skip: we designed, scoped, and in many cases piloted 40+ distinct programs. Most never made it out of the design room. A handful became the programs we’re launching now. The rest taught us something valuable about what works and what doesn’t.

This piece is for anyone — nonprofit operators, board members, funders — interested in how you can deliberately decide what to build instead of reactively launching whatever seems hot.

How we generated the 40

The initial list came from three sources:

1. Founder instinct. Ideas that came out of Bob DeLisa’s experience — things he’d seen work or fail in other organizations, or gaps he’d noticed while operating other businesses.
2. Community input. Conversations with prospective participants, mentors, and local nonprofits about what wasn’t being served.
3. Cross-sector scan. We looked at what was working in adjacent sectors (workforce development, financial literacy, disability services, veteran support) and asked which of those could be adapted to our context.

The result was a long list divided roughly into what we now call PivotBiz (entrepreneurship, business) and PivotLife (personal development, wellness, family, youth).

The scoring framework

We didn’t rank the 40 by gut feel. We built a framework with five criteria, and we scored each program on each:

1. Does the need exist? — Is there evidence that the population we’d serve actually has this need, or are we projecting?
2. Can we serve it without duplicating? — Is there already a strong program in this space, or is this a genuine gap?
3. Can we staff it with Champions? — Is there a population of experienced operators who could lead this program, or would we have to hire specialists?
4. Can AI meaningfully help? — Does this program benefit from the AI-native approach we’re building, or would it be “just another program with a chatbot”?
5. Does it compound with our other programs? — Does this program create a graduate who becomes a candidate for another Pivot program, or does it dead-end?

That last one mattered more than we initially realized. Programs that compound with each other create an ecosystem. Programs that don’t compound create a portfolio of isolated silos.

What made the first cut

Five programs scored highly across all five criteria. These are the ones we’re launching first:

• LaunchPad — Our flagship entrepreneurship training
• Second Chance — For veterans and people rebuilding after setbacks
• BizBoost — Growth-stage business acceleration
• PivotConnect — The mentorship matching platform
• Attracting Money — Financial literacy and wealth-building

Each of these met all five criteria, and — critically — each had a specific Champion candidate in mind before we decided to launch. We don’t launch programs looking for leaders; we launch programs with leaders ready to go.

What didn’t make the cut (and why)

The cuts are more interesting than the keeps. Here are a few instructive cases:

Programs we killed because the need was already well-served:

• A general-purpose small business legal literacy program. Local bar associations and Small Business Development Centers already do this well. Adding a Pivot version would have been duplication for its own sake.
• A basic digital skills bootcamp. Goodwill, public libraries, and dozens of community colleges offer this at no charge. We’d add no value.

Programs we killed because we couldn’t staff them well:

• A program focused on scientific founders (biotech, med-tech, hardware). We couldn’t identify a Champion with the operational background to lead it without us funding a full-time hire. Maybe someday, not now.
• A specialized export-to-Asia program. Real need for some of our target population, but the Champion pool for that specific expertise is tiny and we couldn’t find the right person to lead.

Programs we killed because they didn’t compound:

• A one-off entrepreneurship speaker series. Compelling idea, but graduates of a speaker series are no better set up for a deeper Pivot program afterward than people who never attended. The output was content, not a participant pipeline.
• A content library for aspiring entrepreneurs. Same problem. Consumption doesn’t compound into engagement.

Programs we parked (not killed) for later:

• Several specialized pathways in the PivotLife division — things like Single Parents Hub, Kids Corner, Teen Tribe. Real needs, real audience. But launching them in parallel with our entrepreneurship programs would have spread the team too thin. They’re on the map for future years.

The unexpected winners

Some programs scored surprisingly well and ended up on the launch list partly because they did something the obvious programs didn’t:

• LaunchPad EmpowerAbility — Accessible entrepreneurship for people with disabilities. This program is a clear example of where our AI-native approach genuinely matters. Custom accommodations, adaptive pacing, alternative input methods — all are dramatically easier in a world with modern AI. It scored high on “can AI meaningfully help” in a way that separated it from programs where AI is a nice-to-have.

• Attracting Money — Financial education is widely available, so we expected this to score low on the “already served” criterion. But our target population — people who’ve been historically underserved by traditional financial education — has specific gaps that generic programs don’t address. Plus there’s strong compounding: financial literacy participants frequently become candidates for LaunchPad or BizBoost.

The lessons for other operators

If you run a nonprofit or are considering starting one, a few things we’d pass along:

1. Generate more ideas than you can launch. The act of scoring 40 programs against each other taught us more than if we’d picked five and run with them. The contrast illuminated which programs had real signal.

2. Use a framework, not vibes. Vibes will tell you the flashy programs are the best ones. Frameworks will tell you that the compounding programs are the best ones. Those are different answers.

3. Staffing is a constraint, not a detail. Programs without the right Champion ready to go are prayers, not plans. We killed multiple strong programs because we didn’t have the right person — and that’s been the right call.

4. Compounding beats charisma. A charismatic idea that doesn’t feed into your other work will win attention and lose time. A quieter idea that graduates people into your next program will build the ecosystem you need.

5. Park, don’t discard. Our “not now” list is as important as our “yes” list. Many of those programs will launch in 2027, 2028, 2029 — once we have Champions, have learned from the first wave, and have the capacity to operate them well.

Where this leaves us

We’re launching five programs in 2026, not 40. The long map still exists, and we’ll fill in more of it every year. But we know why those five — and we know what it took to earn each of them a spot.

This is how we’ve learned to decide.

🦅

---

Related: Why We Chose Not to Scale in 2024 · The Champion Model · Why AI + Entrepreneurship Isn’t Optional Anymore