2021-04-12 · 17 min read

Vermont: Building the Ecosystem

Vermont: Building the Ecosystem

I've spent the last several years quietly trying to figure out a question I don't fully know how to answer: how do you grow a real startup ecosystem in Vermont?

Not "how do you copy Silicon Valley." Not "how do you attract one big company." How do you grow the thing (the organism, the network, the long arc) in a place with 645,000 people, a hundred miles of mountain in the way of any drive south, and a deeply, defensibly weird culture?

This post is my working draft of an answer. It's a synthesis of three threads that have been knocking around in my head: Brad Feld's Startup Communities and Startup Community Way, Donella Meadows' Thinking in Systems, and the actual on-the-ground players I've been mapping with help from operators across the state. I'm publishing it because the only way to improve a draft like this is to put it in front of the people who can tell me where it's wrong.

If you live or work in Vermont, as a founder, investor, university administrator, mayor, banker, or interested citizen, I want your feedback.

Key takeaways

• An ecosystem is a complex adaptive system, not a project plan. You guide it; you can't control it. Treat it like a garden, not a factory.
• Founders lead; everyone else feeds. When feeders try to lead, ecosystems calcify.
• The real leverage is high on Meadows' list, information flows (#6) and the paradigm (#2), not the tax credits and grant sizes (#12) where most regions spend their energy.
• Vermont's most active angels are operators from a handful of graduate companies. That alumni-reinvestment loop is the single most important thing to speed up.
• The horizon is twenty years. Any honest plan is a posture, not a roadmap, and its biggest failure mode is feeling productive for years while producing nothing measurable.

I. Why Vermont, and why now

Two reasons the timing is right.

First, the demographic and cultural baseline matters more than it used to. Remote work is real. Founders are no longer required to live within a fifteen-minute drive of Sand Hill Road. The cost-of-living calculus that pulled a generation to the Bay Area has reversed. People are making choices about where to live before they make choices about where to work, and Vermont has the kind of place-quality (landscape, schools, food, civic intimacy) that wins those choices.

Second, the Midwest has shown what regional renewal looks like. A 2020 Economist special report on the Midwest argues that the region's quiet revival rests on five pillars:

1. Revived downtowns with real city living
2. Cultural expansion (food, arts, music, identity)
3. Diversifying economies beyond a single dominant industry
4. Thriving universities that act as both magnets and pipelines
5. Liberalization and openness: drug reform, immigration, a more porous social fabric

Vermont scores well on three of those (downtowns, culture, openness), middle-of-the-pack on universities (we have a lot of them, but they don't yet behave like ecosystem catalysts), and historically weak on economic diversification (we lean hard on tourism, agriculture, and a thin manufacturing base, with one or two outsized employers per region).

The opportunity is in the gap. The pillars work together. A diversified economy without a university talent pipeline stalls. A university without a vibrant downtown loses its graduates to other cities. The interesting move is to invest in the connective tissue, the loops, not any single pillar.

II. What the goal actually is

It's worth being explicit about what we're trying to do, because "support startups" is the kind of phrase that means everything and nothing.

Desired outcomes, in priority order:

• Economic development and diversification. More kinds of companies, in more parts of the state.
• Social impact. More agency, more livelihoods, more young people who choose to stay.
• Job creation. Especially in places that have lost jobs to consolidation or automation.
• Financial return. For local investors, so capital recycles instead of leaking out.

Method: the primary lever is to support founders, entrepreneurs, and the startups they build. Everything else (capital, real estate, services, programming) exists to make founders' lives easier. Not the other way around.

This sounds obvious. It is in fact the thing most ecosystem efforts get backwards. We'll come back to that.

III. Three principles I'm not willing to compromise on

These are stolen wholesale from Brad Feld, who has been writing about regional startup communities for two decades. They aren't original. They are correct.

1. Founders lead. Everyone else feeds.

Feld's central insight in Startup Communities is that an entrepreneurial ecosystem must be led by entrepreneurs themselves. Investors, universities, government agencies, economic-development orgs, service providers, large corporations: these are feeders. Their job is to support, to be useful, to get out of the way. When feeders try to lead, ecosystems calcify. When founders lead, things grow.

This is uncomfortable for some of the most well-meaning organizations in Vermont, because the natural instinct of a feeder is to coordinate, to convene, to offer programs. Coordination is fine when founders ask for it. Programs are fine when founders show up to them. But the ecosystem's center of gravity has to sit with the people doing the actual building.

2. Twenty years.

Real ecosystems run on twenty-year cycles, minimum. Boulder didn't become Boulder in five years. Silicon Valley took thirty before anyone called it that. Anything we build here has to be built by people willing to commit for the long haul, and we have to evaluate ourselves on twenty-year horizons, not on the metrics that fit nicely into a state grant cycle.

This is probably the hardest principle for institutions to swallow. Funders want measurable annual outcomes. Politicians want ribbon-cuttings before the next election. Universities want enrollment lifts. None of those clocks match the clock that ecosystems actually run on.

3. Open and inclusive, by default.

If you want to participate, you can. Founders, students, investors, hobbyists, retirees, immigrants, people who just moved here, people whose families have been here ten generations. All in. Anybody who tries to gatekeep is doing damage, even when the gatekeeping is well-intentioned.

This includes being deliberately welcoming to immigrants and to founders moving from outside the state. The thing that makes a community robust is the rate at which new people can plug in and start contributing. Networks that close themselves off die.

IV. Why systems thinking is not optional

Here's where I want to slow down, because this is the part I think most regional ecosystem efforts get wrong.

A startup ecosystem is a complex adaptive system. That's not a metaphor. It's a technical claim with specific consequences.

Donella Meadows' Thinking in Systems is the canonical primer. The book reframes how to look at any system, biological, economic, social, and a startup ecosystem fits the definition cleanly: a set of interconnected agents (founders, investors, customers, employees, universities, regulators) producing characteristic patterns of behavior over time, through stocks (talent, capital, knowledge), flows (deals, hires, exits, learnings), and feedback loops (success attracting more talent, failure compounding into stigma, etc.).

Four implications that matter for what we do next:

1. You can guide the system, but you can't control it.

Complex systems are nonlinear and evolutionary. You set up initial conditions, you remove obstacles, you reinforce the loops you want, but you cannot dictate outcomes. Anyone who tells you they have a five-year plan to produce N unicorns in Vermont is either lying or doesn't understand what they're doing.

2. The system is emergent.

Emergence is the property of complex systems where the behavior of the whole cannot be derived from studying the parts in isolation. Termites build elaborate cathedrals nobody designed. Traffic jams propagate through networks even though no individual driver wants to be in one. A startup ecosystem produces innovation clusters, founder lineages, capital networks, and cultural norms that no participant invented. They show up because the conditions allowed them to.

Vermont already has a small, concrete instance of this. Nobody sat down in 1990 and decided that dealer.com, IDX, and BioTek would become the source of half the angel checks written in the state thirty years later. That network of operators-turned-investors wasn't designed; it emerged from those companies scaling, exiting, and minting alumni who stayed and reinvested. You can see the loop, but you couldn't have drawn it in advance, and that's exactly the point about what we can and can't engineer.

The practical consequence: you can't predict, and therefore you can't fully measure. The best you can do is run many small experiments, watch the system respond, and double down on what's working.

Treat the ecosystem like a garden, not a factory.

3. Loops, not lines.

Most ecosystem strategy decks I've seen are linear: "Step 1, fund accelerators. Step 2, attract talent. Step 3, exits." Systems don't work like that. Every output is the next input. A successful exit creates angel investors who fund the next cohort. A high-profile founder failure creates lessons (and possibly stigma) that shape who tries next. A new co-working space changes the rate of accidental collisions, which changes the rate of new partnerships, which changes the deal flow seen by local investors. Reasoning in straight lines will systematically mislead you.

4. Leverage is not where you think it is.

Meadows' single most-cited contribution is her twelve leverage points, places to intervene in a system, ranked from least to most powerful. Here's the ranking, simplified:

Meadows' point, and it lines up with everything I've seen in twenty-plus startup-ecosystem case studies, is that almost all of our energy goes into the bottom of this list, and almost all of the leverage is at the top.

Most regional development efforts spend 95% of their time on #12 (parameters: tax credits, subsidy amounts, grant sizes). These are the easiest to change and the least powerful. Real leverage sits in #6 (who has access to what information), #5 (the rules of the game), #4 (whether the system can evolve itself), #3 (what the system is actually trying to do), and #2 (the paradigm, the unspoken shared belief about what's possible here).

A small example. Suppose you want more deals to close in Vermont. The reflexive lever is #12: subsidize the deal, offer a tax credit, write a grant. The high-leverage move is #6: make sure every founder in the state knows every active investor, every weekly office hour, every pitch event, every recently funded peer who'd take their call. Information flow is cheaper, faster, and more durable than money.

A bigger example. The paradigm in Vermont, the unspoken shared belief, is roughly: "This is a quiet, beautiful place where you can have a nice small business and a nice life." That's a lovely paradigm. It's also a ceiling. As long as it's the dominant story, ambitious founders will continue to leave for places where the paradigm is "This is where world-changing companies get built." Shifting that paradigm (slowly, by example, by celebrating the founders who've built outsized things from here) is leverage point #2. It's plausibly worth more than every grant program in the state combined.

V. What is a startup, anyway

Worth defining, because the word has gotten loose.

"A startup is a temporary organization used to search for a repeatable and scalable business model." (Steve Blank)

"A human institution designed to deliver a new product or service under conditions of extreme uncertainty." (Eric Ries)

When I say "startup," I mostly mean software startups, because they have low-to-zero marginal costs and can therefore scale supply quickly if there's demand. That's what makes them distinctive economically. A bookstore can be excellent. A bookstore is not a startup.

This matters in Vermont, because the small-business support infrastructure here is excellent, and largely irrelevant to the kind of company we're trying to grow more of. SCORE, the SBDC, the local credit unions: these are wonderful institutions for the kinds of businesses they serve. They are not the right tool for a software founder who needs $500K of risk capital, technical co-founders, and a path to a $100M+ outcome. We need both kinds of infrastructure. We tend to act like we have both when we have only one.

(I wrote a longer piece on what a startup actually is, if this section is too compressed.)

VI. The external factors that matter

You can't engineer a startup ecosystem, but you can attend to the soil. Five conditions matter more than any program:

Non-extractive partners. Large corporations should treat local startups as collaborators, not as cheap suppliers or acquisition targets. The same goes for landlords, law firms, and consultants. Extractive behavior compounds; so does its opposite.

Labor mobility. People should be able to move easily between startups, big companies, universities, and government. Bad noncompete laws kill ecosystems. Vermont is reasonably good here, but the cultural expectation of long tenure can be its own friction.

A diverse workforce. Different backgrounds, experiences, and viewpoints. Vermont's demographic homogeneity is a real disadvantage, and one that gets worse if we're not deliberate about immigration and inclusion.

Density. Co-working spaces, vibrant downtowns, office parks, neighborhoods where you can run into another founder at the coffee shop. Density doesn't have to mean Manhattan. It means the probability of useful collisions per square mile per hour is high. Burlington has it. Montpelier and Brattleboro can have it. Most of the rest of the state will struggle without deliberate effort.

Quality and love of place. This is Vermont's superpower. Almost every entrepreneur who comes here for a long weekend leaves slightly in love. The pitch writes itself. We just have to make the next step (actually moving here, actually building here) feel possible and welcoming.

VII. Mapping the players

Below is the working market map. Treat this as a snapshot in spring 2021, definitely incomplete, definitely getting updates.

What the categories are

Founders/entrepreneurs. The leaders. Everyone below is a feeder.

Startups, growth-stage companies, large corporations. The core operating layer.

Mentors, coaches, SMEs. Pay-it-forward operators with experience to share.

Investors.

• By stage: friends & family → angels → seed → growth → public.
• By type: equity, debt, grants, revenue-based financing. The Flexible Capital Fund put together a great Vermont-specific overview of this continuum (courtesy of Janice, thank you).

Incubators, accelerators, innovation hubs, hacker/maker spaces, networking and event organizations. Brad Feld defines an incubator as a year-round physical space and advice in exchange for fee or equity (typically nonprofit, often university-attached). An accelerator runs cohort-based programs of fixed length, with funding at the end. Vermont has more of the former than the latter.

Services. Real estate, accounting, legal, banking, compute infrastructure. Less glamorous, more important than founders usually realize.

Universities. Students, professors, researchers, technology-transfer offices, entrepreneurship programs. Crucial as both magnets and pipelines, hard to do well.

Vermont, today

Angels and angel groups (a partial list). The pattern matters more than the roster: most active Vermont angels are operators from a small number of graduate companies, businesses that scaled here, sold or went public, and produced a generation of newly-liquid alumni who reinvest locally. That's the single most important reinforcing feedback loop in the entire ecosystem. The clusters I can map today trace back to a handful of those companies:

There are at least a dozen more I haven't met yet, but the shape holds: liquidity from a graduate company turns into the next round of local risk capital.

Investors / funds. FreshTracks, Hula, VCET (Vermont Center for Emerging Technologies), VCF (Vermont Community Foundation), the Flexible Capital Fund, Helios.vc, Closed Loop Capital, Vernal Ventures, JH Capital, CORI (Center on Rural Innovation), Underdog Ventures, Manchester Capital Management. Different mandates, different stages, different theses. Worth understanding before you pitch any of them.

Universities. UVM, Champlain College, Middlebury, Castleton, Norwich, Vermont Technical College, Landmark, Goddard, Sterling, plus Dartmouth (just over the river but functionally part of the regional brain trust). Each has different entrepreneurship maturity. The biggest opportunity I see is in the connective tissue between schools, which today barely exists.

Startup support orgs. LaunchVT, VCET, Hula (co-working + offices), DeltaClimeVT (climate-focused accelerator), BTVIgnite, Generate, Center for Women & Enterprise. These are the front doors. Founders should know all of them.

Large corporations. This is where the alumni networks come from, eventually. The list, drawing partly from thinkvermont.com's regional employer index, includes the ski resorts, dealer.com, IDX/Athenahealth, GE Aviation, IBM, Green Mountain Coffee, Brueggers, Huber+Suhner, Casella, Mack Group, Burton, Ibex, Ben & Jerry's, BioTek, GW Plastics, King Arthur Flour, Gardener's Supply, SymQuest, Dynapower, Cabot, Mylan, Orvis, Green Mountain Power, Seventh Generation, Chroma, NRG Systems, GlobalFoundries, and Beta Technologies.

A pattern worth noticing: a striking number of these companies are the kind of employer whose alumni will, in fifteen or twenty years, turn into the next wave of Vermont angels and founders. The ecosystem is already running its reinforcing loop. The question is whether we can speed it up.

VIII. The plan

This is the part where I should be most cautious about specifics, because the system will tell me what to do in ways I can't predict from the armchair. So rather than a Gantt chart, here's a checklist, written so anyone trying to seed an ecosystem in their own region can run it, not just me in Vermont.

The ecosystem-seeding checklist

1. Publish a draft thesis and get it argued with. Put your map in front of local founders and investors. Find out what's wrong before you build on it.
2. Educate yourself from other regions. Talk to operators in ecosystems a decade ahead of yours (for me: Boulder, Maine, the Triangle, Pittsburgh, Bend). Write down what you learn.
3. Share what you learn in public. Document it on a blog, a syndicate (mine is Minnow Ventures), and in person. Information flow is the cheap, durable lever.
4. Prioritize by leverage, not by ease. Of everything you could fund, which is highest-leverage now? My current bet is information flows (#6) and the paradigm shift (#2); everything else is downstream.
5. Recruit allies for the long haul. Find the people willing to commit for ten or twenty years and okay being unglamorous about it.
6. Start one small experiment and watch the system respond. Begin on the highest-priority activity, then double down on what works.
7. Repeat for twenty-plus years.

That's it. It's not a plan in the management-consulting sense. It's a posture, a stance toward an enormous, slow-moving, nonlinear thing.

I should name the failure mode of my own thesis honestly, because it's a real one. Betting on paradigm and information flow is close to unfalsifiable: the work can run for years, feel productive, generate good conversations and a tidy blog archive, and still produce nothing measurable. "You can't fully measure it" is true, and it's also exactly the alibi a stalled effort would reach for. The discipline is to keep looking for the small, real signals (a new angel writing her first check, a founder who stayed who would have left) and to treat their absence over a long enough window as evidence I'm wrong.

IX. What I'm asking for

If you've read this far, you probably have an opinion about something I got wrong. Tell me. I want to hear:

• Founders in Vermont: what's the friction in your week that an ecosystem could fix? What's the thing you'd want to exist that doesn't?
• Vermont angels and investors: who am I missing on the map? What deals have happened that I should know about? What's the rate of capital recycling actually doing?
• University people: where's the real entrepreneurial energy at your school, and what's blocking it?
• Operators in other ecosystems: what did your region get right in its first ten years that we should steal? What did it get wrong that we can avoid?
• Skeptics: where is this analysis most likely to be wrong? I'd rather hear it now than five years in.

You can reach me at the email on my contact page, on Twitter (@0xMorgz), or by replying to the Leftovers newsletter.

I'm going to keep documenting this on the blog as I go. If you do one thing, forward this to someone who should be in the conversation. The most important thing about an ecosystem, after all, is the rate at which new people can plug in and start contributing.

Twenty years. Let's get to work.

Sources and further reading

• Brad Feld, Startup Communities and The Startup Community Way
• Donella Meadows, Thinking in Systems: A Primer (Chelsea Green, 2008)
• Donella Meadows, "Leverage Points: Places to Intervene in a System" (1997)
• Steve Blank, "'Startup Communities' delivers roadmap for Silicon Valleys everywhere" (VentureBeat, 2012)
• Bailey Richardson et al., Get Together: How to Build Community with Your People (Stripe Press, 2019); via First Round Review
• The Economist, "A region with outsized punch" (Midwest special report, July 2020)
• Wikipedia: Twelve leverage points, Emergence
• The Flexible Capital Fund's Vermont capital continuum overview (2017)
• ThinkVermont's regional employer index
• My own Startups 101 (forthcoming)