Writing this essay requires a discipline that does not come naturally to investors: honesty about being wrong. Venture capital has an occupational bias toward narrative — the stories we tell about our winners. We celebrate the companies that exceeded our projections, the founders whose vision proved correct, the markets that grew faster than anyone predicted. We talk less about the investments where our original thesis was flawed, the founders who were technically brilliant but commercially underprepared, the impact metrics we agreed on at closing that turned out to measure the wrong things.
Sway for Future has been making impact technology investments since 2022. We have twelve active portfolio companies. We have had two investments that significantly underperformed our original projections, and we have learned more from those two than from the ten that have gone well. This essay is an attempt to share the five most important lessons we have drawn from the full portfolio — including the hard ones.
A Brief Note on Context
Our first investment was ClearFlow, a modular water purification company, in June 2022. We backed the founding team because we believed deeply in the problem — water insecurity affects over 2 billion people globally — and because the co-founders had genuine engineering depth in membrane filtration technology. ClearFlow is now deployed in 120 communities across three continents, and it is one of our most impactful portfolio companies by any objective measure.
But we also made mistakes in the ClearFlow investment that, in retrospect, shaped how we think about nearly every subsequent deal. And those mistakes are the substrate of the lessons that follow.
Founder Resilience Predicts Outcomes More Than Market Size
Every venture investor knows that market size matters. At seed stage, you are making a bet that a large market exists and that this team can capture a meaningful portion of it. We spent enormous energy on market sizing in our first four investments — bottom-up models, customer surveys, competitive landscape analysis. And then we watched a founder in our portfolio navigate a regulatory setback that would have ended most companies, and realized we had systematically underweighted the variable that actually determined whether the company survived: the founder's capacity to endure and adapt.
Resilience is not the same as stubbornness. Stubborn founders hold onto wrong hypotheses too long. Resilient founders hold onto mission and core insight while adapting strategy as information arrives. The distinction is important, and it is observable in how founders talk about their failures. Resilient founders describe past setbacks with specificity and extract learning from them. Stubborn founders describe past setbacks as external factors or bad luck.
We now spend meaningfully more time in our diligence process evaluating founder resilience. We ask directly about the hardest professional moment the founder has faced and what they did. We ask how they have changed their thesis since they started the company and why. We ask what they would do if their primary distribution channel collapsed overnight. The answers to these questions are more predictive of outcomes than any market analysis we have ever done.
"The founders who build the most enduring companies are the ones who are still at it ten years in with the same intensity they had on day one — not because things went well, but because the mission is genuinely more important to them than the outcome." — James Okafor, Partner, Sway for Future
Impact Metrics Must Be Agreed Before the Term Sheet, Not After
In our first two investments, we agreed on impact measurement frameworks after the term sheet was signed — during the post-close onboarding process. This was a mistake. When impact measurement is left to post-close negotiation, it tends to drift toward whatever is easy to measure rather than what actually matters. And it creates a psychological dynamic where the founding team experiences impact measurement as an investor imposition rather than a genuine shared commitment.
The consequences of this mistake became clear about eighteen months into our second investment. The company was technically meeting the impact KPIs we had agreed on post-close, but the metrics were measuring outputs (number of patient consultations conducted through the platform) rather than outcomes (improvement in clinical outcomes for those patients). The company was growing, the output metrics were improving, and the underlying health outcomes were ambiguous. We had inadvertently created a measurement framework that enabled impact washing.
We completely rebuilt our impact measurement process after this experience. Today, every Sway for Future term sheet includes a draft Theory of Change and proposed impact KPIs, developed collaboratively with the founding team during due diligence. The term sheet is not signed until both parties have reached genuine agreement — not just compliance — on the impact framework. The difference in quality and genuine accountability is night and day.
The practical implication: if a founding team is resistant to rigorous impact measurement during due diligence, that is a fundamental signal that the mission alignment is not deep enough to survive the commercial pressures of building a company. We have passed on investments for this reason.
The Best Impact Companies Solve Affordability Gaps, Not Just Access Gaps
Access is necessary but not sufficient. A healthcare telemedicine platform that connects rural patients with urban specialists creates access, but if the consultation costs $150 per session and the patient earns $800 per month, the access is theoretical. The most durable impact businesses solve the affordability problem simultaneously with the access problem — because affordability is what makes adoption self-sustaining rather than dependent on subsidies or NGO distribution.
EduPath taught us this lesson most clearly. The platform's personalized learning technology dramatically improved outcomes in pilot schools — the third-party evaluations were compelling. But the initial product required a tablet device for each student, which was not affordable for Title I schools operating on constrained budgets. The company was creating access to personalized learning quality without creating affordability of the device infrastructure required to deliver it.
The EduPath team solved this by redesigning the product to work on existing Chromebook infrastructure already present in 85% of Title I schools and to function effectively at 10% of the processing power required by the original design. This was a 9-month rebuild that delayed revenue growth but fundamentally changed the unit economics of school adoption. Today EduPath serves 120,000 students across 15 states, and the adoption rate in Title I schools is 3x what it was under the original product design. The affordability problem was the real product problem.
Community and Trusted Distribution Matter More Than Technology in Health, Education, and Agriculture
Impact technology companies operate in sectors where trust is the primary adoption barrier — not awareness, not price, and not feature set. A farmer in rural Ghana adopts a new agricultural technology because a neighbor they trust has already adopted it and can attest to the results. A parent enrolls their child in a new digital learning program because their child's teacher, whom they trust, has recommended it. A patient uses a telemedicine platform because their primary care physician, whom they trust, has referred them.
The companies in our portfolio that have grown most efficiently have all discovered, at some point, that their distribution model needed to be built around trusted intermediaries rather than direct consumer or farmer acquisition. AgriSense's most cost-effective channel is not paid digital advertising — it is their network of community agricultural extension officers who have been trained as AgriSense "farm ambassadors" and who earn referral revenue for every new farm they onboard. EduPath's growth accelerated 4x when they shifted from school administrator sales to teacher champion programs. HealthBridge's rural clinic penetration is almost entirely driven by referrals from the National Rural Health Association network.
The implication for how we think about go-to-market at seed: we now ask every impact founder to identify their three most important trust intermediaries and explain how they plan to partner with them. Founders who have not thought about this have a significant go-to-market gap in their plan.
When to Double Down and When to Accept a Hard Situation
NutriSync, our precision nutrition portfolio company, faced an existential challenge in mid-2023. The original product vision — a comprehensive metabolic health platform combining continuous glucose monitoring, gut microbiome analysis, and dietary recommendations — was technically sound but commercially too complex to sell in a direct-to-consumer channel. Customer acquisition costs were unsustainable, and the onboarding process required more behavioral change than most customers were willing to commit to without clinical guidance.
The founding team brought us a pivot proposal: abandon the direct-to-consumer model entirely and focus on enterprise health benefits, where employers would pay for the platform as a workplace wellness benefit with employer-sponsored onboarding support. The pivot required rebuilding the sales motion, changing the pricing model, and hiring three new enterprise sales representatives. It also required extending the runway by six months and raising a small bridge financing round.
We chose to double down. The underlying technology was validated — the 40% reduction in metabolic disease risk markers in clinical studies was real, reproducible, and compelling to any employer managing healthcare costs. The distribution problem was solvable. The team was exceptional. The pivot made clinical and commercial sense. Today, NutriSync is deployed at 150+ enterprise customers and is on a clear path to its Series A. The decision to double down rather than accept a difficult outcome was right — but it was not obvious at the time.
The framework we have developed for these decisions: we double down when the technology is validated, the founding team has demonstrated the capacity to adapt, and the revised commercial model has a credible path to unit economics. We accept hard situations when any one of those three conditions is missing. Doubling down on fundamentally broken technology or on a team that cannot adapt is rarely a good use of capital, regardless of how much we believe in the mission.
What We Would Do Differently
Three things we would change about our earliest investments:
We would move faster on impact measurement frameworks. The months we spent on post-close impact measurement negotiation were months we could have spent on customer development. If a founding team is not already thinking rigorously about impact measurement, the due diligence period is the right time to develop that rigor together — not after closing.
We would be more honest about go-to-market complexity earlier. In our first two investments, we gave the benefit of the doubt to founders' go-to-market plans because the technology was compelling and the problem was genuine. We should have pushed harder on distribution assumptions, key account relationships, and sales cycle length. A company with extraordinary technology and a flawed go-to-market plan will still struggle.
We would invest more in post-investment operational support. Our initial fund model assumed that a quarterly board meeting and on-call partner access would be sufficient post-investment support. In practice, the portfolio companies that have grown most successfully are the ones where we have been most actively involved in specific functional areas — introductions to customers and partners, hiring support, regulatory navigation. We have significantly increased our operational support capacity in response to this learning.
How Our Thesis Has Evolved
We started Sway for Future with a thesis that we summarized internally as "tech for good." After five years and twelve investments, we have evolved that thesis to something more precise: we back technology that is fundamentally better because of its mission.
The difference is significant. "Tech for good" could describe a competent conventional technology company that donates to environmental causes. Our evolved thesis describes companies where the mission creates structural competitive advantages that conventional competitors cannot replicate: AgriSense's deep domain expertise in soil science that makes its recommendations materially more accurate than any competitor; EduPath's trust relationships with Title I schools built through years of mission alignment that create adoption advantages that no well-funded EdTech competitor can buy; ClearFlow's community-based deployment model in emerging markets that generates trust-based distribution that no corporate competitor can substitute with marketing spend.
Impact, when it is genuinely embedded in the product and business model, is a moat. That is the insight we started with, and five years of portfolio experience has only deepened our conviction.