You’re optimizing a model whose assumptions no longer hold, and every improvement is making you more efficiently wrong. Most advisory engagements default to optimization: improve what exists. Tighten the process. Hire better people. Fix the metrics. And in most situations, that’s the right call — the model works, it just needs refinement. But some problems aren’t optimization problems. They’re transformation problems. The org structure, the go-to-market, the leadership model, the product architecture — the foundational assumptions have shifted and the whole thing needs to change. Optimizing a broken model produces a more efficient version of the wrong thing. I see this constantly in climate tech, where the physics of the situation have changed but the organizational model is still calibrated for conditions that no longer exist.
What it looks like
Execution is improving by every internal metric. The team is more productive. Processes are faster. Costs are lower. And yet outcomes aren’t improving — or they’re getting worse. Market share is flat. Customer retention is dropping. The competition is pulling ahead with a fundamentally different approach while you’re perfecting yours. Internally, there’s a sense that the team is working harder than ever for diminishing returns. Quarterly reviews show progress on every operational KPI, but the strategic trajectory is wrong and nobody can explain why. The usual diagnosis is “execution gap” or “market headwinds.” The actual diagnosis: you’re climbing the wrong hill. Every step upward takes you further from where you need to be, and the operational improvements are masking the strategic error by making each step feel productive. The team is demoralized not because they’re failing, but because they sense the futility that the metrics can’t capture.
The mechanism
Optimization and transformation are different operations that require different diagnostics. Optimization assumes the model is correct and seeks to improve performance within it. Refine the sales process. Improve hiring speed. Reduce costs. These are valid when the underlying model — the business model, the organizational model, the product model — still fits the environment it operates in. Transformation is required when the model’s assumptions no longer match reality. When the market has shifted structurally, when the regulatory landscape has changed fundamentally, when the technology paradigm has moved — no amount of optimization addresses the mismatch. It’s the difference between tuning an engine and realizing you need a different vehicle. From my years in atmospheric physics: when your model assumptions break, incremental refinement doesn’t make the model more accurate. It makes the model a more precise representation of something that doesn’t exist. The same principle applies to organizations.
Why it persists
Optimization is comfortable. It has clear metrics, manageable scope, and visible progress. Transformation is terrifying. It means admitting that the thing you’ve been building — and improving, and investing in — is fundamentally misaligned with reality. Nobody wants to hear that, least of all the people who designed the current model. So organizations optimize by default. Every planning cycle starts from “how do we improve what we have” rather than “should we have this at all.” And the metrics reinforce the bias: optimization shows improvement immediately. Transformation shows disruption immediately and results later. The board sees the operational improvements from optimization and approves more optimization. They see the short-term chaos of transformation and delay it. This creates a ratchet effect where the organization becomes progressively better at something that’s progressively less relevant. By the time the strategic error is undeniable, the optimization investments have created deep structural lock-in that makes transformation even more painful. People, processes, incentives, customer commitments — everything is now calibrated to the old model. The cost of transformation grows with every quarter of optimization, which makes the decision to transform harder every quarter you delay it.
What changes
The first move is diagnostic, not prescriptive: which situation are you actually in? The signal that you need transformation rather than optimization is specific. If performance improvements aren’t translating to strategic progress — you’re faster, cheaper, better, and still losing ground — the model has likely broken. If the competitive landscape has shifted structurally — not a new competitor doing the same thing better, but different players solving the problem in a fundamentally different way — the model is obsolete. If the underlying assumptions about your customer, your market, or your environment have changed — and in climate tech, the physical environment literally changes — optimization is maintenance on a building with a cracked foundation. Distinguishing which situation you’re in requires stepping outside the model’s own metrics — because by definition, the model’s metrics will show improvement even when the model is wrong. You need external reference points: market structure, competitive positioning, customer evolution, regulatory trajectory.
Where this shows up
The optimization-vs-transformation question is acute in climate tech because the sectors themselves are transforming. Earth observation companies trying to optimize their hardware operations when they need to transform into analytics businesses. Clean energy companies optimizing project delivery when they need to transform into platforms. Climate adaptation companies optimizing their enterprise sales process when they need to transform their entire go-to-market for a market that doesn’t buy like enterprise SaaS. Built environment companies optimizing installation efficiency when they need to transform from a service business into a technology company. For investors, the distinction matters because the operating model determines whether optimization or transformation is the right move — and recommending the wrong one accelerates failure. Post-investment support that can’t distinguish between the two will make things worse.
When the model assumptions break, incremental refinement produces a more precise representation of something that doesn’t exist. The question is whether you’re tuning the engine or need a different vehicle entirely. Reach out.