If a successful project leaves more ecology behind than it consumed, and if we can actually measure ecology, then we can do something new: predict which projects will increase ecological capacity, and choose those. That inverts project selection. Instead of only what return will this generate? we ask what capacity will exist afterward that did not exist before? — where “return” now includes what remains in people, institutions, networks, knowledge, infrastructure, trust, access, and future capability.
Measure a capacity stack, not one number
The first move is to stop measuring “ecology” as a single thing and measure a capacity stack — the dimensions along which a project can add: human skill; physical infrastructure; access to tools; reusable knowledge; relationships and networks; stronger institutions; agency; health and viability; resource access; legitimate participation; resilience; reduced single-point dependency; future optionality; new pathways for later participants; and improved ability to detect and correct error.
Then every proposed project gets two maps — what will it consume? and what will remain? — which suggests an equation, conceptually if not literally:
Net Ecological Gain = capacity after − capacity before − capacity irreversibly consumed
Two companies could build the same $1B solar plant producing identical electricity. Project A imports everything, uses temporary labor, patents the operating knowledge, and leaves a working plant. Project B trains local maintenance engineers, connects local suppliers, leaves open documentation and research partnerships, and gives the community access to the new infrastructure. Financially near-identical; ecologically worlds apart. Project B leaves more field behind.
Ten layers of measurement
- Direct capacity created — the obvious new capability (a railway = transport capacity; a hospital = treatment capacity).
- Participant development — who becomes more capable because they took part? Not “did they have a job for three years” but “what can they carry afterward that they couldn’t before” — a participant capacity delta.
- Capacity propagation — do the gains stay confined, or multiply? A technique taught to ten others; a method open-sourced; a supplier that can now serve other industries. Some projects create a capability; others create capabilities that create capabilities.
- Connection effects — how much previously unreachable capacity becomes reachable? A transit line linking neighborhoods to jobs; broadband connecting rural participants; a platform giving a one-person firm corporation-scale tools.
- Option creation — how many new viable paths become possible afterward? A bridge opens geographies; broadband opens new forms of work and education. Option value creation — some projects open enormous option space; others lock society into one expensive path.
- Dependency created (the negative side) — a project can raise immediate capacity while lowering long-term capacity: proprietary tech nobody locally understands, a single foreign supplier, software locked to one vendor, AI that deskills the humans around it. So: capacity gained minus dependency created.
- Knowledge retention — does the knowledge become documentation, methods, standards, teaching material, reusable software — or leave in twenty experts’ heads when they go?
- Resilience gain — does it give the ecology more ways to keep functioning when something fails? Distributed generation, redundant routes, local food, broader competence. Efficiency hates redundancy; ecology sometimes needs it — value the adaptive reserve.
- Correction capacity — one of the most systems-oriented measures: does it make the system better at noticing and fixing mistakes (transparent data, monitoring, independent evaluation, feedback channels, appeal)? It leaves behind a better learning system, not just a thing.
- Geometry activation — the distinctly No Wasted Geometry part: how much previously stranded human capacity becomes available because of this project? Childcare that lets thousands of parents participate; disability access that unlocks blocked participation; apprenticeships that surface capacity invisible to credential-based hiring. Not merely “social benefit” — activation.
The project scorecard
Crude at first, and that’s fine — GDP aggregates wildly unlike things too. The conceptual move is what matters: projects have ecological balance sheets.
| Dimension | Project A | Project B |
|---|---|---|
| Physical capacity created | +80 | +80 |
| Participant development | +10 | +65 |
| Knowledge retained | +5 | +70 |
| Network connectivity | +20 | +75 |
| Future options created | +25 | +80 |
| Resilience | +15 | +60 |
| Stranded geometry activated | +8 | +55 |
| Dependency created | −60 | −15 |
| Ecological gain | 103 | 470 |
Every major project would carry an ecological balance sheet: assets created (skills, infrastructure, knowledge, networks, optionality, institutional capability) against liabilities created (resource depletion, dependency, concentrated power, environmental damage, deskilling, exclusion, fragility, lost alternatives) — netting to an ecological position after completion.
From procurement to capacity return
Now imagine public procurement. Today a government chooses among bids largely on price, technical compliance, schedule, and financial risk. A No Wasted Geometry government adds one question: which proposal leaves the ecology most capable after the asset is delivered? Suddenly the “cheapest” bid may not be cheapest — one bidder saves $100M but leaves no local knowledge and a proprietary dependency; another costs $120M more but leaves thousands of trained people, local supplier capacity, transferable knowledge, and a base capable of generating future projects. The second’s ecological return can dwarf the extra cost.
That points to a metric — capacity return on investment (CROI, to avoid the existing “EROI”): for every dollar, year, unit of energy, acre, and unit of human attention consumed, how much lasting ecological capability was produced? Score that way and early-childhood development, public transit, libraries, and parks may prove extraordinary — while a glamorous megaproject that leaves almost nothing transferable scores poorly.
Predictive modeling, and four levels
Before approving a project you could model it: if we place this here, which capacities propagate through the field over 5, 20, and 50 years? — an input-output model that tracks capacities instead of dollars. A rail line yields transport → labor mobility → educational access → new businesses → new development corridors; a university yields knowledge → trained participants → research → startups → civic leadership → future teachers. With feedback loops — greater capacity creates participants who create more capacity — some projects show ecological compounding.
That is the prize, and it sorts all project effects into four levels:
- Output — what did we make?
- Residual capacity — what capability remains after we made it?
- Propagating capacity — what can that residual capability now create elsewhere?
- Regenerative capacity — does the ecology become better at generating its own future capacity?
The fourth is extraordinary: a regenerative project doesn’t merely solve a problem, it increases the problem-solving capacity of the field. It even inverts a common business logic, where customer dependency is an asset — here, independence created is itself a return. Which sharpens the line you liked into its strongest form:
A successful project leaves more ecology behind than it consumed. A truly exceptional one leaves an ecology more capable of growing itself.
That remains the right test for a project. But scaled up to a whole society, the principle grows a second half, because a field also has to survive the states in which its participants cannot contribute at all. The fuller version: a successful ecology does not merely leave more capacity behind than it consumed — it preserves the conditions by which capacity, relation, and standing can continue across generations and through periods of dependency. That second pillar is ecological continuity, the co-equal partner to everything measured here.