--- name: program-governance-ecological-design description: "Invoke when a program shows strong project-level reporting but weak benefits realization, or when designing governance for a new program that must go beyond oversight and accountability to govern the interaction layer between projects where outcomes live." --- # Program Governance as Ecological Design ## R — Reading > "Governance is an ecological design problem. You are creating the conditions under which the right kind of emergence becomes possible... Not oversight of the components, but stewardship of the conditions." *Source: "Emergence Is Not a Metaphor" (20260418_emergence-is-not-a-metaphor); supporting references: Healthcare.gov repair case (c04), EVM measurement theater pattern (c15); theoretical grounding in Hoel (2026) causal emergence* ## I — Interpretation The conventional model of program governance is built around accountability and oversight: monitor projects, review status reports, track milestones, flag variances. This model produces something that looks like governance — regular reporting, structured escalation, visible controls — while missing where program outcomes actually live. The structural signature is recognizable: project reporting is rigorous and on time, but benefit realization is consistently weak. The governance function is performing correctly, and the program is still failing. The reason is causal location. Program-level outcomes are not produced at the project level — they are produced at the interaction layer between projects. The integration points, dependency relationships, and emergent properties that arise when multiple projects run in parallel are where the program actually happens. Managing components well, even managing them excellently, does not produce outcomes if nobody is governing the interaction layer. Causal emergence science (Hoel 2026) formalizes this: macroscales add genuine causal information not reducible to the properties of components. In practical terms: you cannot measure your way to program outcomes by aggregating project metrics. The ecological design reframe changes the governance question from "are projects performing to plan?" to "have we created the conditions under which the right outcomes can emerge?" This requires attending to four things the oversight model structurally ignores: whether weak signals from the interaction layer can reach someone with authority to act; whether projects have slack to adapt based on what they're learning; whether anyone explicitly owns the interaction layer; and whether governance is maintaining productive boundary conditions rather than eliminating the instability that generates emergence. The disequilibrium insight is counterintuitive but important: programs born from genuine urgency are born into instability. Early governance that tries to eliminate that instability — locking down plans, enforcing original scopes, demanding predictability — may prevent the emergence the program was created to produce. The governance function is not to eliminate productive instability but to be the stabilizing feedback that keeps it from becoming destructive collapse. ## A1 — Past Application The Healthcare.gov repair (c04) is the clearest empirical case. The initial failure was not a failure of project execution — individual workstreams were delivering against their scopes. The failure was at the interaction layer: nobody held a current model of how the components fit together, and integration failures compounded. The repair did not improve project management. It appointed a systems integrator with explicit authority to own the interaction layer — to hold and continuously update a model of the whole, identify integration failures early, and intervene across workstream boundaries. This is interaction layer ownership given governance instruments. The program recovered. The EVM measurement theater pattern (c15) illustrates the opposite: strong component reporting with no corrective authority at the interaction layer. Earned Value Management produces rigorous data about component performance. But when the integration layer is failing, EVM data shows each project on track while the program diverges from its intended outcomes. Governance can detect variance but cannot close the loop because the governance structure is organized around components, not the interaction layer. This is not a measurement problem — it is a governance architecture problem. ## A2 — Future Trigger ★ - A program director is preparing a quarterly review and realizes that all twelve projects are green or amber, but the program's stated benefits are nowhere close to materializing, and nobody can explain the gap. - A PMO is designing governance for a new cross-functional transformation program — multiple departments, multiple parallel workstreams, outcomes that depend on those workstreams coordinating — and the default template is a steering committee plus project status reporting. - A program is six months in and integration issues keep surfacing late, requiring expensive rework. Each workstream PM says they're on track. The issue is never visible until it becomes a crisis. - An executive asks why a program with excellent project management discipline and a capable PMO keeps failing to deliver its intended organizational outcomes, and the PMO's answer is to improve reporting cadence and tighten milestone reviews. - A program governance review reveals that integration risks, cross-stream dependencies, and emergent issues are being raised in workstream meetings but not reaching the steering committee until they've already become decisions rather than options. ## E — Execution 1. **Diagnose whether this is a program or a project.** The ecological design framework applies to programs — multi-project portfolios where outcomes are produced through interaction between projects, not by any single project. If the "program" is actually a large project with sub-tasks, the governance question is different. Ask: Do the intended outcomes require coordination and interaction between multiple workstreams in ways that weren't fully anticipatable at the start? 2. **Apply the interaction layer audit.** Ask three questions: - Who explicitly owns the interaction layer? (Name a person, define their authority, identify what governance instruments they have.) If the answer is "the PMO" or "the steering committee" without a specific accountable individual with cross-workstream authority, this is the primary gap. - Are weak signals from the interaction layer reaching decision-makers with authority to act — before they become crises? (Not aggregated data, but early-warning signals from integration points.) - Do projects have slack to adapt based on what they're learning, or does governance force adherence to the original plan regardless of what's emerging? 3. **Identify the unowned interaction layer gaps.** Map the integration points between projects: handoffs, shared dependencies, outcomes that require two or more workstreams to coordinate. For each integration point, ask: Who is monitoring this? Who has authority to intervene if it degrades? If the answer for any significant integration point is "nobody explicitly," that is a governance gap that cannot be closed by better project reporting. 4. **Design the interaction layer governance function.** Create an explicit role or function responsible for the health of the interaction layer. This person or function needs: visibility into integration points (not just project status), authority to convene cross-workstream conversations and escalate without going through each project's chain, and a regular cadence for updating their model of how the components are interacting (not just whether each component is on track). 5. **Establish the boundary conditions.** Define what productive instability looks like versus destructive collapse — what level of emergent variation is acceptable and generative, and what triggers a governance intervention. The goal is not to eliminate variance but to maintain the conditions under which emergence can produce the intended outcomes. 6. **Build weak signal pathways.** Create mechanisms for early-warning signals from the interaction layer to surface. This is not more reporting — it is specifically structured to capture integration-point health, dependency status, and early friction between workstreams. The signal needs a pathway to someone with authority to act, not just a place in a report. 7. **Review governance design, not just project status.** On a regular cadence (quarterly at minimum), ask whether the governance structure is still appropriate for where the program is. Early-stage programs may need more slack for recombination; mature programs approaching integration may need tighter boundary conditions. The ecological design requires maintenance, not just initial setup. ## B — Boundary This skill applies to programs — multi-project portfolios producing outcomes through interaction — not to individual projects. A large, complex single project with strong dependencies may benefit from some of these concepts, but the theoretical grounding (causal emergence at the interaction layer) requires genuine interdependence and emergence dynamics, not just complexity. The skill diagnoses governance architecture gaps and provides design principles. It does not prescribe specific governance instruments. What interaction layer ownership looks like depends enormously on organizational culture, program scale, domain, and available authority structures. The framework tells you what function needs to exist; it does not tell you exactly how to build it for your organization. The ecological framing can be misread as permission for unstructured programs. It is not. The framework explicitly requires stabilizing feedback and boundary conditions — the argument is against eliminating productive instability, not against governance discipline. Programs still need clear accountability, milestone structures, and escalation paths. What changes is the *target* of governance attention, not the rigor of governance. The theoretical grounding in Hoel (2026) causal emergence science is newer and carries the usual caveats of emerging academic consensus. The practical implications are well-supported by case evidence (Healthcare.gov, EVM patterns), but the specific theoretical framing should be treated as a useful lens, not an established axiom. Ecological design requires someone with sufficient organizational standing and authority to own the interaction layer across workstream boundaries. In organizations where authority is tightly siloed by project or department, creating this function may require political groundwork before governance redesign is possible. ## Related Skills - composes-with: `emergence-conditions-audit` — audit first to identify which conditions are missing; ecological design framework then prescribes the governance architecture - **governance-model-adequacy-test** — *relates*: ecological design requires interaction layer ownership; Conant-Ashby specifies who must hold the model to make that ownership real - informs: `sv-trend-diagnostic` — SV measurement theater is a symptom of ecological design failure; redesigning governance structure addresses the cause ______________________________________________________________________ ## Provenance - **Source:** Project Management Research and the Critical Path, Nicole Williams, 2026