Cross-sector delivery becomes fragile when partners are funded and performance-managed in silos. The result is predictable: activity looks healthy, but outcomes stall and no one can credibly explain why. This article sits within System Leadership & Cross-Sector Governance and should be governed through Board Governance & Accountability, because commissioners and boards need defensible answers to three questions: what is being achieved, who is accountable for what, and what changes when performance slips.
Why shared outcomes fail in practice
Most partnerships agree the ânorth starâ (reduced crises, better stability, improved access), but then operate with incompatible measurement, separate funding levers, and different definitions of success. A housing partner may measure tenancy sustainment, a health partner may measure appointment adherence, and a social care partner may measure service hours. None of these are wrongâbut without a shared outcome framework, they do not add up to system impact. The solution is not one giant spreadsheet; it is a small set of shared outcomes with clear attribution rules, practical data standards, and a routine governance cadence that forces decisions.
Two oversight expectations you should be able to evidence
Expectation 1: Performance measures are actionable and attributable. Funders and system leaders expect that each metric has an owner, a threshold, and a defined management action when it goes off track (not just reporting).
Expectation 2: Funding and contract levers align with the operating model. Boards and commissioners typically expect incentives do not reward âhand-offsâ and cost-shifting. If one partner improves their internal metrics by pushing risk elsewhere, governance should detect and correct it.
Design principles that hold up under scrutiny
Effective cross-sector contracting starts with a few practical principles: (1) separate shared outcomes from partner-specific deliverables, (2) define minimum data quality and timeliness standards for any metric used for payment or escalation, and (3) ensure performance review includes âwhyâ analysis (root causes, interface failures, and capacity constraints) rather than only âwhat happened.â In many systems, the first maturity step is to treat outcomes as a governance deviceâused to drive decisionsâbefore using them for payment.
Operational Example 1: Shared crisis reduction metric with a clear âinterface failureâ review
What happens in day-to-day delivery. The partnership defines a shared metric such as âavoidable crisis contactsâ (e.g., ED visits, crisis line calls, urgent placements) for an agreed cohort. A monthly review pack lists the top repeat utilizers and flags where there was a prior contact opportunity (missed follow-up, delayed handoff, unresolved safeguarding plan). A cross-partner panel reviews a sample of cases using a standard template: timeline, decisions made, handoffs attempted, and where the pathway broke. Each case produces specific actions with named owners (e.g., strengthen follow-up within 72 hours, tighten escalation triggers, improve medication reconciliation at discharge).
Why the practice exists (failure mode it addresses). The failure mode is âoutcomes theaterââeveryone agrees crises are bad, but no one can identify the operational causes in a way that changes practice. Another failure mode is blaming a single partner when the real issue is an interface failure across multiple handoffs.
What goes wrong if it is absent. Without a structured review, crisis metrics become a dashboard that generates concern but not learning. Partners can defend themselves with plausible explanations (âwe did our partâ), and commissioners cannot see whether the pathway is being improved. Over time, trust erodes and funding discussions become adversarial because nobody can show credible control.
What observable outcome it produces. A structured âinterface failureâ review produces measurable system improvement: fewer repeat crises for the same individuals, faster time-to-follow-up after high-risk events, and documented actions that governance can track to completion. It also produces defensibility: leaders can show they understand why crises occurred and what they changed as a result.
Operational Example 2: Attribution rules that prevent cost shifting between partners
What happens in day-to-day delivery. The partnership defines attribution rules for shared outcomes. For example, if a housing breakdown occurs within a defined window after an unmet clinical follow-up, the review attributes part of the failure to the follow-up gap rather than treating the housing provider as solely responsible. A quarterly âcost-shift checkâ compares changes in one partnerâs internal metrics with changes in system metrics (e.g., reduced program contacts but increased crisis use). If patterns suggest cost shifting, the tactical governance group triggers a joint improvement plan, and the contract includes a mechanism to rebalance responsibilities and funding for the next quarter.
Why the practice exists (failure mode it addresses). The failure mode is perverse incentives: one partner improves their performance by narrowing eligibility, discharging complex clients, or reducing contact frequencyâimproving internal activity measures while system harm increases.
What goes wrong if it is absent. Without attribution rules, partnerships become fragile. Frontline staff experience constant referral âping-pong.â Commissioners see inconsistent narratives, and the system ends up paying more through downstream crisis services. Accountability becomes politicized rather than evidence-based.
What observable outcome it produces. With attribution rules, governance can spot and correct misalignment early. Outcomes include fewer circular referrals, clearer escalation routes for âstuckâ cases, and improved stability indicators that track across partners rather than within one silo. Importantly, leaders can evidence that incentives support collaboration instead of creating avoidance behavior.
Operational Example 3: A performance cadence that produces decisions, not just reporting
What happens in day-to-day delivery. The partnership runs a three-level cadence: weekly operational huddles (capacity, urgent risks, immediate barriers), monthly tactical performance reviews (cohort outcomes, case sampling, improvement actions), and quarterly strategic governance (contract levers, funding shifts, redesign decisions). Each cadence uses a âdecision logâ format: what was decided, why, the evidence used, the owner, and the deadline. Data used for performance is subject to minimum timeliness and quality checks; metrics are not presented without clear denominators and definitions.
Why the practice exists (failure mode it addresses). The failure mode is drift: governance meetings become presentations, problems are noted repeatedly, and no one owns corrective action. Another failure mode is inconsistent data leading to debates about validity rather than decisions about change.
What goes wrong if it is absent. Without a cadence that forces decisions, issues lingerâmissed follow-ups, delays in referrals, unclear escalation authorityâand frontline teams create workarounds that are invisible to leadership. Commissioners lose confidence, because performance discussions never translate into action that can be tracked and audited.
What observable outcome it produces. A decision-driven cadence produces visible governance signals: higher completion rates for improvement actions, faster resolution of recurring barriers, and better trend performance over time. The audit trail also supports external scrutinyâleaders can show that when performance slipped, they acted, and they can evidence whether the actions worked.
Putting it together: align incentives while keeping accountability clear
The goal is not to eliminate partner accountability; it is to make shared outcomes meaningful. Keep shared outcomes few, define them tightly, establish attribution rules, and run an assurance cadence that creates decisions and learning. Done well, cross-sector contracts become a mechanism for operational improvement and public accountabilityârather than a source of blame and defensiveness.