Charting the path towards a long-term knowledge brokerage function: an ecosystems view

The paper addresses how evidence use in policymaking, drawing on multiple sources, benefits from but also requires enhanced collaboration across domains. The authors focus on ecotones—transitional interfaces between the academic research and policy ecosystems—as hybrid networks that facilitate knowledge exchange. They pose the research question: How does the knowledge brokerage function of ecotones evolve to meet the needs of academic research and policy ecosystems? Adopting an ecosystem perspective and a phenomenological approach, the study seeks to describe the lifecycle evolution of ecotones’ brokerage functions and their implications for evidence generation and use.

Three units of analysis in brokerage literature are outlined: (1) individual brokers, emphasizing skills and roles at the science-policy interface; (2) boundary organizations, enabling collaboration among actors with divergent interests but facing politics, power tensions, and definitional struggles over evidence; and (3) ecosystems/ecotones, hybrid networks at interfaces of science with policy/business that promote reciprocal cross-pollination. While prior work examines broker archetypes, network governance, and static views of brokerage (e.g., ARCs), gaps remain on how ecotone brokerage functions evolve over time and adapt to political/institutional barriers. The paper argues that network brokers operating across ecotones can leverage diverse relationships to overcome resistance to external organizations, motivating an ecosystem lens on brokerage evolution.

Design: Qualitative phenomenological study (Moustakas, 1994; Creswell & Poth, 2016) focusing on participants’ lived experiences of knowledge brokerage within ecotones at the research–policy interface. Researchers adopted reflexivity and bracketing to minimize preconceptions. Data collection: Mixed sources over March 2022–March 2023.

• Survey (N=127): purposive sampling across policy and academic communities (London policymakers=67; academic researchers=60). Used to inform interview protocol.
• Semi-structured interviews (N=107): 30–90 minutes each, June–September 2022 (London policymakers=49; academic researchers=58). Transcribed via MS Teams and manually checked; pseudonymized and analyzed in NVivo. Sampling spanned disciplines, institutions/teams, and policy areas; data saturation guided stopping.
• Direct observations (~1950 hours): Authors served as Policy Fellows embedded in the Greater London Authority’s City Intelligence Unit and as part of the London Research and Policy Partnership (LRaPP) executive team during a 12-month pilot to develop a brokerage function. Observations focused on broader experiences of research/policy engagement.
• Archival data: Programme booklets, presentations, terms of reference, calls for proposals, policy engagement mapping reports, websites, and blogs to map actors and partnerships. Analysis: Inductive thematic analysis (Braun & Clarke, 2006) in NVivo. Open coding generated 613 nodes, aggregated to 74 second-order codes, then axial coding to aggregate dimensions informed by organization design and knowledge co-production/brokering literature. Three clusters structured the framework: (1) Collaborative Knowledge Exchange and Facilitation; (2) Resources, capabilities, and practices; (3) Brokerage focus, response, and knowledge utilization. Validation included 13 feedback sessions (workshops, boards, seminars, conferences) with actors from both ecosystems.

The study proposes a three-stage maturity transitions framework for ecotones’ knowledge brokerage functions, tracing evolution from basic services to complex, multi-network coordination. Stage 1: Service-function model

• Brokerage mode: Foundational, basic brokerage (mapping interests, raising awareness), light-touch convening, informal logging of requests and connections.
• Funding: Minimal/seed funding (e.g., small HEIF-backed projects), often ad-hoc.
• Resources and practices: Resource acquisition via informal networks; immature capability building without broad buy-in; engagement resembles consultancy and relies on existing relationships.
• Focus/response/knowledge use: Emphasis on urgent priorities and quick wins; responses are ad-hoc and reactive; knowledge base is fragmented with disparate contributions and episodic momentum. Stage 2: Programme-partnership model
• Brokerage mode: Programme management structure with aligned programmes, boards/steering groups, and delivery teams; proximity between researchers and policy-makers is valued.
• Funding: Government or ecosystem funding; multiple funders can create governance/priority asymmetries.
• Resources and practices: Significant investment in professional services for management, events, and communications; capability building through legitimacy of recognized programmes; systematic engagement (e.g., speed-dating, roundtables, matching services, joint problem statements).
• Focus/response/knowledge use: Strategic, narrowly defined foci (e.g., climate adaptation, AI/data science, ageing in cities); proactive, ongoing collaboration; accumulation of multi-disciplinary evidence toward tightly defined areas. Stage 3: Network of networks model
• Brokerage mode: Hub-and-spoke with an umbrella hub coordinating multiple self-regulated spokes; roles include collaborator, communicator, and campaigner (advocacy).
• Funding: Mixed revenue—initial government support, hub subscriptions, spokes’ independent funds.
• Resources and practices: Intensive mapping and sharing of resources across networks; capability building emphasizes replication and scaling across policy domains; digital brokerage tools (e.g., NCUB Konfer) and Areas of Research Interest (ARIs) enable sophisticated matching and agenda-setting.
• Focus/response/knowledge use: Strategic and evolving foci addressing emergent public-interest challenges; co-production and challenge-led approaches; diversified, multi-disciplinary evidence feeding multiple policy domains simultaneously (e.g., net zero, EDI, mental health). Outputs to research and evidence systems
• Stage 1 produces evidence pools (disparate, limited repositories for specific issues).
• Stage 2 develops evidence lakes (comprehensive, organized repositories for well-defined policy areas).
• Stage 3 enables cross-pollination of evidence across domains, integrating insights and applying lessons between fields. Validation: Temporal bracketing frames maturity progression; 13 expert review sessions supported construct validity. Comparisons with Best & Holmes (2010) suggest complementarity between systems thinking and ecotone approaches.

Findings directly address the research question by explicating how ecotone brokerage functions evolve across maturity stages with distinct configurations in convening power, funding structures, resources/capabilities, engagement practices, and knowledge utilization. The framework demonstrates that maturing brokerage shifts from reactive, fragmented knowledge exchanges to proactive, coordinated programmes and ultimately to orchestrated, multi-network collaboration supporting co-production and diversified evidence use. Theoretical contributions include: (1) extending knowledge brokering theory with an ecosystem/ecotone perspective that emphasizes gradual maturation and dynamic governance beyond individual brokers and boundary organizations; and (2) articulating how brokerage maturity transforms research evidence outputs—from isolated evidence pools to evidence lakes and cross-pollinated knowledge—addressing fragmentation in evidence use. The study situates its model alongside systems thinking (Best & Holmes, 2010), suggesting both are complementary: systems thinking illuminates complexity; ecotone approaches operationalize stakeholder engagement and evidence mobilization. The discussion also recognizes potential biases and constraints within ecosystem approaches (disciplinary fit to policy issues, institutional reputation effects, interdisciplinary power asymmetries, political context complexity) and underscores the need to manage increasing complexity as brokerage scales.

The paper contributes a three-stage transitions framework for ecotones’ knowledge brokerage functions: Stage 1 service-function, Stage 2 programme-partnership, and Stage 3 network of networks. This model clarifies how ecotones can systematically build from basic brokering to structured programmes and ultimately to orchestrating multiple networks to co-produce and mobilize evidence across domains. Practically, recognizing maturity stages can guide policymakers, funders, and research organizations in aligning investments, governance, and engagement strategies with ecotone development. Future research directions include: specifying milestones and timescales for transitions between stages and identifying enabling factors (e.g., leadership, project successes, committed individuals); examining contextual contingencies (political/physical geography, institutional proximity, ecotone thematic focus); and deepening analysis of individual brokers’ roles, leadership, and management of activities, meetings, and project calls within ecotones.

• Temporal and milestone specificity: The study identifies stages but does not determine precise timescales or milestones for transitions, nor isolate causal enablers (e.g., leadership, project successes).
• Context dependence: Transferability may vary with political and physical geography, institutional proximity, and thematic focus.
• Potential biases within ecosystem approaches: Some disciplines may have weaker policy linkages; institutional prestige may skew participation; interdisciplinary collaboration risks dominant disciplines overpowering others; political complexity may constrain brokerage effectiveness.
• Data sensitivity and scope: While extensive, observation and interview data focus on London-based contexts and are not publicly shareable; initial fellowship observations emphasized broader experiences rather than the final operational brokerage outcomes.