After more than a decade of experience implementing the market systems development approach, key lessons have emerged globally and from within DFAT's portfolio, including from CAVAC, MDF, and most recently, AIP-Rural. The AIP-Rural Mid-Term Review outlines a range of lessons learned from implementation of AIP-Rural that have informed the design of AIP-Rural II. Key lessons from DFAT's portfolio of market systems development programs include:

• Market systems development works, including in fragmented and isolated contexts such as eastern Indonesia or the Pacific, and can benefit poor people and disadvantaged groups, such as women. To do so requires rigorous analysis and measurement, identification of viable business cases that can attract investment, and willingness to use a mix of intervention types and tactics. It has also been recognised that operating in these challenging contexts can adversely affect value for money, although the benefit-cost ratio of intervention typically remains positive.
  
  
• Key factors that determine successful intervention include: analysis that identifies the root causes of market constraints and viable opportunities for change; robust scrutiny of potential partners' incentives and capacity to change; negotiation of partnership deals that involve substantial co-investment and test and build partners' ownership over time; early identification of 'scale agents' that can contribute significant impact and act as role models for peers to copy; and intervention teams of sufficient calibre and skills mix do the analysis, deal making, due diligence and intervention management.
  
  
• Talent is the vital ingredient: market systems development is a distinctive capacity requirement. Sectoral specialisation has been found to be less critical than staff with attributes suited to market systems development: analytical, entrepreneurial, creativity, negotiation and change management skills. Staff are the most important way in which a program adds value; it is their intelligence, insights, ideas and influence that stimulate change. Programs have had to invest significantly in developing effective teams early in the life of a program, and then strive to retain talented staff. Since staff usually constitute the largest element of intervention costs, management systems must be capable of tracking staff time utilisation and allocating costs to interventions costs accurately, and not just record staff costs as part of overhead.
  
  
• Time and patience is required: market systems development programs tend to be constrained by time rather than funds. Building effective understanding, teams and partnerships takes patience. Programs can only move as fast as co-investors are willing to move, if their ownership is to be strengthened rather than diluted. Piloting innovations, measuring them, adapting them and then scaling them up can take several years, especially in seasonal agriculture. Early identification of scale agents can speed up this process, but all market systems development programs have followed an 'S-curve' trajectory towards sustainability and scale – starting small and slowly, then escalating significantly – that is markedly different from the linear pathway of direct delivery [REF]. This trajectory should be understood from the outset. Programs need to measure and report 'leading' indicators that can capture short term progress (e.g. partner co-investment and target group rates of access) and not rely solely on reporting 'lagging' indicators (such as change in income) that take a long time to materialise.
  
  
• Use rigorous portfolio management: the pathway to large-scale change is not predictable. Programs cannot control all factors affecting interventions, partners and sector conditions. Innovations, by definition, are new; therefore some will fail. A rigorous portfolio management approach has been found to be the best way to mitigate this uncertainty and ensure that aggregate performance delivers impact, by reducing dependency on a single sector, partner or intervention, i.e. by spreading the 'bets'. Successful programs have tended to focus on a blend of sectors, with different levels of risk and probabilities of impact; they have worked with a diversity of partners; and they have intervened through a range of entry points. The portfolio must be managed rigorously through routine and transparent mechanisms (e.g. monthly or quarterly team meetings). Strategies need to be re-examined regularly and the performance of interventions and partnerships monitored carefully. Underperforming or failing initiatives should be adapted, de-prioritised or dropped. Resources (people and finance) need to re-allocated to those initiatives that are demonstrating the best potential for progress towards impact.
  
  
• Retain as much flexibility as possible: market systems development programs work in dynamic environments, through independent market partners and do not deliver prescribed solutions. Innovation requires flexible and pragmatic contractual arrangements. Rigid, overly-detailed planning, budgeting and reporting frameworks have been found to restrict successful performance; they inhibit programs from responding to emerging opportunities and reacting to changing conditions. Programs require adaptive management systems, which provide accountability to funders but also cope with change. Typically, funders set the broad strategic parameters of program: a well-understood theory of change, high-level objectives and principles and associated targets (e.g. for goal and outcome level impact and outreach, sustainability, co-investment). Programs are then afforded considerable operational flexibility. This flexibility is managed using rigorous results measurement, incorporating short-term intervention planning and feedback loops, and regular oversight processes involving the funder, usually on a quarterly and annual basis, to set work plans and budgets and assess progress to date.
  
  
• Understand stakeholder requirements and tailor communications: market systems development programs are unconventional for many stakeholders (e.g. funders, government, politicians, media). The theory of change is complicated, the approach is sophisticated and the impact trajectory takes longer than direct delivery programs. Smart stakeholder communications is therefore a key role of program management - understanding stakeholders' interests, distilling key program information so that it is accessible and responds to these interests, and managing expectations. It requires a purposive communications strategy, tailored to different audiences, that goes beyond routine reporting.
  
  
• Influence public policy opportunistically: most market development system programs have had limited success in achieving specific policy or regulatory goals established at the outset of the program, even when they have had dedicated policy components. The political economy is too unpredictable and efforts have often been disconnected from mainstream interventions in sectors. More success has been achieved by being opportunistic, responding to openings as they emerge, usually as a result of sector-focused activities. It has been found that the most effective way market systems development programmes influence policy is through the evidence they generate through their analysis, the demonstration of alternative models as a result of successful interventions ('what works'), and through relationships with key market partners, who in turn have influential connections to regulators and policy makers (i.e. 'influencing the influencers').
  
  
• Supportive leadership and program culture: the lessons documented above demand high-calibre program leadership and a conducive program culture. Successful team leaders tend to be those experienced in change management processes, not just the technical delivery of prescribed tasks. They also tend to be skilled in human resource management, capable of building and leading a team of diverse talents. They need to be experienced in establishing and overseeing adaptive management processes that balance accountability to funders with the flexibility necessary to operate in dynamic market systems. They also need to be capable of shaping a program culture that is sufficiently open to encourage creativity and entrepreneurism, sufficiently rigorous to scrutinise information, partners and interventions, and sufficiently supportive to foster learning and development.