Farming projects must be able to access research at any point along the innovation trajectory, say Rasheed Sulaiman V. and colleagues.
But new evidence by the Research Into Use (RIU) programme in Asia, funded by the UK Department for International Development, suggests this is rarely the case. 
Instead, research works best when it can respond to needs for knowledge at any stage of the innovation process.
A learning curve
RIU's 'post-research' projects, mainly led by non-research organisations, aimed to broaden the use of products and processes derived from agricultural research.
The assumption was that new ideas and technologies from completed research simply required dissemination activities for them to be taken up by the farming community. But it quickly became apparent that the process of applying research ideas — innovation — often needs further research support.
Sometimes innovation means using research expertise for technical training. For example, in a nongovernmental organisation-led project in Bangladesh, which promoted a decentralised technique for fish seed production, the researchers who developed the technique organised training, managed troubleshooting and provided technical support for brood producers and seasonal pond farmers.
Sometimes existing technologies need further research to adapt them to new applications. For example, when community-based organisations involved in a project to manage a flood plain in Bangladesh started diversifying into enterprises — duck rearing and bee keeping — they found they needed research to help them adapt their production techniques to the flood plain environment.
And sometimes new research questions come to light as a result of bottlenecks during the innovation process. For example, in the same flood plain management project, farmers needed market research to learn how to sell sunflower and other crops that were new to the region.
Implications for investment
These cases suggest that there is an ongoing need for research at different points along the innovation trajectory: sometimes as a source of expertise, sometimes to adapt existing techniques, and sometimes to solve a new problem or to learn how to do something new.
The findings have important implications for the way agricultural research is put into use, but also for how projects are financed, managed and evaluated.
National and international development investors need to recognise that research and innovation are not separate, but inter-connected activities that feed into each other as an innovation trajectory progresses.
To put this into practice, researchers need new incentives to encourage them to support others involved in the innovation process. RIU projects led by researchers found that complementary expertise had to be brought in to bring about innovation and impact.
For example, plant breeders in India and Nepal trained community-based groups to help them develop new rice and chickpea varieties, but found they also needed entrepreneurial and marketing skills to promote these varieties. This expertise was brought in by involving specialist development agencies.
Irrespective of what sort of organisation leads a 'research into use' project, success depends on combining research with development expertise of different types.
One way of achieving this is to reward researchers for the outcomes of innovation processes they have supported, in addition to rewarding them for generating new ideas and publishing papers.
Accountable for development
To finance this way of working, countries like India and Papua New Guinea are experimenting with innovation challenge funds and competitive grants mechanisms to support innovation processes that create opportunities for small-scale farmers and agricultural entrepreneurs.
In Papua New Guinea, for example, a grants scheme for agricultural innovation, funded by the Australian government's overseas aid program AusAID, proved to be an effective mechanism for delivering research services. The government is scaling it up as a national competitive grants scheme to support innovation.
Projects funded in this way are often led by a development organisation with researchers as partners. They range from initiatives to introduce coffee production techniques into the school curriculum, to working with women's groups to improve the quality of their processed coconut products.
The starting point for such projects might be new findings from research, but equally it could be new market opportunities, new technology supply chains or social innovations in development practice.
Their defining feature should be that they are accountable for development outcomes, but include resources and partnerships to access research services as and when they are needed as the innovation trajectory evolves.
Rasheed Sulaiman V. is director at the Centre for Research on Innovation and Science Policy in Hyderabad, India. Andy Hall is director of LINK. T.S. Vamsidhar Reddy is visiting scientist at AVRDC (The World Vegetable Center), Hyderabad, India. Rasheed can be contacted at Rasheed.email@example.com.
 Vamsidhar Reddy, T.S. et al. The when and where of research in agricultural innovation trajectories: Evidence and implications from RIU's South Asia projects [704kB] (UNU-MERIT, 2011)