Table 1.

Research and Development recommendations for the six groups identified by cluster analysis.

Cluster	Research and Development recommendation
Cluster 1: High Yield and High N application rate (HYHN)	Little room for yield improvement. Focus on increasing nutrient use efficiency.
Cluster 2: Low Yield and Low N application rate (LYLN)	Low socio-economic relevance i.e., Research and Development actions have a limited impact on farmers' livelihoods and the economic development of the community. Unraveling the causes of low yields could provide the way forward.
Cluster 3: Low Yield and Medium N application rate (LYMN)	High socio-economic relevance i.e., Research and Development actions could have substantial implications for improving farmers' well-being, addressing socio-economic inequalities, and contributing to economic growth in the area. Research institutes should investigate the yield gap, causes of low yield, and possible solutions. Site-specific nutrient management (SSNM) solutions should be tested. Then, governments/agencies should disseminate and promote good agricultural practices and SSNM solutions.
Cluster 4: Medium Yield and High N application rate (MYHN)	High socio-economic relevance. Research institutes should investigate the yield gap, causes of medium yield, and possible solutions. Site-specific nutrient management (SSNM) solutions should be tested. Then, governments/agencies should disseminate and promote good agricultural practices and SSNM solutions.
Cluster 5: Medium Yield and Low N application rate (MYLN)	Potential for further increasing fertilizer inputs; Research institutes should investigate reasons for low fertilizer inputs. Then, if relevant and according to the country's priorities, governments could establish policies facilitating access to fertilizer for smallholder farmers.
Cluster 6: Very High Yield and Very High N application rate (VHYVHN)	Little room for yield improvement. Focus on increasing input use efficiency.