Although Bangladesh made some remarkable achievements in reducing poverty and in improving social and economic outcomes in recent decades, about one-third of the rural population still lives below the upper poverty line most of whom depend on agriculture as their primary source of income. One of the reasons for their poverty is the low productivity that results from sub-optimal use of inputs and other technology. To foster agricultural productivity and rural growth, technology innovations have to reach all strata of the poor among small farming communities in rural Bangladesh. For that purpose, technology opportunities need to be brought together with systematic and location-specific actions related to technology needs, agricultural systems, ecological resources and poverty characteristics to overcome the barriers that economic, social, ecological and cultural conditions can create. The first step towards this is to identify underperforming areas, i.e. rural areas in which the prevalence of poverty and other dimensions of marginality are high and agricultural potential is also high since in such areas yield gaps (potential minus actual yields) are high and productivity gains (of main staple crops) are likely to be achieved. The marginality mapping presented in this paper has attempted to identify areas with high prevalence of societal and spatial marginality - based on proxies for marginality dimensions representing different spheres of life - and high (un/der utilized) agricultural (cereal) potentials. The overlap between the marginality hotspots and the high (un/der utilized) agricultural potentials shows that Rajibpur (Kurigram), Dowarabazar (Sunamgonj), Porsha (Naogaon), Damurhuda (Chuadanga), Hizla (Barisal), Mehendigonj (Barisal), Bauphal (Patuakhali) and Bhandaria (Pirojpur) are the marginal areas where most productivity gains could be achieved.