BackgroundBehavioral change communication (BCC) promotes skills and knowledge to improve infant and young child feeding, but without additional material inputs, recipients must develop strategies to translate knowledge into action. Using data from the Alive & Thrive initiative in Bangladesh (2010–2014), we aimed to test whether households receiving the intensive intervention (opposed to the nonintensive intervention) increased expenditures on key foods for mothers and children (e.g., foods that were promoted by the intervention and also changed in maternal and child diets).MethodsThe intensive intervention provided interpersonal counseling, community mobilization, and mass media campaigns to promote breastfeeding and complementary feeding. A cluster-randomized design compared 20 subdistricts randomly assigned to the intensive (4281 households) or nonintensive (4284 households) intervention. Measures included food and nonfood expenditures, dietary diversity, and women's economic resources. Linear and logistic regression tested difference-in-differences (DD) in expenditures and dietary diversity, accounting for subdistricts as clusters, and the association between maternal and child consumption of specific food groups and corresponding food expenditures.ResultsExpenditures on eggs and flesh foods increased more in intensive areas than in nonintensive areas by 53 (P < 0.01) and 471 (P < 0.01) taka/mo, respectively. Household food expenditures increased more in intensive areas by 832 taka (P = 0.02), whereas changes in nonfood expenditures did not differ. Women's employment and control of income increased more in intensive areas by 12 (P = 0.03) and 13 (P < 0.01) percentage points, respectively, while jewelry ownership decreased more by 23 percentage points (P < 0.01). Higher expenditures on food groups were reflected in higher consumption by women and children.ConclusionsRecipients in the intensive intervention mobilized additional resources to improve diets, reflected in increased expenditures and consumption of promoted foods. BCC interventions should document how recipients produce desired results without additional material inputs, particularly for behaviors that likely require additional resources. This trial was registered at clinicaltrials.gov as NCT01678716.