Most children in the UK are not eating enough fruit and vegetables to support optimum health. Evidence-based interventions are needed to change this trend. In the present pilot study, effectiveness of simple behavioural nudges on children’s lunchtime consumption of fruit and vegetables was tested in two primary (elementary) schools. Children’s (n = 107) lunchtime consumption was measured directly through the use of a validated digital photography protocol; measures were taken at baseline and again after a 3-week long intervention. Changes to the choice architecture of dining rooms included improved presentation and provision of target foods, attractive advertisements and labelling, and prompting by staff. For children who took school lunches (n = 67), both selection and consumption of fruit increased as the result of the intervention. Their selection of vegetables did not change over time, however, and their consumption ether remained unchanged or declined. No changes were observed in the comparison group (n = 40), who brought their lunch boxes from home. These results caution against using selection (serving sizes), as estimates of consumption or a measure of behaviour changes. They also show that all evaluations should consider the effects of each intervention on children’s eating overall, rather than just report changes in target item consumption, to check for any unintended consequences of the intervention.