While much research has focused on understanding the process by which young children learn to count, little work has explored the effects of direct instruction on this process. In the current study, we explored the impacts of training children in an explicit counting procedure on two distinct cardinality tasks. Two- to 5-year-old children first participated in a Give-N task in which counting proficiency was assessed, and then participated in a short instruction session where explicit counting was modeled and encouraged. Following training, children were significantly better at identifying which of two cards contained a set size outside of their range of counting mastery (Huang, Spelke, & Snedeker, 2010) and were more likely to improve on a secondary numerical production task (Give-N; Wynn, 1990, 1992) compared with children in the control group. Not surprisingly, a greater proportion of children in the count training condition overtly counted during the cardinality task, a strategy that was found to be the strongest indicator of performance. Together, results reveal that even 5 min of counting instruction greatly increases the likelihood that a child will engage in counting behavior and results in improvements in cardinality judgments in two distinct numerical tasks.