In 2014-2015 and 2015-2016, a metropolitan school system in the southern United States embarked on a unique mission to improve the quality of its public prekindergarten programs through a partnership with a group of developmental researchers in an iterative, data-based venture. Data on 407 children in Year 1 and 433 in Year 2 (who were enrolled in 26 classrooms and extensively observed) are presented from the first 2 years of the ongoing partnership. All children were 4 years of age. Variability in classroom practices, measured empirically, and variability in child outcomes provided the means to examine the relations between children's gains in academic and social-emotional areas and major areas of classroom practices. Lessons learned, the eight identified significant practices, implications, and next steps in the partnership are addressed. (author abstract)