The recent demand for schools to respond to external accountability measures challenges school leaders to create school instructional systems that use data to guide the practices of teaching and learning. This paper considers how local school leaders build data-driven instructional systems (DDIS) by developing new programs and using existing school functions to create an information flow through a school. The paper considers how leaders work with teachers and students to create DDIS to intentionally and systematically improve student learning. The paper begins by presenting a theoretical and analytical framework for understanding the systems that local school leaders create to develop a DDIS to facilitate an information flow about student achievement in schools. The DDIS is presented as six component functions involving (1) data acquisition, (2) data reflection, (3) program alignment and integration, (4) program design, (5) formative feedback, and (6) test preparation. The second part of the paper reviews data collected in a year-long study of four schools, using the DDIS cycle as a framework for describing how school leaders structure opportunities to engage in data-driven decision making. The data provide rich examples of how schools facilitate and impede the flow of student achievement information, and call into question any simple definition of data-based decision making in schools. The following are appended: (1) DDIS Interview Protocol; and (2) DDIS Coding Scheme. (Contains 1 table, 1 figure, and 2 footnotes.) [This paper was prepared for the Annual meeting of the National Council of Professors of Educational Administration (Washington, DC, July 2005).]