For most of the history of school leadership, data arrived slowly enough that a decision cycle could keep pace with it. A test given in spring, scored over summer, reported to the board in the fall, acted on the following year. That rhythm assumed data was scarce, so nobody built a structure for moving fast on it. Neither condition holds anymore. Real-time dashboards, predictive risk models, and continuous formative assessment now generate more signal than most leadership teams can absorb on the old annual cycle (Mandinach, 2012).

I spent the better part of my career trying to close the gap between what the research says and what districts actually do. That gap was rarely about knowledge. It was almost always about will.

The Data Isn't the Problem

Ellen Mandinach made this point plainly back in 2012: policy and technology had outpaced the profession's capacity to use data well, and the gap has only widened since (Mandinach, 2012). Districts don't lack information. Most sit on more of it than any single leader can read closely: attendance systems, benchmark assessments, behavior logs, early warning dashboards, all updating in something close to real time. The bottleneck was never collection. It's judgment, and the structure that makes judgment possible at the speed the data now moves.

Amanda Datnow and Vicki Park drew a distinction in their research that matters more now than when they wrote it. Data-driven leadership treats the number as the answer. Data-informed leadership treats the number as one input a leader still has to weigh against context, experience, and judgment (Datnow & Park, 2014). The difference sounds small. In practice, it's the difference between a district that reacts to every dashboard alert and a district that has built the capacity to ask what the alert actually means before it acts.

That capacity doesn't show up automatically just because the data does. Mandinach and Gummer's research on data literacy found a persistent gap between the volume of data flowing into schools and the skill level required to read it responsibly, a gap that shows up at every level of the system, not just the classroom (Mandinach & Gummer, 2016).

I believe data-driven decision-making oversells itself, because predictive tools are not as clean as the interface makes them look. Alex Bowers and his colleagues reviewed 110 different dropout-risk indicators across 36 studies and found something worth sitting with: most of these flags are high-precision but low-accuracy. They catch a lot of students, but a meaningful share of the students flagged were never actually going to drop out (Bowers, Sprott, & Taff, 2013). A predictive model that's wrong in that direction doesn't just waste a counselor's time, but can put a student into an intervention they didn't need, based on a system that looked more certain than it was.

That's not an argument against predictive analytics. It's an argument for treating the output the way Datnow and Park describe: informed, not driven. The model does not replace the looking, but it does narrow where a leader should look.

If your district has invested in dashboards, early warning systems, or predictive tools over the past few years, the question worth asking this week isn't whether the data is good enough. It's whether your team has the structure, the protocols, and the standing time, to sit with what the data is actually saying before acting on it. Most systems built the pipeline first and the judgment layer second, if at all. But the ones that will hold up over time are building both at once.

A few practices separate the districts that turn data into judgment from the ones still drowning in alerts. Build a standing data conversation protocol, a structured set of questions a team works through before discussing any data set, so the conversation moves past the raw number into causes and next steps rather than stopping at the alert itself. Assign a specific person or team the job of translating data into plain language for the people who have to act on it, not just the people who requested it. That translation step is where Mandinach and Gummer's data literacy gap actually gets closed, one report at a time.

Pair every quantitative flag with a qualitative check, a conversation with a teacher, a classroom visit, a look at the student's own account, before treating the number as settled, especially given how often precision and accuracy pull apart in practice. Put a standing review on the calendar, not just an ad hoc meeting triggered when a threshold is crossed, to look at what the data have been saying over the last month rather than reacting to the latest alert in isolation. None of this requires new software. What it does require, however, is the understanding that judgment is a formal step in the process, not an afterthought squeezed in after the dashboard already made the call.

Bowers, A. J., Sprott, R., & Taff, S. A. (2013). Do we know who will drop out? A review of the predictors of dropping out of high school: Precision, sensitivity, and specificity. The High School Journal, 96(2), 77-100. https://doi.org/10.1353/hsj.2013.0000

Datnow, A., & Park, V. (2014). Data-driven leadership. Jossey-Bass.

Mandinach, E. B. (2012). A perfect time for data use: Using data-driven decision making to inform practice. Educational Psychologist, 47(2), 71-85. https://doi.org/10.1080/00461520.2012.667064

Mandinach, E. B., & Gummer, E. S. (2016). Every teacher should succeed with data literacy. Phi Delta Kappan, 97(8), 43-46. https://doi.org/10.1177/0031721716647018

This week's Edupreneurs Network essay makes the same argument this issue makes about data, just applied to purchasing. It argues that the lowest bid, the number that looks like the answer, isn't actually the decision, and that total cost of ownership, implementation quality, and stakeholder fit matter more than the sticker price. That's the same instinct behind treating a dashboard number as one input rather than a verdict. If data-informed leadership resonated this week, this essay shows what that same judgment looks like at the negotiating table.

Thought Leadership in Education: "Digital Transformation and Educational Thought Leadership"

Chapter 5 explores the same terrain this issue does, from a wider lens. It traces how digital tools, from learning management systems to predictive analytics, have reshaped not just how educators teach but how institutions decide. The chapter's section on ethical questions in educational technology, particularly "Privacy: Who Owns the Data?," sits close to this week's argument about precision versus accuracy: technology that looks authoritative isn't the same as technology that's right.

This week: Read "Conceptualizing Technology's Role in Learning and Teaching" and "Addressing Ethical Questions in Educational Technology" in Chapter 5. Then ask yourself: where in your system does a dashboard's authority go unquestioned simply because it's a dashboard?

You don't have to scroll every AI thread, track every new tool, or watch every demo. 

The Rundown AI breaks it all down for you — the latest AI news, tools, and tutorials in one free 5-minute email every morning. 

Trusted by 2M+ professionals at Apple, Google, and NASA.

Sign up to start learning.