Start a learning analytics discussion with governance

Learning analytics becomes risky when a buyer hears prediction and assumes certainty. An account executive should ask what decision a team hopes to improve, which data represents the students involved, and who governs use. A model cannot repair missing sections or resolve a faculty concern by itself.

Provost Hale says, "Can your analytics tell us which students will fail? Faculty are worried about bias." Account executive Imani replies, "Analytics can surface patterns from the data available, and your institution decides how those patterns are used. Which decision would an alert inform, and which course sections are absent from the current data?" Hale says adjunct led sections are missing. Imani continues, "That gap needs to be visible in any evaluation. Could we meet with institutional research, privacy, and faculty senate to define a limited question and the review criteria? We can show how inputs and alerts are documented during that session."

Imani avoids describing the tool as an authority over students. She makes governance part of the evaluation rather than a late approval step. The institution can then decide whether a limited use case has enough reliable information.

Take a proposed dashboard and write the decision it would support in one sentence. Ask a colleague to name one missing data source. In roleplay, acknowledge the gap and invite the owner of that data into the evaluation meeting.