When tied to learning goals at the classroom level, grades are certainly appropriate indicators of learning. In the context of program assessment, however, grades can be problematic. A grade, by itself, does not communicate what a student has learned, and only a little about how much.
Grades only communicate information that is only indirectly linked to learning, like attendance, effort, participation, or test-taking skills. These indicators are certainly related to learning, but in many cases, students can achieve high grades due to factors that are only indirectly related to learning outcomes.
Using grades in program assessment become more problematic in discussions with colleagues about program improvement. Assessment should be a collaborative process, and it is difficult to see what an A, B, C, D, or F – buy itself and with no supporting information or context – communicates, or how the information can be used for improvement.
There are effective strategies for using grades in assessment. A good resource is Linda Nilson's Specifications Grading: Restoring Rigor, Motivating Students, and Saving Faculty Time. Further strategies and case studies are presented in Effective Grading: A Tool for Learning and Assessment in College.
Elements adapted from Suskie, L. (2009). Assessing Student Learning: A Common Sense Guide. San Francisco: Wiley.