Why are so many educators willing to give credence to observations based on announced visits? Perhaps it's avoidance or a failure to distinguish between good teachers and good teaching, or perhaps it's the way the conventional teacher-evaluation model limits administrators' options. To put it bluntly, an evaluation process that relies on announced visits is inaccurate, dishonest, and ineffective. A growing number of principals around the country are experimenting with an alternative to the dog-and-pony show--an approach intended to win teachers' trust, make better use of time, enhance instructional leadership and collegiality, and usher in improvements in teaching and learning. It has three layers: changing the structure, the human element, and management details.