Lesson Observation vs Learning Evidence: Two Lessons a Year
Lesson observation sees one performed lesson. Learning evidence shows a unit. Why we keep observations to two a year and where the evidence already sits.
A school with sixty teachers and two lesson observation slots each spends about a hundred hours a year watching lessons. Each hour sees fifty minutes of a lesson that was prepared for a visitor. Heads tell me they learn something from it, and they do. But when I ask "What did the students learn this term?" the observation notes cannot answer. They were never about that.
This post is about the difference between a lesson observation and learning evidence, why we would keep observations to two a year, and where the evidence of learning already sits in your school.
What a lesson observation can and cannot tell you
| It can tell you | It cannot tell you |
|---|---|
| How a teacher explains one idea on one day | Whether the class understood it a week later |
| Whether the room is calm and the routines work | What the quiet students in row four were thinking |
| Whether the plan was followed | Whether the plan was right for this class |
| What the teacher does under a visitor's eye | What the teacher does on a normal Thursday |
None of this makes observation useless. It makes it a tool for one job: a developmental conversation about teaching. It is a poor tool for the other job schools keep asking of it, which is knowing whether learning happened.
Learning walks and lesson observation tools have the same limit
Learning walks (short, unannounced visits to several classrooms) were meant to fix the performance problem. They are shorter and less staged, and they do give a feel for a corridor. But a ten-minute visit is still a snapshot of teaching, not a record of learning.
The same is true of a lesson observation tool with forty criteria on a tablet. More boxes do not change what is being looked at. The teacher is still the subject. The learners are still the background.
Our position at TeachBuddy AI is simple. We don't observe teachers. We listen to learners.
Learning evidence: five sources your school already has
Learning evidence is anything a student produced or said that shows what they can now do. Most schools already have five sources and look at none of them together.
- Student work over weeks. A worksheet from week 2 beside one from week 6, same student, same objective.
- Quick checks and quiz results. An exit ticket or a short auto-marked quiz after each lesson, per objective, per class.
- The teacher's own reflection. Two minutes after the lesson: what was used, what landed, what did not. Written the same day, not for a visitor.
- What students say. A short survey each term, or every student recording a one-sentence answer to "explain today's idea in your own words".
- Learner notes. What the teacher recorded about a student's needs and progress, approved by the teacher, changing over time.
Put these five beside each other for one class and one unit, and you know more about learning than ten observations could tell you. The teacher did not perform anything. The students did the showing.
Two observed lessons a year, and what to do with them
Keep lesson observation. Keep it to two a year per teacher, one chosen by the teacher and one agreed together. Make both developmental: the observer comes to help with something the teacher named in advance.
The lesson observation form then needs three boxes, not forty:
- What the teacher asked me to look for.
- What I saw the students do when that happened.
- One thing to try next, agreed in the conversation afterwards.
Separate this entirely from any decision about a teacher's job. The moment a lesson observation carries a consequence, you will see a performed lesson, and you will learn less. We write more about that in teacher evaluation software: why we would rather listen to learners.
An example: Grade 9 photosynthesis
A Grade 9 science teacher is observed teaching photosynthesis. The observer notes a clear explanation, a good diagram, students on task. The lesson observation is positive. Three weeks later the end-of-unit test shows that half the class still thinks plants get their food from the soil.
Now the same unit with learning evidence. After the lesson, a five-question quiz: two thirds of the class answer the "where does the mass come from" question wrong. The teacher's reflection says "the diagram landed; the soil misconception did not move". Her next plan starts with that misconception. A class voice activity a week later asks every student to explain, in one recorded sentence, where a tree's mass comes from. She listens to thirty answers in ten minutes and hears which students still need the idea. The unit test is better. Nobody visited. For quick checks that do this, see formative assessment examples: 15 quick checks.
How TeachBuddy AI helps
TeachBuddy AI is built around the five sources. The two-minute reflection after teaching asks what was used and what landed, then proposes updates to the class description and learner notes that the teacher approves item by item. Quizzes are auto-graded. In Spaces, class voice lets every student record an answer, interactive worksheets are marked automatically for objective questions, and the AI chat tutor keeps learner notes the teacher can read. Surveys ask students and parents directly. Read how class voice changes who gets heard in class voice: hear from every student.
For a school leader, My School shows what was planned and submitted, and Buddy can answer "who has not submitted this week". It does not record teachers. It records what learners were given and what they showed.
Frequently asked questions
What is a lesson observation tool?
Software or a form that an observer fills in while watching a lesson, usually against a list of criteria about the teacher. It records teaching on one day; it does not record learning over a unit.
How many lesson observations should a teacher have a year?
Two is enough when they are developmental and the teacher chooses the focus. More observations add performance, not information. Put the extra time into looking at student work and quiz results together.
Are learning walks better than formal observations?
They are shorter and less staged, which helps. They are still a snapshot of teaching. Pair them with learning evidence from the same classes or they tell you about corridors, not learning.
If your school wants to spend less time watching lessons and more time reading what learners show, talk to us about your school.
Hammad advises school leaders and education organisations on AI strategy, growth and community. He writes about how schools adopt AI without losing sight of learners.