Education app: a completed lesson is not proof of learning

A lesson marked complete is a record of a visit, not a record of skill. If an education MVP treats those as the same number, teachers and parents will trust a dashboard that only knows someone reached the last screen. For a first release, decide what completion means and what evidence of practice you will actually store.

1. 01

   Consider a short reading exercise for a class of thirty. A student can open the lesson, skip the questions and still trigger a completion event if the product only watches the last page. The teacher then sees a green row and assumes the work was done. The useful first workflow is smaller: the student attempts the practice, the product stores the attempt, and the teacher can see which items were answered, skipped or not started.

2. 02

   Name three states and keep them separate. Not started means the activity has not produced an attempt. In progress means an attempt exists and is unfinished. Completed means the student submitted the practice under the rules you chose, such as answering every required item or spending a minimum useful attempt. A fourth state, reviewed, is only needed if a teacher must confirm the work. Do not collapse reviewed into completed to make the dashboard look tidy.

3. 03

   Define what happens when a student retries. A second attempt can replace the first, sit beside it, or count only after a cooldown. Pick one rule and show it. If a parent can see scores, say whether they are looking at the latest attempt or the first pass. Changing that rule later will rewrite history unless each attempt has a stable identity.

4. 04

   The product can stay useful with a thin teacher queue. List the attempts that need attention, such as skipped required items or repeated low scores on the same skill, instead of a class-wide percentage. A percentage based on mixed events is hard to explain: it may combine opened lessons, submitted practice and teacher marks. Report the numerator and the rule, not a blended progress ring.

5. 05

   An MVP can delay adaptive paths and badges. The tradeoff is less apparent engagement in exchange for a number people can defend. Add a next-lesson recommendation only when it is based on stored attempts, not on the student having tapped Continue. If you later introduce mastery, define the evidence window and keep earlier completions available as history rather than silently upgrading them.

6. 06

   Test the ordinary failures. Open a lesson and leave it, submit with one required item blank, retry after a completed attempt, and open the teacher view before the student returns to the app. Check that a completed row cannot appear from a skipped last screen. Include two students on a shared device: the second session must not inherit the first student's completion. Time-zone changes should not restamp an attempt as a different school day without a recorded reason.

7. 07

   For the pilot, measure submitted attempts, skipped required items, teacher-review backlog and contacts that say the dashboard disagrees with the work. Do not treat a completion event as a learning outcome, and do not collect more student detail than the practice record needs. Review a sample of green rows with the teacher. The useful result is a class view that distinguishes opened, attempted and submitted work so a follow-up conversation has something real to start from.