A cohort of twelve senior individual contributors. A twelve-week remote program. By week five, four people had stopped attending and two had formally withdrawn.
What the data showed
Exit interviews revealed three separate problems: sessions ran too long, group size made it hard to speak, and participants did not see how the content connected to their actual work.
The rebuild process
- Cut session length from ninety minutes to sixty. Remove the content delivery portion and move it to async pre-reading.
- Split the cohort into smaller groups of four for discussion segments. Rotate the groups every two weeks.
- At the start of each session, ask one participant to name a real problem they are facing. Build the session around it.
- Send a short written summary after each session with one question for the week. Keep it under one hundred words.
- Check attendance patterns weekly, not monthly. A two-week drop is a signal, not a coincidence.
The remaining eight participants completed the program. Three extended individually. The format changes were not creative — they were corrections to obvious structural errors I had missed during design.
Group remote programs fail at the format level more often than at the content level.