# Live member sessions ("office hours") for a small paid academy: honest benchmarks and a minimal format

Date: 30 Oct 2026 · Status: PROPOSAL only (0 Academy members per the member-sync log; build nothing until members exist)
Topic: whether and how to run live sessions for the Cedwyn Academy (£49.99/month, 7 graded quizzes, 90% pass mark)

## Benchmarks (what is real, what is not)

1. **altMBA, 96% completion** (self-reported). Wes Kao, then altMBA Director, interviewed by Dr Steph Habif (Stanford d.school), Medium, 19 Sep 2016:
   - 100–130 students per session, 15 coaches, a 10:1 student-to-coach ratio, 4 weeks;
   - **required** live Zoom sessions Tue/Thu/Sun; "Synchronous times for students getting together is critical, and students are required to attend. People push back in the beginning, but then they do very well";
   - coach support on Slack.

   Caveats: the program is selective, expensive and short, so the 96% isn't a causal effect of live sessions. Not checked against altMBA data.
2. **Office-hours attendance barriers.** Hsu et al., CBE-Life Sciences Education 2022 (500+ students, full text read via PMC):
   - the top barrier students cited was a **schedule conflict, 40.9%** (instructors guessed only 16.4%);
   - "intimidated" was 10.9%, and "got help elsewhere" was 11.7%;
   - asked about future attendance, students split: 40.5% more likely in person, **38.1% more likely online**, and women more often preferred online (42.9% vs 27.6% of men, p<0.01).
3. **Accountability lasts longer than plans.** In Kizilcec et al., PNAS 2020 (250,000 learners), the only sustained engagement effect came from "social accountability" (people checking in on progress). Live sessions are one form of that. This is our inference; the study did not test live sessions.
4. No primary benchmark was found for live sessions and **monthly churn** in a paid membership. Vendor blogs (membership platforms, LMS sellers) give no figures with sources. Treat any churn claim as not checked.

## Minimal format (generic)

- **Fixed slot, rotating time.** Schedule conflict is the top barrier, so alternate two slots: an early evening one weekday and a weekend morning one (UK time). Members can make at least one.
- **Online, camera optional, questions can be typed.** This covers both the "intimidated" group and those who prefer online.
- **Structure each session around one record, about 30–40 minutes:**
  - 10 minutes walking through one catalogue item (evidence ID on screen);
  - 15 minutes on member questions, collected in advance in the reading-room thread (academy-discussion-seeded-prompts-instructor-posts.md);
  - 5 minutes on "what record would settle it" (Grade 7 style).
- **Never reveal quiz answers live.** Discuss only material from grades the member has already passed, so the 90% gate stays honest.
- **Recording for members only**, chaptered (youtube-chapters-captions-transcripts-archive.md pattern). The recording is the main product for the 40.9% who can't attend.
- **Measure:** attendance as a share of members, recording views, and first-renewal rate of attenders against non-attenders (observational; say so).
- **Cadence gate:** monthly until 25 members, fortnightly after that, and stop if attendance stays under 3 people for 3 sessions running (send a recording-only "record of the month" instead).

## How we apply it to Cedwyn (proposal)

- Name: "Reading Room Live" (monthly). Members only. Each session takes one GHF-E catalogue item and reads it with members. It stays neutral and third-person, never names private people, uses no banned words and states nothing about family holdings.
- Host: the site's archivist voice, or the project lead if he chooses. Running it with the project lead is his decision; ask only once there are paying members.
- Tooling at zero spend: a free video-call tier for the call, the recording uploaded as an unlisted video linked from /learn/ behind the member gate, and an .ics invite through the existing add-to-calendar pattern.
- Put the first session on the roadmap only after the member-sync log shows 10 paying members. Until then, a recorded "record of the month" video gives the same content without depending on attendance.
- Don't market "96% completion". Cite nothing about live sessions on public pages beyond what is actually run.

## Sources

- Habif, S., interview with Wes Kao, "How to Design an Online Course with a 96% Completion Rate", Medium (Behavior Design), 19 Sep 2016. medium.com (read via a reader proxy; Medium returned 403)
- Hsu, J.L., Rowland-Goldsmith, M. & Schwartz, E.B. (2022), "Student Motivations and Barriers toward Online and In-Person Office Hours in STEM Courses", CBE-Life Sci Educ 21(4). lifescied.org · Europe PMC PMC9727601
- Kizilcec, R.F. et al. (2020), PNAS 117(26). pnas.org · PMC7334459