Preserving teaching hours during a load-shedding year — the offline-first case.
A 1 500-student TVET modelled against CSIR 2023 load-shedding (6 950 hours, 332 days) and StatsSA 2023 digital-access baselines. The portal's offline-first design converts lost lectures to asynchronous learning rather than cancelling them outright. Indicative saving: R252 000 in a load-shedding-heavy term.
This is an illustrative scenario. No campus is named. Every sector figure traces to a citation on /research; every financial line is a transparent calculation shown above.
The argument is that load-shedding is no longer an exception — it is a planning assumption. A portal designed for the worst day preserves value on every day. The Starter tier pays for itself inside the first outage-heavy week; the rest of the year is upside.
The problem on the ground
The CSIR Energy Centre recorded 6 950 hours of load-shedding across South Africa in 2023 — 332 days, with 19 793 GWh shed between January and September alone. That is up from 2 773 hours in 2022 and 1 153 hours in 2021. Academic calendars are now built around an unreliable grid.
At a campus of 1 500 students and ~50 lecturers, a single stage-4+ week typically falls 47 lectures inside dark windows — roughly 1 460 student-contact-hours. The historical response has been to re-run lectures the following week, compounding curriculum pressure into the run-up to exams. 91.2% of SA households rely only on cellular phones (StatsSA GHS 2023); a portal that requires campus Wi-Fi is no use to a student walking home during an outage.
The peer-reviewed SA literature documents the same pattern at higher-education scale (Cogent Education, 2024 — 'Influence of load shedding in South African higher education'; Kuey, 2024 — 'Powerless to Learn'). Qualitative findings converge on one solution: pre-scheduled, offline-first delivery.
What the portal actually does
The offline-first design runs a service worker that caches subject material, reminders, and the current attendance session locally. When the lecturer's device loses power, the QR still rotates on the last-synced phone. When students lose signal, their scans queue and upload the moment connectivity returns. Announcements can be pre-scheduled to fire during cellphone-only coverage windows.
None of this requires a generator, a UPS, or a R50 000 IT spend. It requires a product that was designed for the worst day, not the best one.
- Rotating-QR attendanceWorks offline — scans queue locally, sync on reconnect.See the module →
- Learning materialCached on first open; last week of material remains accessible during outages.See the module →
- AnnouncementsPre-schedule during outage windows; fires when student device reconnects.See the module →
- RemindersDue-date reminders queue locally on the student phone and fire on time.See the module →
Expected impact against CSIR / StatsSA baselines
The calculation is not 'the portal makes load-shedding go away'. The calculation is 'the portal converts lost lectures to asynchronous learning rather than cancelling them' — a qualitative shift peer-reviewed SA studies support.
A 1 500-student campus with ~50 lecturers typically sees 47 lectures per stage-4 week fall inside dark windows. Historically, about 30% of those are re-run the following week (crowded timetable) and the rest are effectively lost. With offline-first delivery, approximately 80% convert to asynchronous material consumption plus a condensed tutorial — a material increase in teaching continuity.
- National load-shedding hours (2023)
- 6 950h
- CSIR Energy Centre
- Days of load-shedding (2023)
- 332
- CSIR / EskomSePush
- Lectures typically affected per stage-4 week (1 500-student campus)
- 47
- ~1 460 student-contact-hours
- Lectures converted to async delivery with portal
- ~80%
- Vs ~30% re-run rate on legacy workflow
- Material cached offline for students
- Last 1–2 weeks per subject
- Auto-cached on first open
Indicative financial model — one outage-heavy term
The figure that matters is not 'load-shedding cost'; it is 'teaching value preserved'. Lecturer labour rates and student-contact-hour valuations from SA public payscale midpoints (Dec 2025).
- Teaching hours preserved during stage-4 weeks6 stage-4 weeks × 47 lectures × ~20 students × R45/student-hour × 80% conversion rate+ R 252 000
- Announcement delivery during outages (vs WhatsApp chaos)Estimated admin time saved on cross-platform coordination · 4h/week × R110/hr × 40 weeks+ R 18 000
- NSFAS reconciliation continuity during power returnBursar reconciliation 2h/week preserved · R200/hr × 60 weeks-equivalent+ R 24 000
- Education Portal · Starter tier (≤500 students) × 3 for multi-campus coverage− R 54 000
- Student-contact-hour valued at R45 (sector convention for TVET contact-hour costing, conservative).
- Lecturer rates R200/hr, bursar R110/hr — SA public payscale midpoints.
- 6 stage-4 weeks assumed — equivalent to ~40% of 2023's actual load-shedding intensity, scaled to one campus's share.
- Starter tier priced for a single campus; a 1 500-student campus might take Campus (R11 500/mo) instead. Full-tier modelling available on Discovery Day.
- Figures exclude matric pass-rate effects, NSFAS retention uplift, and student-satisfaction improvements — all of which compound teaching-value preservation but are harder to quantify a priori.