FlySafair vs NCC — Overbooking: Claims vs Actuals

Source: FlySafair Sales system (QlikView), 145,138 flights, one row per flight per departure date, Jan 2024 – May 2026. “Oversold” = booked passengers > seats (capacity) on the flight. Oversold-seat revenue is a derived estimate (oversold seats × average fare), shown both excl. and incl. tax. Denied-boarding counts are the NCC’s figures — not derivable from this dataset.

The headline numbers

Penalty NCC seeks

= 10% of FY2025 turnover (R13.29bn)

NCC overbooking revenue claim

R114m/yr

Our actual: /yr — overstated

Overbooking stopped

Apr 2025

Oversold flights 16.0% → 2.8%

Load-factor effect of stopping

Flown LF, pre vs post — negligible

Claims vs actuals — reconciliation (NCC referral window)

Overbooked-flight counts

Our count from the Sales system reproduces FlySafair’s reported figures to within 0–2 flights. The NCC’s counts run 12–15% higher.

What the NCC claims vs what we find

Metric	NCC	Actual	Gap

Overbooking revenue — the R114m test

Monthly overbooking-seat revenue: NCC vs actual

NCC alleges ~R9.5m/month. Actual oversold-seat revenue (even incl. tax) is ~4× lower. The driver is the passenger count, not tax.

Overbooked passengers: NCC vs seats actually sold over capacity

The NCC counts ~5× more overbooked pax than seats actually sold beyond capacity (net, at departure) — the single assumption that inflates the revenue.

Overbooking since January 2024 — and the April 2025 stop

% of flights oversold, by month

Practice ran at ~16% of flights through Mar 2025, then collapses to near zero from April 2025.

Effect on FlySafair of stopping overbooking

Are we losing money by not overbooking? — demand-controlled spoilage

Flown load factor at equal demand: overbooking ON vs OFF

Flights bucketed by booked-demand level, then we compare how full they actually flew pre vs post-stop. This strips out market/seasonal demand swings that the raw before/after number can’t.

Booked-demand bucket	Flown LF — OB on	Flown LF — OB off	Diff

The verdict

High-demand flights (≥97% booked)

flown LF, OB-on vs OB-off — within noise

Demand-controlled revenue loss

≈ R0

planes fly just as full without overbooking

Denied-boarding comp avoided

+ reaccommodation, meals, admin (ops data)

Gross “R25m forgone” overstates: denied pax were largely reaccommodated (revenue kept) and no-show seats weren’t filled at volume. Netted against avoided costs + the R1.33bn / brand exposure removed → stopping is net-positive, not a loss. Confirm with denied-boarding spill/recapture data from ops.

Monthly overbooking statistics (Jan 2024 – May 2026)

Month	Flights	Oversold	% oversold	Oversold seats	Max oversell	Booked LF	Flown LF	No-show %	OB rev (ex-tax)

Prepared for the FlySafair / NCC overbooking matter. Operational actuals from the FlySafair Sales system; oversold-seat revenue is an estimate (oversold seats × avg fare). For legal use, confirm the NCC’s overbooked-passenger definition against the referral (see ncc-v-flysafair-referral-FULL.md) before relying on the revenue comparison.