NP-RM-14 · O&D Revenue Management · Airlines Process Wiki

📊

Process Flow Diagram (BPMN)

📋

L4 Process Steps

Step	Step Name	Role / Swim Lane	System	Input	Output	KPI	Dec?	Exc?
Phase 1 1.1	Extract O&D booking history from data lake	Revenue Management Analyst	AWS Redshift	Daily booking feed from Amadeus Altéa PSS (12-month rolling window)	Cleaned O&D demand dataset segmented by cabin, point-of-sale, and booking lead time	Data completeness ≥99%; extraction latency <2 hours from departure day-close	N	N
1.2	Segment O&D demand by fare class and channel	Revenue Management Analyst	Amadeus Revenue Management (NRM)	Cleaned O&D demand dataset from AWS Redshift	Demand segments by fare class (Y/B/M/H/Q/V/W/L/K/G), booking channel, and travel purpose (leisure/business)	Segment classification accuracy ≥92%; misclassified itineraries <5% of O&D pairs	N	N
1.3	Generate 90-day O&D demand forecast	Senior Revenue Management Analyst	Amadeus SkyCAST	Segmented demand history, seasonal indices, special-events calendar	90-day O&D demand forecast by fare class and day-of-week; confidence intervals at 80% and 95%	Mean Absolute Percentage Error (MAPE) ≤8% at 60-day horizon; ≤12% at 90-day horizon	N	N
1.4	Validate forecast against booking pace signals	Senior Revenue Management Analyst	Amadeus Revenue Management (NRM)	90-day O&D demand forecast; live booking-curve actuals from Altéa PSS	Validated forecast or rework flag with variance root-cause note	Forecast validation cycle <4 hours; >95% of O&D pairs accepted without manual override	Y	Y
Phase 2 2.1	Calculate network contribution per itinerary	Revenue Management Analyst	Amadeus Revenue Management (NRM)	Validated O&D demand forecast; filed fare inventory from ATPCO; connection rules from Amadeus Altéa	O&D contribution matrix: revenue yield, spill cost, and recapture probability per itinerary	Network revenue contribution accuracy within ±3% of post-departure actuals; matrix refresh latency ≤6 hours	N	N
2.2	Benchmark competitive fares by O&D market	Revenue Management Analyst	OAG Schedule Analyser	Competitor schedule data (OAG), ATPCO published fares, Sabre GDS fare feeds	Competitive gap analysis: own-fare vs. lowest available competitor fare by O&D and booking lead time	Competitive fare index maintained within ±5% of market median for top-50 revenue O&D pairs	N	N
2.3	Assess O&D contribution validity	Senior Revenue Management Analyst	Amadeus Revenue Management (NRM)	O&D contribution matrix; competitive gap analysis	Approved contribution matrix or rework instruction with adjustment rationale	Gate review cycle ≤2 hours; <10% of O&D pairs requiring manual contribution override	Y	N
Phase 3 3.1	Run LP network optimisation for bid prices	Revenue Management System (automated)	Amadeus Revenue Management (NRM)	Approved O&D contribution matrix; 90-day demand forecast; current flight leg capacities from Altéa	Shadow prices (bid prices) per leg-class combination for each active departure window	Optimisation solve time ≤45 minutes for full network; bid price solution feasibility ≥99.5% of legs	N	N
3.2	Review bid price reasonableness	Senior Revenue Management Analyst	Amadeus Revenue Management (NRM)	Generated bid prices; prior-day bid prices; yield targets by O&D	Validated bid prices or escalation to RM Manager with variance explanation	Bid price variance vs. prior day flagged if >20%; analyst review cycle ≤90 minutes	Y	Y
3.3	Publish approved bid prices to inventory	Revenue Management System (automated)	Amadeus Revenue Management (NRM)	Validated bid prices	Bid prices committed to Altéa PSS availability engine; audit log entry with timestamp and analyst sign-off	Bid price publication latency <15 minutes from approval; 100% of active departures covered within the 90-day window	N	N
Phase 4 4.1	Apply O&D class controls to Altéa PSS	Revenue Management Analyst	Amadeus Altéa PSS	Published bid prices; authorisation levels (AU) per fare class from NRM	Updated fare class open/close status in Altéa inventory; availability snapshot distributed to GDS and NDC channels	Class control accuracy ≥99.8% (no misapplied open/close); distribution sync to all GDS channels <5 minutes	N	N
4.2	Monitor intraday booking pace vs. forecast	Revenue Management Analyst	Amadeus Revenue Management (NRM)	Real-time bookings from Altéa PSS; validated O&D demand forecast	Intraday booking pace report; pace-vs-forecast variance by O&D and fare class	Monitoring refresh interval ≤30 minutes; pace deviation alerts triggered at ±10% threshold	N	N
4.3	Evaluate booking pace deviation trigger	Senior Revenue Management Analyst	Amadeus Revenue Management (NRM)	Intraday pace report; alert flags for O&Ds exceeding ±10% deviation	Decision: hold current controls, tighten availability, or open additional classes	Override response time <60 minutes from alert; revenue impact of late override <0.5% of flight revenue	Y	Y
4.4	Execute manual availability override	Senior Revenue Management Analyst	Amadeus Altéa PSS	Override instruction (tighten or open classes); fare class control screen in Altéa	Revised class availability posted to Altéa; override action logged with business justification	Override-to-publish latency <10 minutes; 100% of overrides logged with analyst ID and rationale	N	Y
Phase 5 5.1	Capture post-departure O&D revenue actuals	Revenue Management Analyst	AWS Redshift	Final departure manifest from Altéa; ticketing revenue records from BSP / ARC settlement	Post-departure O&D revenue report: RASM, yield, load factor, spill, and spoilage by fare class	Report availability ≤24 hours post-departure; O&D revenue reconciliation variance vs. settlement <0.2%	N	N
5.2	Evaluate demand model accuracy	Senior Revenue Management Analyst	Amadeus Revenue Management (NRM)	Post-departure actuals; corresponding pre-departure forecasts	Model accuracy scorecard (MAPE by O&D, phase, and departure horizon); recalibration flag if accuracy <85%	Model accuracy review completed within 5 business days of departure; ≥85% of O&D pairs meeting MAPE target	Y	N
5.3	Recalibrate O&D demand models	Senior Revenue Management Analyst	Amadeus Revenue Management (NRM)	Recalibration flag; post-departure actuals; external demand signals (GDP, events, competitor changes)	Updated model parameters committed to NRM; change log with effective date and recalibration rationale	Recalibration turnaround ≤3 business days; post-recalibration MAPE improvement ≥15% on flagged O&Ds	N	N

📋

Process Attributes

Identification

Process IDNP-RM-14

L1 DomainNetwork Planning & Scheduling

L2 ProcessRevenue Management

L3 NameO&D Revenue Management

L4 Steps17 across 5 phases

Decision Gates5 (all with iteration loops)

Exceptions4 documented

Swim Lanes (Roles)

Revenue Management Analyst

Senior Revenue Management Analyst

Revenue Management System (automated)

Systems & Tools

AWS RedshiftAmadeus Revenue Management (NRM)Amadeus SkyCASTOAG Schedule AnalyserAmadeus Altéa PSS

Key Performance Indicators

Extract O&D booking history from data lakeData completeness ≥99%; extraction latency <2 hours from departure day-close

Segment O&D demand by fare class and channelSegment classification accuracy ≥92%; misclassified itineraries <5% of O&D pairs

Generate 90-day O&D demand forecastMean Absolute Percentage Error (MAPE) ≤8% at 60-day horizon; ≤12% at 90-day horizon

Validate forecast against booking pace signalsForecast validation cycle <4 hours; >95% of O&D pairs accepted without manual override

Calculate network contribution per itineraryNetwork revenue contribution accuracy within ±3% of post-departure actuals; matrix refresh latency ≤6 hours

Benchmark competitive fares by O&D marketCompetitive fare index maintained within ±5% of market median for top-50 revenue O&D pairs

Assess O&D contribution validityGate review cycle ≤2 hours; <10% of O&D pairs requiring manual contribution override

Run LP network optimisation for bid pricesOptimisation solve time ≤45 minutes for full network; bid price solution feasibility ≥99.5% of legs

Airline-Specific Risks & Pain Points

Connecting itineraries booked via GDS often log leg-level data only; O&D reconstruction requires custom passenger name record (PNR) stitching logic that breaks during irregular operations

Carriers without seat assignment (open-boarding models) lack seat-map signals for travel-purpose inference, reducing segmentation accuracy vs. assigned-seat competitors

SkyCAST demand models require minimum 18 months of stable history per O&D; new routes or routes disrupted by major events (e.g., network collapse, weather) cause cold-start forecast degradation

Sudden demand spikes driven by flash sales or competitor capacity withdrawal are not captured by statistical models, requiring analyst intervention outside automated tolerance bands

Point-to-point network topology limits connecting itinerary combinations available for O&D value netting; contribution calculation must account for high-frequency same-city multi-flight options

Competitor dynamic pricing and NDC fare products are not consistently visible through ATPCO filings; private corporate fares and carrier-direct channel fares create persistent blind spots in competitive benchmarking

Inputs / Outputs

Primary InputDaily booking feed from Amadeus Altéa PSS (12-month rolling window)

Primary OutputUpdated model parameters committed to NRM; change log with effective date and recalibration rationale

←PreviousNP-RM-13 · Revenue Integrity & Ticketing Compliance NextNP-RM-15 · LCC Competitive Pricing Strategy→