Operational Cost Model
Transforming Operational Decision Making
Operational decision making can be transformed using a ground-breaking operational cost model, giving airlines potential for substantive performance and operational cost management.
The primary goal of a flight operations center is to operate an airline’s flight network efficiently and safely. This task involves many decisions that have to be made every day, which affect long-term planning goals, in addition to short-term planning and execution.
Results can be measured in terms of cost (either absolute or relative). Traditionally, cost factors contributing to the overall flight-network costs have been considered as separate factors. However, to ensure a consistent cost picture, they should all be viewed holistically.
In a completely static environment with a disconnected view on flight-operations efficiency targets, efficient flight operations have been achieved at least partially by applying strict on-time performance, assuming that the different kinds of disruption costs are statistically the main cost drivers. And this is fundamentally true.
However, airlines don’t have the luxury of dealing with a static environment. Operations planners, therefore, need to embrace holistic data and move closer to making the right decisions based on individual scenarios.
Without a holistic view on cost data and the support of sophisticated systems, highly capable and responsible people in operations control are mainly relegated to basing their decisions on guy feelings and simple indicators marking certain flights as "critical."
Passenger Misconnect Cost Reduction
An arrival-delay cost function for a single flight, including speed-up costs with a 20-minute departure delay while operating at maximum speed, would allow the flight to arrive with a 10-minute delay and reduce overall passenger misconnect costs.
With the initial huge rise in fuel costs, generally associated with the first oil crisis of the 1970s, fuel started to become a significant factor in flight-operations costs. Up sprang the concept of the cost index, which measures fuel and time costs in relation to one another.
In the late 1980s, aircraft flight management systems introduced cost-index functionality. The cost index was also later implemented in computerized flight-planning systems.
Unfortunately, the only precisely known costs have been fuel costs and air traffic control charges, because sophisticated tools have not been available to capture and assess other important cost factors.
For time-dependent maintenance costs, estimation tools were developed, but they were not equipped to deal with other cost factors (particularly disruption-recovery costs).
A simple approach that is still widely applied for consideration of such costs is to create a default cost-per-delay-minute value that can be inserted to statistically reduce overall flight-operations costs.
Speed-up Option Increases Costs
The arrival-delay function for a heavily delayed flight with around 45 minutes departure delay would not reduce costs even with the maximum speed-up opportunity since the first group of connecting passengers would already miss their flight, hence the flight is planned with ECON cost index.
However, since delay costs are highly dependent on many factors such as passengers, connecting flights and crew rosters, a simple cost-per-delay-minute value is insufficient to correctly consider such costs.
To support this holistic view and initiate precise decision making all the way from flight operations to airline management, a sophisticated operational cost model could provide a place for (and significantly improve the quality of) all cost parameters for flight operations center systems. These cost parameters must represent real values from real-world operations scenarios.
An effective and reliable model must also include both input and output parameters that accept only realistic values that validate a specific real-life scenario. Because of its centralized structure, the operational cost model should enable full control over processes and cost governance.
Centralized storage of cost data is just one part of the story. Obtaining reliable data can be very tricky. So the design of the operational cost model must be capable of handling different approaches to data resolution, meaning it’s best to use whatever constitutes the best data available. Therefore, management of this data-resolution mix should be implemented as a “return-best-available-data” strategy.
A-DCF Including Individual Cost Components
The arrival-delay cost function (A-DCF) shows the individual cost components including passenger misconnection costs, passenger loss-of-good-will costs, speed-up costs and the linear approximation of the A-DCF.
This means, depending on the precise level and type of data available, that the system should be able to handle data all the way from global values, over regional and city-pair-specific data, down to flight-specific data, thus ensuring that heterogeneous data feeds can be used. In addition, the operational cost model should have a “learning” capability, thereby allowing gradual implementation as the system “learns” more information.
Though the following does not constitute an exhaustive list, the operational cost model should be designed to integrate several cost domains:
- Passenger trip-interruption costs
- Passenger loss-of-goodwill costs
- Fuel costs (including tanker fees)
- Time-dependent aircraft maintenance costs (TDMC)
- Air-navigation service-provider costs (such as en route charges, landing fees, etc.)
- Flight-operations costs (fuel and time required for different speed scenarios).
Also, the operational cost model should be capable of integrating other data (apart from cost data) to enable precise predictions, such as:
- Airport gate-to-gate transfer times,
- Air traffic control behavior,
- Airport curfew times,
- Regulatory costs (such as EU261 passenger regulatory costs).
Such data can be used, for example, to compose a flight- and scenario-specific arrival-delay-cost forecast (A-DCF).
Improved Decision-Making Result
The operational cost model enabled by Sabre Intelligence Exchange provides a consolidated view of operational data and enables integrated processes with closed transactions between individual Sabre solutions and any external applications. This leads to more accurate decision-making results in commercial and operational planning.
To help explain why on-the-spot data and analysis are required for efficient flight operations, here are a few examples.
First is a scenario in which a flight is delayed at departure by 20 minutes, which (with tight passenger connections) will cause a majority of passengers to miss their connecting flights at the hub airport.
An operations control center enabled with the operational cost model can effectively manage the passenger-connection stream by using the A-DCF to confirm that going to maximum speed would allow the flight to arrive with only a 10-minute delay and, thereby, to protect the majority of the connecting passengers while simultaneously reducing overall costs (taking into consideration speedup fuel costs, aircraft knock-on delay costs and passenger re-accommodation costs).
In a situation in which a flight is already considerably delayed at departure (such that, even when planned with maximum speed, connecting passengers will miss their connecting flights), it makes sense to completely ignore the delay costs (as well as the knock-on delay effects) since the delay costs cannot be recovered in-flight, and plan the flight with baseline cost-index parameters.
This is definitely an extreme and isolated example, but it illustrates the non-linear influence of irregularity-recovery costs on overall flight costs.
As mentioned, flight-operations costs may also be brought into the picture. A higher speed naturally leads to a higher fuel burn and a lower trip time. Integrating this cost factor into the overall A-DCF will ensure that the desired arrival time determined by the use of data from the operational cost model will be optimized for total costs.
A-DCF can also be analyzed (in any particular scenario) in greater detail. A good example involves the various costs associated with misconnecting passengers.
Reduce Uncertainty In Decision Making
The operational cost model enables OCC operations controllers and integrated Sabre solutions to retrieve precise data related to passenger, maintenance and speed–up costs and reduce decision-making uncertainty, as well as improve the overall operational efficiency of the airline.
Based on the amount of time by which various passenger groups miss their connections, cost amounts can be derived. There’s also an estimate of costs associated with the loss of passenger goodwill. All these costs can be calculated based on the airline’s profile, various passengers’ loyalty status and other factors. In addition, costs are analyzed for speeding up the flight (or slowing down).
Then all cost functions can be summed up, including analysis of disruption-recovery costs. One salient caution involves using a linear approximation of the A-DCF, which can lead to situations in which disruption-recovery costs are either significantly underestimated or significantly overestimated, causing wrong decisions.
Currently, airlines operate based on metrics including on-time performance, which does not provide a good understanding with regard to the overall cost deviations for the airline operation.
An effective operations cost model should enable airlines to get a better understanding with on-cost key performance indicators, which will exert direct impact toward understanding and improving flight profitability and the overall operation.
Another important area within flight operations in which lack of data-sharing and communication can lead to unpredictable and unfavorable results is commercial scheduling, as described in “Redefining Optimal” from Ascend magazine 2015, issue No. 2.
As indicated in the article, data-sharing beyond flight operations should be regarded as a beneficial means to improve performance and efficiency. However, integrating all airline systems with each other in a point-to-point manner does not result in easy governing processes or finding solutions to commercial or operational problems.
Only a system-wide information layer would enable a consolidated view on data, as well as enable integrated processes with closed transactions.
Operational Cost Model
Sabre Airline Solutions has built an operational cost model that includes backend and client applications on the Sabre Intelligence Exchange platform. The model integrates operations information from Sabre AirCentre Movement Manager and Sabre AirCentre Flight Plan Manager, as well as reservations, bookings and customer information from SabreSonic Customer Sales & Service. The operational cost model enables airline OCCs to make cost-based operational decisions considering passenger, maintenance and speed-up opportunities to improve the overall operational efficiency of the airline.
Especially when it comes to commercial planning, which is typically performed months ahead of the actual flight operation, the uncertainty of what is going to happen on the day of operations is inherently high.
Even the most effective operational cost model cannot remove all uncertainty, either within commercial planning, operational planning or execution. But what a robust, reliable operational cost model should significantly improve is the ratio of how often a decision will turn out in the end to be the right decision.
Apart from the increase in the probability of having made the right decision, the increased “traceability” of decisions should better enable post-decision analysis, and should definitely be used to improve future predictions and decision support.
The opportunity to have more information enhances the accuracy ratio of predictions, even when outcomes are by nature uncertain and cannot be precisely predicted. Simply having more data is not necessarily good. Presenting an airline analyst with more data could also potentially lead to less-than-desirable decisions.
Considering all these parameters, special care must be taken in the design of user interfaces, and in the actual use of data, to avoid presenting analysts with too much data which, mixed with biased gut feeling, may lead to wrong interpretation.
An effective operational cost model should enable other software to retrieve precise data for use in solving problems. The end result can then be presented to the analyst as decision support.
The operational cost model GUI enables airline analysts to easily visualize the delay-cost forecast so they can quickly understand the impact delays will have on vital aspects of the operation, such as passenger connections, speed-up opportunities, airport curfews and missed maintenance windows, providing operational cost centers with enhanced decision support.
In partnership with an industry-leading airline, Sabre Airline Solutions is building a proof-of-concept for a sophisticated, fully capable and functional operational cost model. The solution will be implemented on the Sabre Intelligence Exchange platform, integrating reservations and operations data.
The operational cost model will provide decision support to operations controllers for managing delays through an interactive user interface. It will also provide visualization of the delay-cost forecast so operations controllers can quickly understand the impact of delays on costs coming from passenger connections, speedup opportunities, knock-on delays, airport curfews, missing maintenance windows and other factors to help the airline operate on a lowest-cost basis.
Using the new operational cost model, which will be in the market in the coming months, airlines can expect to enhance their operating performance, as well as improve operational cost management.