30 October 2021

Yield pricing (or yield management) is a variable pricing strategy applied to managing goods or services that have a fixed maximum capacity and relatively fixed cost.

Common examples are Airlines, Accommodation, Advertising Inventory, Rental Car Hire, Equipment Hire, and Live Performances.

The operating cost of flying an aircraft route is largely fixed regardless if the aircraft is empty or full of passengers.

There may be more fuel burnt on take off and holding altitude at cruise because the plane is heavier. However, the majority of the cost is airplane lease cost, maintenance, crew cost, and airline overhead recovery. These costs comprise the majority of operating cost and remain the same regardless of passenger loading. And this fixed-cost-regardless-of-utilization problem is at the heart of the yield pricing strategy.

The aim of yield management is to maximize revenue through adjusting pricing in response to the interaction between supply and demand.

Flight schedules are planned around reasonably predictable passenger demand, so the aircraft is very unlikely to fly empty, however if close to the departure date and time there are still empty seats, it makes sense to discount because even selling a seat for $1.00 is better than an empty seat (in practice, government taxes, and airport fees are included in the ticket price, so the minimum ticket price will always be at least enough to cover these variable costs).

In practice this means that every seat on the flight can be sold at a different price (price discrimination) because each seat is sold at a different time, through various channels, and to different customer segments...

• First Class and Business Class seats are premium priced.
• Seats are sold to the travel industry months in advance to be bundled into travel packages.
• Seats are sold to travel groups (bulk bookings) at a discount
• Sold to international airlines as part of the final leg of the journey.
• Seats are allocated to casual booking systems, accessed by travel agents and direct online booking where pricing is frequently updated based on supply and demand.
• Online bookings provide a range of booking types: heavily discounted seats that are non-transferable and non-refundable, and higher priced seats that allow transferring the cost to another flight if last minute changes are needed (fully flexible fares).
• Some seats are allocated to frequent flyer programs.
• Some seats are allocated to flight booking portals (for example WebJet).
• Popular routes may be initially higher priced weeks out from the flight to cater for people who wish to book well in advance (price skimmed) and closer to the flight prices may be further increased if demand looks strong or lowered if a high number of seats are still available.
• However, pricing flights in the future is very much dependent on predicted demand at that time. Events like holiday seasons, special events at the destination city, and popular business commuter days of the week (Monday mornings and Friday afternoons for example) are known times of high demand and advanced booking prices will be set high. Conversely, for periods of low demand Airlines will offer discounted prices for advanced bookings.

The airline monitors the Yield from each flight which is = [ Average Revenue Per seat ] X [ Number of Seats booked ].

Yield is often expressed as a percentage; [ Actual Revenue ] divided by [ Max Possible Target Revenue ] X 100.

The Yield Manager constantly monitors the forward bookings for each flight and regularly decides price and seat availability to each of the above sales channels in an effort to maximise revenue per flight

In an ideal world, every aircraft would be fitted with First Class Seats and sold at their maximum price, in reality there simply aren't enough passengers willing to pay (demand) first class ticket prices, so configuring the aircraft with different classes of seating is the first step toward price differentiation and yield management is the second.

Yield management and competition

Airlines usually compete for passengers on the same routes and many consumers routinely compare prices online before booking.

Airlines use Frequent Flyer Programs (loyalty programs) to combat this, and also use algorithms to tailor pricing offered through online booking systems based on individual profiles.

Knowing that a loyalty card holder is less likely to switch to the opposition airline, the pricing offered at that immediate moment may be slightly higher.

To manage this highly complex and dynamic situation, airlines assign a Yield Manager to each airline route to constantly review forward bookings for each flight, continuously monitor forecast yields, and making seat allocations to different market segments, distribution channels, and adjusting pricing.

Some airlines may even over-sell tickets (continue to sell seats at higher prices after all seats have been booked) and then bump-off the lowest yielding ticket holders and put them on a later, less-heavily booked flight.

This is one of the risks of buying discount tickets.

The decision as to who to bump could be algorithmically selected based on...

• What did they pay for their ticket?
• Are they a corporate account customer?
• Are they a frequent flyer or club member?
• Do they have a connecting flight at the other end?
• Have they checked-in luggage?
• Are they flying with others?
• Are they an airline employee flying FOC (the first to get bumped).

If the whole flight is under-booked and they can fit the passengers (and any crew that have to get home) on the next flight - they may bump the whole flight.
If you have flown on a near empty flight and wondered why they didn't cancel the flight it would be because they needed the aircraft at the other end by a certain time.

Artificial Intelligence and yield management

Most airlines now use Artificial Intelligence systems to either assist the Yield Manager or fully automate yield management decisions.

AI algorithms can now dynamically (and in an instant) take into account a huge range of additional factors to determine the precise price and the precise time to offer a customer looking to book a flight.

• Location: Airlines may charge different prices for the same flight depending on where the customer is booking from. This is known as price discrimination or dynamic pricing, and it is based on the assumption that customers in different markets have different willingness to pay.
• Past purchase: Airlines may use the customer's purchase history and preferences to offer personalized prices or recommendations. This is known as behavioral pricing or targeted pricing, and it is based on the use of data and analytics to segment and profile customers. For example, an airline may offer a lower price to a customer who has searched for the same flight multiple times, or a higher price to a customer who has booked a premium seat in the past. The system will also look at your tendency to book cheap flights, or how long you take to make decisions - do you book on average within 10 minutes of first searching or several days? Are you booking for yourself or is your booking being done by a secretary or purchasing officer?
• Competitive searches: The algorithm may even offer a different price if they can see you have being comparing pricing with different airlines.
• Who knows what else: the thing about Artificial Intelligence is that it gets smarter with experience. It accumulates massive data about its past price offers with the outcome (booked/didn't book) and matches that with hundreds of other stored characteristics.
  
  It doesn't care if the identified factors make no sense, it just identifies what they are. For example, it might determine that maximum price can be obtained from people booking between 11:00 PM and Midnight 6 days before a flight, when there is a full moon. It will also scan back through all of the past bookings at the higher pricing and identify the common features of those people and any other information peculiar to those bookings and look for similar patterns when offering pricing.
  
  It might also work out that people buying at the low end of a price bracket (example: $302) are just as likley to book at $322. But, crossing into a new price bracket (say: from $199 to $200) significantly reduces the chance of booking. Hence, the yield management system will just jump straight to $222 if it perceives that higher pricing is possible.
  
  Again, AI systems don't need to understand WHY, they work by discerning patterns.

Yield Management, price positioning, and brand image

While the aim of yield management is to maximise the revenue for each flight on the route, the yield manager must also consider the longer term effects of continuously discounting seats to maximise yields and how the airline is positioned against its competitors. Overtime regular flyers will adjust their buying strategies if they can discern a pattern to the pricing.

Sometimes long term yields can be maximised through maintaining higher average seat prices even if some seats remain un-sold.

Airlines have one considerable advantage over yield management in other industries which is the ability to fill seats from frequent flyer programs. These seats are selectively released for frequent flyer bookings using a secondary pricing system - varying the number of points needed to book a seat. It is much harder for customers to convert points to dollars particularly when the relationship between the two is dynamic.

Similarly, airlines can also sell seats through channels (travel operators, booking portals, etc.) which helps prevent online comparison.

Where an airline has a higher quality reputation against competitors (more leg room, higher in-cabin service, faster check in times, better customer service) customers will discern a worthwhile difference and pay on-average higher prices. The Yield Manager must not interfere with this positioning through over discounting or being too aggressive with bumping customers.

Yield management and AI (artificial intelligence) as the ultimate test of brand building

I speculate that if an airline switched over to fully automatic price setting (leave it up to the AI system) then over time the average yields would start to increase IF service improvements and perceived brand image were working.

That would be the ultimate test - much more meaningful than market research.

People will pay more for a product (or service) they perceive as more desirable.

The A and B brand strategy

Some airlines operate two brands that offer different levels of comfort and service.

QANTAS for example flies under both QANTAS and JETSTAR.

The JETSTAR brand is set-up with lower operating cost, lower standards of inflight service, longer check-in times, operates from gates at far ends of the terminal and often without aero-bridges, has less generous luggage allowances built in to the price, tighter seating configurations, and no first class or business class seating.

QANTAS works hard to maintain a clear perceived quality difference between the two airlines thus demonstrating to the customer why they should pay more to fly QANTAS. In addition it limits the ability of their main competitor VIRGIN to gain market share through discounting because JETSTAR is clearly positioned as the low price alternative.

Having two brands operated by the same airline provides greater market coverage and control of market pricing.

In addition, further cost savings are achieved for both airlines through shared IT systems (reservation and ticketing for example), shared back-office functions, and greater purchasing power for goods and services through having higher purchasing volumes.

The JETSTAR brand also presents a significant barrier to entry for a new low cost carrier that may upset the dual-airline duopoly (QANTAS and VIRGIN).

By Justin Wearne

Other pricing strategies:

Cost Plus Pricing
Penetration Pricing
Price Skimming
Premium Pricing
Price Bundling
Discriminatory Pricing

Pricing strategy overview:
Pricing Strategy