Dynamic Pricing For Airlines

Its Role In Digital Retail Thinking

By Alan Dunne - chief innovation officer at Datalex

What is dynamic pricing in 2019?

Airlines are often acknowledged as the trailblazers of dynamic pricing, since they started to offer the same product at multiple different price points in the 1980s and 1990s. However, over the last decade other online retailers have taken dynamic pricing to a whole new level, leveraging new technologies and utilizing new data-driven strategies.

While dynamic pricing is often thought of as the ability to offer a product at different price points, it is really a set of evolving pricing strategies that enables the price to be dynamically determined in real-time based on various criteria.

While many airlines have a relatively high level of maturity in how they price their tickets, their level of pricing maturity for some of their new product offerings is much less developed. This is mainly due to a lack of capabilities in their current systems. These were designed for traditional airline ticket sales and don’t typically lend themselves to the pricing of multiple digital products and service offerings.

Even for airline ticket pricing, airlines are beginning to explore ways to leverage new technologies to enhance current demand-based pricing and experiment with personalized pricing techniques.

Traditionally dynamic pricing was calculated based on a retrospective analysis of sales data across various sales channels. This enabled the dynamic pricing system to forecast market demand by observing cyclical or seasonal patterns. This technique was effective in comparison to static pricing but had some shortcomings.

Estimates of market demand lagged the actual demand in the market. Sales data, while a good estimator of market demand, does not capture all the factors that might influence demand at any given time.

It also relied heavily on cyclical and seasonal patterns which are reliable when the market is stable. But these become less reliable when the market and/or competition become more volatile and hence less predictable.

Modern dynamic pricing solutions use the latest technologies and algorithms to generate real-time dynamic prices that leverage multiple sources of data to more accurately estimate market demand.

Performing these calculations in real-time enables the current market demand to be used to determine the current offered price. The elimination of the time lag between changes in market demand and the reciprocal pricing changes, combined with the more accurate estimation of current market demand, provides a competitive advantage to those who are first to embrace the latest innovations.

Technological advancements changing dynamic pricing

With technology advancements in recent years, approaches to dynamic pricing have changed both in terms of what’s possible from a scalability point of view, and what machine learning algorithms can be leveraged to improve demand estimation.

For example, the emergence of distributed cluster computing platforms has enabled machine learning algorithms, which were previously prohibitively expensive, to be brought into mainstream applications.

These platforms allow the required computations to be broken down into smaller chunks that can then be processed in a massively parallel fashion, leveraging a modern cloud-based infrastructure. This means that proven academic techniques for machine learning such as supervised learning, unsupervised learning, reinforcement learning and deep learning, can now be leveraged in production systems in a timely and cost-effective manner.

So how can machine learning be applied in the context of dynamic pricing?

At a fundamental level, machine learning is the practice of using statistical analysis techniques to make predictions based on data.

There are several categories of machine learning algorithms, each with many individual algorithms to choose from. The key categories of machine learning algorithms include:

* Supervised Learning: Based on known outcomes, the system is trained to identify similar scenarios in new data and predict future outcomes.

* Unsupervised Learning: Leverages collected data to identify repeating patterns to make predictions about both past and current data sets.

* Reinforcement Learning: An autonomous learning technique, learning on the fly using continuous feedback to understand what actions provide a positive or negative reward.

* Deep Learning: Analyses collected data to autonomously identify attributes used to predict an outcome that can be applied to future data sets.

Each of the above categories of machine learning research offers exciting possibilities for the area of dynamic pricing. Supervised and unsupervised learning provide the basis for new dynamic pricing algorithms where expert pricing knowledge can be combined with machine learning automation and scalability to transform how we price today.

Reinforcement learning and deep learning take us a step further, reducing our reliance on manual expert pricing knowledge and enabling the use of autonomous pricing algorithms to maximize the automation and scalability of our dynamic pricing solutions.

While manual oversight and tuning of dynamic pricing decisions will always be required, the more automated and autonomous the solution, the broader the scope that it can be applied to.

How leading retailers use price to drive more than just revenue

Many leading retailers leverage price across verticals to enable successful strategies that could equally be applied in an airline context.

Successful retailers understand that while price is important, it means nothing on its own. An offer ties the product, the price, the value proposition, the customer and the retail context together.

While identifying the right price is key to maximizing revenues and optimizing margins, it can also be used as a lever to influence other important aspects of the business:

* Price as a revenue driver: Increasing prices when demand is strong, and running a promotion when demand is weak; a basic form of dynamic pricing strategy.

* Price as a customer experience driver: Disney’s theme parks use dynamic pricing to impact the customer experience, applying high prices on peak days and discounts off-peak. All customers have a better customer experience with shorter queues, better value for their stay, and more time spent spending more money.

* Price as a lever to control supply and demand: This is common in the consumer power industry where dynamic pricing is used to dampen demand at peak times, on toll roads to manage peak traffic volumes, and at Uber which uses surge pricing to match demand with supply.

* Price as a lever to create a winning business environment: In Major League Baseball in the US, ticket prices will vary based on stadium location, weather, team performance, quality of the opposition, and more. The League has improved its attendances particularly for less popular games. This has resulted in a better atmosphere generated by larger crowds, leading to higher revenues, rising league positions and more high-demand games in contention as the season draws to a close.

As we can see from these examples, dynamic pricing is an effective way to improve lots of common key performance indicators within a business above and beyond the typical revenue-generation use cases.

And all are relevant to airlines, as they strive to manage the finite supply of lounge space, airport check-in areas and premium on-board seating. All these products can be managed through dynamic pricing strategies to guarantee supply for their most valuable customers.

The future of dynamic pricing for airlines

Pricing revolution or evolution?

For airlines, making the transition from their existing pricing systems, which have remained broadly unchanged for decades, to new dynamic pricing solutions can be a daunting prospect. For their ticket pricing, the current pricing systems and standards are deeply embedded in the airline’s sales channels, both direct and indirect.

For other products, airlines often have relatively basic pricing systems that can’t support the technological requirements of dynamic pricing.

We see a couple of trends emerging in this area. For deeply embedded pricing systems, dynamic pricing can be implemented as an add-on capability to the current solutions. In these cases, the dynamic price can be represented as a dynamic mark-up or mark-down from the standard public price. This gives all the benefits of dynamic pricing without the disruption and complexity of swapping out deeply integrated systems.

For other products where the current pricing systems are basic or not deeply integrated in the legacy infrastructure of the airline, it can be more feasible to upgrade the pricing capability by replacing the current solution with a new dynamic pricing solution.

So why is dynamic pricing so important to airline revenues and margins?

For leading retailers, optimizing the price for a given transaction to maximize margin is just the first rung on the ladder to profitability.

These retailers play the long game with their most valuable customers, measuring things such as share of wallet and customer lifetime value to maximize long-term customer revenue share.

But airlines were the original trailblazers of dynamic pricing – the purchase of an air ticket is driven by necessity.

So why would an airline look to other retail verticals for dynamic pricing inspiration?

Your next steps...

To learn more, click here to download the white paper Dynamic Pricing – Its Role in Digital Retail Thinking.

The white paper continues this conversation by exploring what airlines can learn from dynamic pricing in retail:

* The concept of strategic personalized pricing.

* Pricing for discretionary purchases.

* Optimizing revenue using "continuous pricing".

* How AI data-drive insights contribute to the optimized dynamic offer.

Machine learning algorithms and breakthroughs in supporting technologies have extended exponentially the new opportunities afforded by dynamic pricing. Leveraging these opportunities allows the airline to drive innovative and unique product offers seamlessly.

But the game-changer will be the lightning speed with which they will be able to manage their pricing strategy, providing them with the flexibility and control they’ll need to remain competitive well into the future.