Why do restaurants hold happy hours and late-night deals? Why does Uber surge? Why can the same airline trip vary by hundreds of dollars depending on when you book it? It all boils down to dynamic pricing, or changing prices based on market demand and contributing factors.
Dynamic pricing helps airlines maximize revenue. Cheaper seats are available to the customer willing to book way ahead of time (or willing to take the center seat in the last row). And now, air cargo carriers are looking at and adopting dynamic pricing methodologies. So let’s take a closer look at dynamic pricing techniques and benefits.
Dynamic Pricing: Good for the Buyer and Seller
Critics of dynamic pricing attack sellers, claiming they take advantage of the customer’s lack of pricing knowledge. Price gouging and monopolies, for example, can result in sellers setting higher prices than expected.
However, dynamic pricing can also benefit the customer. Standard shipping rates are lower than rush rates, resulting in cheaper prices for the customer who doesn’t mind waiting an extra few days for delivery. The key is for air cargo companies to clearly lay out their pricing to customers. If customers understand how the final price tag is calculated, they’re more likely to cooperate. That way, the customer has a choice.
What Influences Dynamic Pricing?
Dynamic pricing gives air cargo companies the flexibility needed to adjust their rates according to a variety of factors, including:
- Buying Power of the Targeted Segment: When demand from the carrier’s customer base rises, the carrier can charge higher prices.
- Market Trends: Overall fluctuations in the air cargo industry will influence pricing. For example, surging demand will drive prices up across all companies.
- Seasonal Variations: Fluctuations in demand, such as around the holidays, can drive up prices as air cargo space becomes limited.
- Historic Demand: Forecasting algorithms rely in part on historical data to determine current prices.
To go one step further, the factors above influence dynamic pricing because they themselves are influenced by variable costs. Unlike fixed costs, variable costs will fluctuate depending on the air cargo carrier’s shipping volume. The higher the volume, the higher the variable costs. Some examples include:
- Fuel Surcharges
- Origin-Destination (O-D) Pair Transport Cost
- Truck Cost
- Warehousing Costs
It’s important to note that many of these factors have little to do with the actual delivery of the product. The delivery of a product by air exists apart from this week’s fuel prices. To get the most out of dynamic pricing, air cargo carriers need to turn to resources that pool all these factors together to determine the final pricing. That’s where algorithms and analytics come in.
The Random Forest Algorithm
The air cargo industry uses dynamic algorithms that predict and recommend costing and pricing based on historical, current, and forecasted data. Actual costs are next compared to these predicted values. Any price deviations are used to update the algorithm for more accurate estimations in the future.
One common algorithm is the random forest, where a decision (in this case, the final price) is made based on a series of factors answered in a specified order. The decision tree is the first step toward creating a random tree algorithm. Let’s take a look at the chart below:
Based on the factors of weather, humidity, and wind, the decision tree tells us whether or not to play a scheduled game. Of course, in real-world applications there are many factors that affect a desired result, so you’d want to include them in a decision tree as well.
Yet this presents another problem: listing the same factors in the same order would always lead to the same outcome. Dynamic pricing is just that, dynamic. So what can be done to ensure different yet accurate results?
Enter in the forest, a series of decisions trees whose factors and data are chosen at random. One tree might have weather, temperature, and humidity. Another might have weather, field condition, and time of day. The first tree might have more detailed weather data than the second. A third tree may have general humidity data. Based on the decisions made in each tree, the forest as a whole can make the best decision.
For dynamic pricing, the trees that examine random sets of variables, from historical data to fuel surcharges, come together to determine the final price. Companies like Airbnb already use an algorithm similar to the random forest to help hosts price their homes.
Moving Forward with Dynamic Pricing
Today’s dynamic pricing algorithms are self-learning, meaning they can take new data to ensure outcomes are current and reliable. For air cargo carriers, these dynamic pricing models will help them identify the most profitable customer niches. For customers, this means more detailed pricing information.
Please note this doesn’t always mean lower prices. Customers in a more exclusive niche, for instance, are likely more willing to pay higher costs. Rather dynamic pricing allows for a more fair and transparent relationship between both parties. The mutual benefits cannot be underestimated.