W-Moszczynski Data science
Fixed prices for goods and services were introduced during the industrial revolution in the late 19th century. Before that, each shopkeeper negotiated prices with buyers, judging the wealth of the customer by appearance and behavior to offer appropriate prices.
With industrial development, markets became flooded with cheap products sold in mass in numerous stores and stalls. The increased volume of goods and shortened transaction times required more skilled sellers who could set prices. However, such specialists were scarce in the labor market, and the potential for malfeasance with fluctuating prices became a concern. Thus, the traditional haggling system quickly proved inefficient, and sales took on the form we know today: a single price for everyone.
Variable Pricing as a Source of Competitive Advantage
Fixed prices are not universally applied today. There are many sectors where prices change dynamically, reflecting market conditions. We recall that price is the optimal value at which a buyer agrees to buy, and a seller agrees to sell. Prices fluctuate on commodity and capital markets. The hotel booking market, airline ticketing, and taxi services (e.g., Uber) change constantly. Similarly, dynamic pricing occurs on the shipping market for sea, rail, and road transport. According to McKinsey, logistics companies that adjust prices can increase revenue by 2 to 4%, resulting in a 30-60% boost in operational profit. Their report suggests that a variable pricing strategy is the most profitable direction for development. Other online sales sectors have drawn similar conclusions: price is a significant area for increasing competitiveness. This doesn’t refer to merely lowering prices or implementing random discounts or loyalty programs. Competition in variable pricing primarily involves offering effective, intelligent prices based on in-depth econometric analyses—smart pricing dynamically calculated from market behavior and economic conditions. To compete in this area, companies need to invest in analytical capabilities and delve into machine learning algorithms.
Building an Organization Capable of Competing with Smart Pricing
Until recently, the approach to building a pricing-competent organization involved prioritizing strategic goals and developing relevant skills among sales staff. McKinsey’s 2018 report emphasized establishing local pricing offices, standardizing market validation processes, and fostering dynamic communication with clients as essential elements.
However, after the pandemic-accelerated digitization and the development of artificial intelligence, we can view these recommendations as somewhat outdated. Prioritizing goals and proper organizational structure remain essential, but the role of specialized sales personnel has diminished.
Replacing People with Robots
Just as prices were standardized in the 19th century due to a lack of skilled sales personnel, modern intelligent pricing cannot rely on specialized professionals. Such an approach is simply inefficient.
The world has long sought to simplify processes to avoid the unstable and costly human factor. For some time, advanced systems of automatic process control, commonly known as artificial intelligence, have been taking over human roles. These „robots” operate around the clock, know multiple languages, and are not subject to labor protection laws. Although it sounds futuristic, I stayed at a hotel in Tallinn last year where check-in and check-out were handled by a tablet, with the entire establishment managed remotely from Helsinki. In intelligent pricing for customers, automation is not about replacing people but creating mechanisms that exceed human capabilities. According to this approach, the worker’s role is merely to receive information generated by advanced AI algorithms.
It is an exaggeration to call a simple reception tablet or a voice information portal “artificial intelligence.” Yet, there are relatively simple mathematical algorithms that mimic human decisions, often surpassing human choices. In most cases, these are optimal decisions—the best possible in given conditions. Therefore, the term “artificial intelligence” is somewhat justified for more advanced automation systems.
Mechanisms for Dynamic Smart Pricing
Below are the primary methods of dynamic smart pricing, which can be applied simultaneously.
Bilateral Rule-Based System
Rule-Based pricing is the simplest and probably most widespread dynamic pricing system for goods and services. It appears when a customer configures a range of goods or services while shopping. For example, additional fees may apply to a basic airline ticket for priority boarding, extra baggage, or reservation flexibility. This is not intelligent pricing but rather a method of adding costs for additional options. The rule-based system is rigid and cannot respond to external changes.
Externally Referenced Rule-Based System
There are automated pricing rules that respond to market situations. In the airline ticket reservation example, a rule-based system could adjust prices based on the number of seats available on a flight. A rule linked to cargo space occupancy in sea containers is another example. If the cargo space is 50% occupied, the per cubic meter price is low, possibly at the average minimum price of the previous quarter for that voyage. To adjust this price to market value, the algorithm could offer a 10% discount compared to similar prices for similar routes on the current shipping exchange. Depending on space occupancy, the price changes in relation to market prices. When occupancy reaches 75%, the system suggests a price set at 90% of the market average for comparable transport. Thus, prices are continuously updated based on market rates with flexible rules relative to container space occupancy.
Customer Segmentation-Based Dynamic Pricing
In the past, respected shopkeepers knew their customers, met their needs, and supplied appropriate goods, each with specific pricing. In some sense, those days are returning. In online sales, customers are identifiable, having their own login and ID numbers, and transactions are logged in registers, allowing each customer to be characterized and grouped. Effective clustering requires identifying key characteristics, such as purchase frequency and transaction value. It’s also essential to determine how long since a customer’s last activity. This approach is known as RFM (recency, frequency, monetary). From these attributes, classes can be created, such as „spends a lot,” „spends moderately,” or „spends little.” Using Apriori or K-means clustering, customers can be grouped and assigned to classes. Based on sales strategy, the system can offer specific discounts or higher prices.
Key Performance Indicators (KPI)
Prices can be configured based on defined performance indicators. A business can set limits for price adjustments. In this approach, prices become part of equations determining profit, costs, and asset utilization. Most companies have defined performance indicators guiding priorities and actions, used to evaluate work efficiency and project effectiveness. KPI indicators can serve as additional constraints in an automated pricing system, leading to optimized outcomes.
Operational Research Algorithms
Companies can set goals such as maximizing profits or minimizing customer churn risk, while defining economic constraints like driver hours, storage space, or raw material quantities. In optimization equations, a goal function is defined, such as cost minimization, work hour maximization, or customer growth. The operational research algorithm proposes an optimal solution, where price is a variable in achieving the goal function. Dynamic pricing based on operational research algorithms is among the most advanced analytical methods, requiring a well-organized management system.
Market Forecasts
This method requires historical data on market behavior over time. For instance, ice cream sales vary between summer and fall, with prices adjusting based on demand and product availability. Market data is available ex-post, i.e., post-sale. To create intelligent, dynamic pricing, prices are forecasted based on historical data, using models from machine learning. For example, a model predicting ice cream demand based on temperature and rainfall from previous years also considers variables like location, competition, and past advertising effectiveness. In market forecasting, regression models based on Thomas Bayes’ conditional probability or decision tree models like Random Forest Regressor can be applied. Recurrent neural networks can be used in deep learning for market forecasts, paving the way for intelligent pricing.
Dynamic Pricing Based on Demand Elasticity
Demand elasticity refers to customer sensitivity to price changes. Some sectors experience fixed demand where customers purchase regardless of price, while in others, prices align with perfect market competition. For dynamic pricing algorithms, specific historical data is needed, accounting for sales and frequency of purchase cancellation due to price changes, such as those available on commodity exchanges.
Summary
Dynamic pricing has always been a source of potential benefits, strengthening customer relations and providing additional income. However, this form of transaction has evolved into mass sales with fixed prices. Today’s growth in online sales and the development of advanced data processing technologies allow for a return to dynamic, intelligent pricing.
Consulting firm analyses indicate that competition based on dynamic, intelligent pricing will be a primary competitive advantage in select industries in the future.
Wojciech Moszczyński – A graduate of the Department of Econometrics and Statistics at Nicolaus Copernicus University in Toruń, specializing in econometrics, data science, and management accounting. He focuses on optimizing production and logistics processes and is involved in promoting econometrics and data science in business environments.
At Global FCL, we are dedicated to delivering exceptional service and expertise to our clients.
As a leading U.S.-based logistics provider, we are proud of our strong reputation for
excellence in the industry. Backed by years of experience,
we offer reliable and efficient commercial freight transportation solutions
customized to meet your specific needs.