Did you search for a product online in the morning and check back in the evening to see that the price has changed? In that case, you may have been subject to the retailer's pricing algorithm.
When determining the price of a product, marketers traditionally take into account their value to the buyer and the cost that a similar product costs, and determine whether potential buyers are sensitive to price changes. In today's technology-driven market, things have changed. Price algorithms usually perform these activities and set the price of products in the digital environment. In addition, these algorithms can actually work together in a way that is bad for the consumers.
Originally, online shopping was seen as an advantage for consumers because it allowed them to easily compare prices. The resulting increase in competition (along with the growing number of retailers) would also lower prices. The so-called revenue management pricing systems have allowed online merchants to use market data to predict demand and determine prices to maximize profits.
These systems have been extremely popular in the hotel and tourism industry, especially as hotels have fixed costs, perishable inventory (food that must be consumed before it runs out), and fluctuating demand. In most cases, hotels with the revenue management system can quickly and accurately calculate ideal room rates using sophisticated algorithms, past performance data, and current market data. The room rates can then be easily adapted wherever they are advertised.
These income tax systems have led to the term "dynamic pricing model". This refers to the ability of online providers to immediately change the price of goods or services to the slightest change in supply and demand, whether it be an unpopular product in a full warehouse or an uber Ride during a nocturnal wave. Accordingly, today's consumers are more content with the idea that online prices can and do fluctuate not just at the time of sale, but several times over the course of a single day.
However, new algorithmic pricing programs are becoming more complex than the original revenue management systems due to developments in artificial intelligence. Humans still played an important role in revenue management systems by analyzing the data collected and making the final decision on prices. Algorithmic pricing systems, however, work largely by themselves.
Just as in-home language assistants such as Amazon Echo experience their users over time and change their way of working, algorithmic pricing programs learn from market experience. 1
The algorithms examine the activity of online shops to determine the economic dynamics of the market (such as prices for products, normal consumption patterns, supply and demand levels). But you can also unintentionally "talk" to other pricing programs by constantly monitoring the price points of other sellers to see what works in the marketplace.
These algorithms are not necessarily programmed to monitor other algorithms in this way. However, they learn that it is best to achieve their profit maximization goal. This leads to an unintentional price agreement, in which the prices are within a very narrow limit. If a company raises prices, the competitor systems will respond immediately by raising their systems and creating a common, non-competitive market.
Monitoring competitors' prices and responding to price changes is a normal and legal activity for businesses. Algorithmic pricing systems, however, can go one step further by setting prices above what they would otherwise be in a competitive marketplace, as they all work in the same way to maximize profits.
This may be good from a business perspective A problem for consumers who have to pay the same everywhere, even if prices are lower. Non-competitive markets also lead to less innovation, lower productivity and ultimately lower economic growth.
What can we do?
This raises an interesting question. If programmers (unintentionally) could not prevent this collusion, what should happen? In most countries, tacit collusion (where companies do not communicate directly with each other) is not currently considered illegal activity.
However, companies and their developers could still be held accountable because these algorithms are programmed by humans and have the ability to learn how to communicate with competitor algorithms and share information. The European Commission has warned that the widespread use of price algorithms in e-commerce could lead to artificially high prices across the market, and that the software should be built in such a way that it does not allow collaboration.
However, as long as the algorithms are programmed to maximize their profits and are able to do so independently, programmers may not be able to overcome those agreements. Despite some limitations, the algorithms may well find ways to overcome them as they look for new ways to reach their destination.
Attempting to control the market environment to prevent deliberate price monitoring or market transparency will undoubtedly lead to further questions and new problems. It is against this background that we need to better understand this kind of machine learning and its capabilities before introducing new regulations.
This article was again published by The Conversation by Graeme McLean, a lecturer in marketing at the University of Strathclyde under a Creative Commons license. Read the original article.