Is Price Control by Humans?
I bet everyone who clicks on this article finds themselves involved in the price optimization problem in mathematics, we once were studying in math class or more specific calculus class and asked “what is the use of this optimization problem?”, we were curious, as we seek the application and empirical study from the learning process where we had been suffered from the unknown use of the thing we tried to achieve. If you are involved in Computer science, economics, applied mathematics, engineer degree, or the mathematical application toward the real world you get a pretty much chance facing this problem more than others. I have the answer you once sought to find by providing evidence of the price optimization application.
What is price optimization? Its definition from Wikipedia is…
“ Price optimization is the use of mathematical analysis by a company to determine how customers will respond to different prices for its products and services through different channels. It is also used to determine the prices that the company determines will best meet its objectives such as maximizing operating profit. The data used in price optimization can include survey data, operating costs, inventories, and historic prices & sales. Price optimization practice has been implemented in industries including retail, banking, airlines, casinos, hotels, car rental, cruise lines, and insurance industries.” …Wikipedia…
To put it simply, it is the price which is the maximum it should be charge according to the demand, supply, and the externality which can affect the profitability of the company, the price can be high if the profit can be an increase from the increase in price and low if it is otherwise. There are many factors that affect profitability( i.e. the number of firms, substitute products, brand loyalty, quality and etc.) and it is seemingly impossible that humans can control the price to its best potential result, assume the math is correct, that’s right! someone which more caliber has to step in and take this job, so humans create them, the invention of the price optimization software. It is a simple process, you can try in Microsoft Excel, given that you have data you can optimize it manually with machine computation power but recently, the era of the data-centric create a more powerful technology than just a computation power but the ability to compute and operate with dynamic self-operation while improve itself by an increase in training data or so-called machine learning and Artificial intelligence that create software of algorithm pricing.
Why we should concern?
First of all, everyone has to admit pricing is one of the most crucial pillars in the business world. A lot of tasks we had to perform for many decades had been replaced by the better repetitive and brute force species, the robot. Those are the tasks where they require brute force or require pure computation power, the introduction of the MACHINE which will replace the job is simply because MACHINE can achieve what humans can’t in those tasks. The algorithm pricing is the software that serves the order by maximizing the given objective( exposure, profit, market share and etc. ), this introduction of algorithm pricing is changing the whole commerce in many industries, mostly e-commerce and data-driven business. It has been for a long time ago that the businessman charges the price according to their knowledge, some charge from guessing between a certain range, some imitate the competitor. There are many pricing strategies out there but not a single one choose the price that optimization throughout time according to all relevant factors that updated as that factor change, price normally, that handled by human, changes period by period with the stability, but if AI is the one who controls the price, it will change as the information its receive updated the knowledge to the best outcome. Imagine that businessman had charge you a high price, but that high price you are still willing to pay, if AI charge that price with respect to profit maximization price, the outcome would be they trying to rip every piece of benefits it can gain and making sure people still buy it. It is not it is influenced to the daily business life at a slow pace but now many companies and start-ups had found the price optimization software company and compete which each other to provide a better algorithm pricing at better quality and lower price to the level that any small company can buy. For example Pricefx for price competition, IBM Omnichannel intelligence pricing specialize in pricing and promotion and, IntelligenceNode for e-commerce business and Technology.
How Pricing Algorithm work
One empirical algorithm pricing that monopolizes its platform is Airline flight tickets algorithm pricing. The airline industry is one of the most innovative adaptive industries, but this innovation does not help the consumer for the better, faster, or cheaper service, it helps the airline achieve the impossible task for human, pricing strategy. It works to serve the objective, maximize profit from the demographics and available data.
First, it defines the price range from the airfare class such as business, economics, and premium. The more privileged class the more money its charges to the consumer, this is basic human still able to perform, the second step is selling according to profit maximization. To achieve maximum profit, AI has no need to charge one consumer at the same price as other consumers, it keeps the number of available seats. When the available seat starts to drop or scarce, it charges a higher price, for every ticket bought, the price of other tickets rise and if you try to buy a lot of ticket at one time, all of your tickets are averagely charged higher than buying one ticket. Not only its charge price discrimination, the notion that charging a different price to a different consumer with the same product is provided. Ok, this might arguably that they might cost more if the consumer buys the ticket the closer to the day airline departure, it might increase the management, ordering and preparing food, human resource, and asset allocation or another argument as the closer the fight date the higher the demand. The third one is optional refundable add-on price, it plays with the uncertainty preference of the consumer but a still a minor one, the fourth one is much more clever, a golden combination of the profit is the price and cost, thus what it did is restrict the available seats so that it can both charge the high price of monopoly and reserve the fuel for the connecting route, for example, you travel from London to Miami at $1000 the which Miami is connected route to Bogota, but the AI price the ticket from London to Miami more than one trip direct to Bogota, even though the time travel is lesser. What it did is to discourage people from reserve those seats and keeping that seat to lower the fuel usage and restrict the output it maximizes profit while at the same time keep the seat available so people from Miami can reserve the seat support with the prediction analysis they perform, the amount it should reserve for a future seat that can jack up the price to maximize profit, for further read.
Not only the airline price ticket is controlled by the AI, but the e-commerce platform, Retail store that uses ePayment, Supply Chain industry, Insurance, and Healthcare. Not only it can perform price optimization to the level that is far superior to human-being, but it can also use the tactic of collusion or corporate pricing. In economics, corporate pricing is the price where the firm decides to corporate and lessen the competition in the market so that all the firms can enjoy the high price with no consideration of price competition or price war. The research from Calvano, Calzolari, Denicolò, and Pastorello run experiments with AI agents powered by reinforcement learning in controlled environments (computer simulations) and find the uncommunicative AI agent perform the tactic collusion. Even though there are still no evidence support of tactic collusion in the real world or we still lack the ability to detect it.
Can algorithm pricing be good for society?
All of these are controlling be algorithm pricing and the antitrust or the Federal Trade Commission still not able to find the proper way to against the technology as they cannot simply stop the technology if they don’t know the consequence of the regulation. It looks like I have said the bad stuff about the Algorithm pricing, the algorithm is not a bad machine that should be destroyed or keep in inventory or the manipulation of an evil action, it is tools which depends on the use. as we know it solves the optimization toward the objective function which I had referred to the bad side of the story, the profit side so if I told only the side that the technology is bad by injecting the greedy objective the whole society will be greedy pricing, but if the algorithm can be regulated using the social welfare function or the function which in economics stand for the social benefit of the whole population in the market rather than maximizing the private benefit but the social benefit. But who would like to maximize their profit by restricting to the social benefits, here is not rational for the businessman to do so but if the government regulation to fixed the level of profit maximization or they create AI that can detect any collusion or the monopoly abusive pricing and punish that could be a solution for controlling the Algorithm pricing.
Originally published at https://www.linkedin.com.
