Management Science Review

Machine learning (ML) and data analytics are rapidly changing today’s business landscape.  With the power of big data, Amazon can find products you like based on your purchase history, Netflix can make recommendations from a library of thousands of movies, and Uber can adjust prices instantly to ensure there is enough driver supply during peak hours.  ML’s success is mostly due to its ability to make accurate predictions, which is made possible by a combination of smart machine learning algorithms, massive datasets, and a lot of computational power.

However, while making good predictions are important, real-world business problems also require ML algorithms to make good decisions, which ultimately affect companies’ bottom line.  In applications such as displaying products, recommending movies and setting prices, an ML algorithm needs to make a sequence of decisions to interact with customers. 

For example, suppose you are an online fashion retailer who wants to use ML to set prices for men’s shirts. When a customer arrives to the website, a pricing algorithm suggests prices for all the shirt styles. The customer may choose to buy one shirt, multiple shirts of different styles, or leave without any purchase.  Because the customer’s choice tells us something about your prices (e.g., perhaps they are set too high) as well as something about seasonal and market trends, the ML algorithm must take the customer’s response into account. It is also important to notice that the customer’s response may change the behavior of ML algorithm, thus the prices received by future customers. The whole sequence can be thought of as a “closed loop feedback” process shown below.

The Caveat of Wrong Models

When an ML algorithm makes a decision, it often relies on some model, which is a way for computers to understand our world.  In practice, managers or business software often use incorrect models that don’t precisely reflect the real business problems. This issue is known as model misspecification.

There are several reasons for model misspecification: perhaps the manager does not fully understand the problem, or perhaps the software needs to use a simplifying model to save computation time.  Another reason for model misspecification is caused by the sheer size of data we have today: large-scale dataset may contain various kinds of information such as product characteristics, customer types, and economic conditions of the market. These features can affect customer choice in a complex way that we may never fully understand, so we inevitably use some wrong model.

The caveat of using a wrong model is that it will affect the “closed loop feedback” process we just showed. The new process is shown in the picture below. The computer algorithm will keep updating the wrong model, unaware that it may be far away from the truth.  In the meantime, of course, data are still generated by the true model (i.e., the real world). The discrepancy between the two models will cause bias in the learning process, which leads to poor decisions made by ML algorithms.

New closed loop process

Solving the Model Misspecification Challenge

In a recent paper appeared in Management Science, Dynamic learning and pricing with model misspecification, Nambiar, Simchi-Levi and Wang try to tackle the task of learning with misspecified models.  Even though sometimes a wrong model is unavoidable, the key question is: can an algorithm learn from data and get as close as possible to the true model?  If so, this implies that the decisions produced by the algorithm are almost as good as the optimal decision, the decisions that we would have made if we knew the true model.

Using dynamic pricing as problem context, the authors have proposed to inject random price shocks to their algorithm’s pricing decision. The intuition behind these price shocks is related to the concept of “instrumental variables” that are widely used in econometrics to correct for biased estimations.

Their analysis shows that random shocks does indeed exhibit strong numerical and theoretical performance. Moreover, the authors have also shown that the algorithm is versatile and can be adapted to a number of common business settings: for example, the seller may specify some business constraints on what prices are allowed to be offered.

The authors also demonstrated the real-world applicability of their algorithm through a case study in collaboration with Oracle Retail, involving a large fashion retail dataset.  Using historical data, they have performed tests gauging the performance of the random price shock algorithm.  The results are very promising and show that the algorithm can earn 8-20% more revenue than competing algorithms over a period of 35 weeks. 

Read the full article at

Nambiar M, Simchi-Levi D, Wang H (2019). Dynamic learning and pricing with model misspecification. Management Science 65(11):4980-5000.

The answer given by Rui Albuquerque, Yrjo Koskinen and Chendi Zhang in their paper “Corporate Social Responsibility and Firm Risk: Theory and Empirical Evidence,” is that CSR initiatives reduce firms’ systematic risk and increase valuations. This is because firms investing in CSR face relatively less price-elastic demand – that is demand for their goods does not fall that much with a price hike - so they can charge higher prices for their products and retain loftier profit margins. Customers become more loyal because they appreciate firms’ green credentials and socially responsible products, which are in line with their values and concerns about sustainability, so they are not so swayed by prices.

View Full Post »

The last fifteen years have seen a surge in patent litigation, driven predominantly by non-practicing entities (NPEs), firms that amass patents just to pursue license fees and/or litigation, rather than for the sake of producing commercial products. The impact of NPEs on innovation is a question of some controversy:

Proponents argue that NPEs are efficient financial intermediaries that police against well-funded firms that would otherwise infringe upon small inventors' intellectual property.
Opponents argue that NPEs simply raise the costs of innovation by wielding the threat and uncertainty of the legal process against otherwise innovative firms.

View Full Post »

When an earthquake strikes, its relatively mild P-waves travel through the earth faster than the powerful S-waves that cause the severe shaking. Taking advantage of this speed difference, usually just a few seconds, earthquake early warnings systems have alerted people so that they could take shelter.

View Full Post »

Learning consumer demand from early sales is a complex but a potentially rewarding practice for retailers of fashion products. With learning capability, a retailer can adjust prices during a sales season in response to sales observations; i.e., the retailer can benefit from the ability to act upon updated information. The potential value of demand learning under dynamic pricing could be significant and has been the preeminent motivation for retailers to acquire such capability. Nonetheless, in fashion retail environments, consumers may be cognizant of a retailer’s learn-and-price strategy; as a result, they may strategically adjust their purchase timing in anticipation of price changes driven by the dynamics of the retailer’s learning process. This gives rise to an important question: Is the adoption of learning technology truly beneficial for a retailer that sells fashion goods to strategic consumers? In the paper “Responsive Pricing of Fashion Products: The Effects of Demand Learning and Strategic Consumer Behavior”, Yossi Aviv, Mike Mingcheng Wei, and Fuqiang Zhang address this question.

View Full Post »

It has been argued, that the amazing expansion that many digital markets, and especially the market for mobile smartphone apps, have seen over the past decade was fueled by the possibility to collect previously unseen amounts of user data.

View Full Post »

In the book The Innovator’s Prescription, Harvard’s Clayton Christensen and colleagues have described hospitals as “some of the most managerially intractable institutions in the annals of capitalism”. While there are many factors that make hospitals difficult to manage, they argue that an operational dilemma lies at the heart of this managerial complexity: The tension between standardization and compliance on the one hand and creative autonomous solution finding on the other. These two operational modes require very different management styles and mixing them in the same organizational unit leads to dissonance and confusion. It is no coincidence that R&D and manufacturing are clearly separated in organizations that do both.

View Full Post »

Existing knowledge suggests that State-owned enterprises (SOEs) are managed inefficiently and influenced by inadequate political goals. This conventional thought, however, contradicts with the observation that state capitalism in China has been one of the most important contributors to its miracle economic growth in the past 4 decades. This contradiction is especially puzzling given that CEOs in Chinese SOEs are poorly incentivized using the traditional managerial incentives measures, i.e., monetary compensation.

View Full Post »

Service sectors are very labor intensive – they hire millions of workers, but generally suffer from low labor productivity (about three fourth of the productivity of industry sectors). Effective team management has the potential to serve as a powerful operational lever to improve labor productivity, not only because service workers often work in teams/groups, but also because effective teams often create value more than the sum of their individual parts.

View Full Post »