The End of Bargaining
In rent-driven industries, humans will no longer set prices – algorithms will.
Dataism, the emerging philosophy created by the rise of big data, is changing the world at a dizzying pace. A main conviction of this philosophy is that all structures and systems can be viewed as data processing systems, and that the universe consists of data flows. As is explained by Yuval Noah Harari in his 2017 book, “Homo Deus,” Dataism posits that the value of any phenomenon or entity is determined by its contribution to data processing [1].
In this view, data is seen as only the first step in a long chain of activities. In fact, for as long as we’ve been around, humans have used biological data processing algorithms (e.g., gut feel, experience, intuition, judgment) to distill data into information, information into knowledge and knowledge into wisdom. However, in a world of increasing complexity, we are no longer able to rely on these natural mechanisms to process vast amounts of data into information, let alone knowledge into wisdom. Today, the work of data processing must be entrusted to electronic algorithms.
Revenue management (RM), the best data processor for pricing, uses electronic algorithms to consistently find the optimum price for a product or service under multiple market conditions. This consistency enables sellers to sustain long-term revenue growth, and buyers to be content with their purchase decisions. As data becomes ubiquitous, RM is becoming a sophisticated management discipline, upending entire industries by altering the way prices are set.
One of the latest industries affected by the rise of RM is rent-driven industries such as multifamily housing (apartment), self-storage, and senior living. Not long ago, pricing was done purely by a local manager’s gut feel. Management used manual, reactive, and archaic methods to set prices that rarely resulted in optimal revenue. Typically, an operator started with a budget and used occupancy levels as triggers to set rents. If prices were too low, the operator ended up with waiting lists of interested customers, but no available units. If prices were too high, the operator ended up with unsustainably low occupancies.
This rent-occupancy seesaw problem, together with outdated pricing practices, has created market demands and opportunities for RM systems to thrive. Today, more than half of the industry has already adapted RM and is transitioning from manual to automated pricing practices. In fact, the near-universal adoption of RM in rent-driven industries is an inevitable outcome – all noteworthy operators will either adapt to RM or end up being purchased by others who do [2].
A Short History of RM in the Rental Industry
The use of RM first emerged following the deregulation of the U.S. airline industry. American Airlines hit the jackpot in the 1980s with a new RM system and reported annual incremental revenue of $500 million. Adoption of RM subsequently revolutionized the entire travel and hospitality industries, and slowly disseminated into other commercial enterprises.
During the late 1990s, forward-looking real estate executives noticed the RM revolution in travel and hospitality and embarked on implementing similar RM systems for pricing apartments. Real estate rental operators face many challenges to maximize their profits. First among them is assuring that their units are priced correctly. In fact, no process is more foundational than the method an operator uses to make pricing decisions.
The importance of RM for real estate rental operators becomes evident if you listen to any earnings call of a publicly-traded company. Questions from investor analysts revolve around the company’s pricing policies, rent and occupancy levels, existing customer rent changes, and the speed at which new properties are leasing. These questions are not surprising; the assets already exist, fixed costs have been incurred, and operators’ success relies entirely on how much profit they can extract from their existing assets.
Real estate owner and operator Archstone-Smith was the first company to implement RM for pricing apartments during the late 1990s. They partnered with Talus Solutions (now JDA) to develop the first RM system. However, data quality was exceedingly poor and needed to be cleaned before it could become usable for modeling. At the time, I was part of the development team and, as a lead operations research (O.R.) scientist, was tasked to prototype, design, and validate its implementation. However, because the data was not in good shape, to get started, we had to make assumptions and simulated the data. In parallel, the team was manually cleansing, filling gaps, and preparing the actual data for modeling. I must admit that at times it felt like we were being thrown off a cliff while being asked to assemble an airplane on the way down.
The resulting RM system we designed was enormously successful. Archstone-Smith outperformed its peers during the economic slowdown of the early 2000s and attributed this success to its new RM program. Before long, in 2003, the second multifamily operator rolled out an RM solution. Other operators quickly followed suit, and by 2008, RM had become a mainstream business strategy. Today, hundreds of RM professionals use sophisticated methodologies to price millions of apartments daily.
One key takeaway was that, like any other scientific discipline, RM is an empirical endeavor. Fixing data issues is critical, but you don’t often have the luxury to wait for data to be perfect before you begin working. Modeling work may also uncover unforeseen data problems. Like crude oil, even data with poor quality might still contain a lot of valuable information to be explored.
In the 2010s, large self-storage operators followed in the footsteps of the multi-family housing industry, with most of the top 10 operators embracing the use of RM systems. The speed at which RM systems spread to other self-storage operators was much faster due to intense competition and higher fluctuations in rent, demand, and lengths of stay. In addition, the need for change management is minimal compared to other rent-driven industries. Today, the RM practice is considered a strategic imperative for self-storage operators of all sizes.
In 2011, Prorize, the company I now lead, partnered with Holiday Retirement to pilot a new RM system for the senior living industry. We conducted live pricing experiments with randomized A/B testing, extensively evaluated the algorithms, and confirmed that RM could generate more than a 9 percent revenue uplift. Holiday Retirement rolled out RM across its entire portfolio in 2014, subsequently achieving record levels of revenue growth.
Our groundbreaking work with Holiday Retirement won the prestigious Franz Edelman Award in 2017 [3]. This was remarkable since no RM system had received this recognition since American Airlines’ seminal RM work in 1991.
Other major senior living operators have evaluated RM algorithms, confirming similar revenue uplifts from randomized trials. However, the spread of RM has been slow. Price bargaining is still the dominant form of exchange coupled with misaligned salesforce compensation schemes based on volume. In some cases, salesforces do not work in the best interest of corporate objectives; rather they consistently give away margins to meet their personal sales objectives. Nonetheless, as the senior living industry continues to grow exponentially along with an aging population, RM will become unavoidable in this industry, too.
In summary, the use of RM in rent-driven industries is in different stages of adaption. If one considers the product lifecycle, RM technology is currently in a mature stage for the multifamily housing industry, a growth stage for the self-storage industry, and an introductory stage for the senior living industry.
The rest of this article presents three main lessons that could provide invaluable insights for future RM initiatives as it spreads into other industries and economies.
Transitioning from Bargaining to Fixed Pricing
For much of human history, bargaining was the dominant way to determine a price. This is hardly surprising as our natural instinct is to bargain anything we need to purchase, perhaps as an evolutionary legacy of the barter trade. Fixed or haggle-free pricing, where sellers unilaterally choose prices and/or discounts and do not negotiate with buyers, started to appear in various urban institutions during the 18th century [4]. Department stores were the first to use fixed pricing policies. While the transition from bargaining to fixed pricing has been taking place over the past 200 years, bargaining still survives in some businesses across the world.
The common practice is that headquarters create “asking prices” and grant local sales staff discretion to decide on the final rent based on local conditions or deal-specific information. The number one reason quoted for bargaining in the real estate rental industry is that everyone else is doing it. There is a fear that customers would abandon the operator who unilaterally imposes take-it-or-leave-it prices. In addition, allowing local staff to negotiate prices provides customer or market-specific information that could be exploited.
However, there are many downsides to bargaining. First, it favors customers who are most willing to haggle. When price bargaining is the norm, no one actually wins. Customers are always left feeling that if they only negotiated harder, they could have received better discounts, and end up permanently unsatisfied. Second, bargaining creates a significant gap between “asking price” and “actual price.” This often results in a suboptimal price due to the price gap exposed to human biases and judgments. Third, bargaining lacks the consistency, control, and superior efficiency that fixed pricing provides. Operators can use enterprise-wide data to accurately estimate customer preferences and promote fairness by stopping to compete with their customers over price. They should instead compete with rivals overvalue. Lastly, our extensive pricing experiments in multiple industries indicate that fixed pricing provides consistently higher revenues.
The increasing use of RM systems has enabled the multifamily housing and self-storage industries to completely eliminate bargaining. Manual pricing processes are also removed or minimized. Today, only the senior living industry is still in the early stages of RM adoption, and bargaining is still the dominant form of exchange. This is particularly unacceptable for senior housing where price haggling reduces customers’ positive anticipation of choosing good care and a well-suited home to a process that can feel like buying a used car.
As the digital age and RM take over, companies will move toward centralization, human feelings will be out of pricing decisions, and automated fixed pricing will continue to gain ground at the expense of bargaining and manual pricing processes. As a result, salespeople will stop selling on price, but rather sell on value.
Measuring RM Performance
An RM system sits at the heart of any business. It requires enormous investment and resources, and the pricing recommendations it produces can make or break the company. It is vital to have a yardstick that closely monitors the profit performance that an RM system delivers. Black boxes must turn into glass boxes with complete visibility. In addition, the RM system must have the ability to constantly assess itself and make self-correcting adjustments.
However, measuring the RM performance is rather challenging for rent-driven companies since the profitability of a customer depends on a lifetime of transactions. In fact, the main objective of an RM system is to maximize the net present value of each customer, which is extremely difficult to calculate with confidence due to high uncertainty in length of stay.
The “gold standard” of measuring RM performance is to compare alternatives with randomized trials. However, this is something only large companies can do prior to rolling out a new RM system. The reasons for this are simple; the sample size of smaller companies would not be big enough and, once fully deployed, conducting randomized trials is no longer feasible.
Instead, the makeup of a performance measurement module includes monitoring key performance indicators (KPIs) and predicting demand and revenue contributions. Sample KPIs include rent, occupancy, revenue, and user compliance indices. Each KPI is continuously evaluated using multiple statistics such as average, trend, dispersions, percentiles, skewness, and kurtosis. Multiple KPIs could be modeled or reviewed simultaneously over time to measure their influence on each other.
An operator rarely observes overnight financial success from an RM program as it is necessary to have an adequate sample size of new customers with optimized prices. As a result, it is most critical that senior executives sustain their commitment to the RM initiative over a sufficient period of time, which leads us to the third and final lesson.
Change Management and Leadership Buy-in
When successfully implemented, an RM solution is truly transformative. It is never a “plug and play” solution. It requires enterprise-wide changes in business processes, mindsets, culture, work roles, responsibilities, and how business operates. More often than not, cultural warfare arises with “old guards” as a result of the changes that come with a new RM solution. There is no technical solution to emotional or cultural problems. In our experience, success starts from the top; unless the corporate executive team makes it a topmost priority, an RM program will be restricted, rerouted, and eventually terminated at any perceived speed bump.
In fact, the number one competitor of an RM system is intuition, gut-based decision-making, and human experience. These are popular because they are cheap and readily available; they represent our biological impulses. However, they are not superior to fact-based algorithms; the capacity of RM far exceeds that of the human brain in processing a large amount of data. We, humans, are lousy forecasters and are equipped with unintentional biases. For example, we give more weight to recent events although earlier events might be more relevant. Our intuition does not account for underlying changes, and we tend to be risk-averse.
In the real estate rental industry, there are many “old-school” business processes. These include focusing on occupancy rather than revenue, using budgets for pricing and measuring success, and having volume-based sales bonuses. For RM to be successful, these must be eliminated or at the very least enhanced.
An effective RM program does not guarantee long-term success. The success of any business also relies on good management, marketing, communication, and a host of other factors. RM must integrate into the existing processes, like HR training, certification, and reward metrics based on system compliance.
The personal support and commitment of the CEO is also critical, albeit not enough by itself. The implementation requires a long-standing commitment of key executives to favor fact-based decisions over intuitive conclusions. From day one, the CEO needs to send a powerful message to the entire organization and ask that all subordinates support the new data-driven pricing direction.
Finally, key stakeholders should be regularly informed on the performance of RM system and be provided forums where feedback can be given. Whether through periodic meetings, newsletters, emails, feedback sessions, or any other means, it is essential to regularly communicate what is working, what is not working, and what the future plans are.
What Lies Ahead
Now that “Dataism” has arrived for the rent-driven industries, there is no going back. The relentless flow of data will continue to increase, giving us ever greater insight into human nature, better predictions, and, ultimately, accurate pricing.
Data has the freedom and capacity to tell us the truth. Those companies that connect data and use data-driven pricing for their advantage will overrun their respective industries. Manual pricing processes and bargaining in the real estate rental business will soon be found only in history books. Employment of RM will separate the right and wrong, and will surely extend, deepen and spread to the entire real estate rental industry.
The business imperative is strong. Potential business results from optimized pricing are too great to be left to gut feelings and human experience. The importance of RM as a business discipline will increase, and sophisticated tools will continue to be developed for new industries. The authority of pricing will finally shift from humans to data-driven computer algorithms. As data quality, quantity, flows, and connections improve, RM will virtually free humans from having to find the right price or think about whether we’ve paid too much or too little. We will instead focus on the value we give or receive from any exchange – a much more worthwhile endeavor.
Notes and References
1. For example, American, United, and Delta Airlines embraced RM, thereby achieving significant growth in revenue, and ended up purchasing giants like Pan Am and TWA, which were extremely slow to adapt to RM.
2. Harari, N. H., 2017, “Homo Deus, A Brief History of Tomorrow,” Harper Collins Publishers.
3. Kuyumcu, A., U. Yildirim, A. Hyde, S. Shanaberger, K. Hsiao, S. Donahoe, S. Wu, M. Murray, M. B. Maron, 2018, “Revenue Management Delivers Significant Revenue Lift for Holiday Retirement,” Interfaces, Vol. 48, No. 1, pp. 7-23.
4. Phillips, R. L., 2012, “Why are prices set the way they are?” Özalp Özer, R. Phillips, eds., “The Oxford Handbook of Pricing Management,” Oxford University Press, Oxford, U.K.
Source: ORMS Today
