Overbooking in the Hotel Industry

Introduction
Over the past 50 years, revenue management has played a major role in many markets, including hospitality and tourism, expanding and growing. Defined as “sell[ing] the right inventory item to right customer at right time at right price” (Smith, Leimkuhler and Darrow 1992:8), revenue management uses a number of tools for maximising revenue, such as pricing, stock, marketing and distribution networks. It was initially developed to maximise the price of perishable products, the development of its concept and application coincided with its growth in the aviation, hotel, cruise and other tourism sectors. Revenue management research started in the airlines and gradually spread to other related tourism industries. The theory behind the topic has continued to evolve with innovations that integrate customer behaviour and relationship management; improving the accuracy of prediction and predictive pricing models; incorporating all revenue streams within the company; and focusing on gross operation (Milla and Shoemaker, 2007). Overbooking represents an important technique that extends revenue management for many service providers. Many companies in the service sector such as hotels systematically overbooking capacity of rooms in order to maximize their revenue. It can lead to customer denials, which then increases customer complaining behaviour, and decrease customer spending behaviour and overall satisfaction (Noone and Lee, 2010). The aim of this essay is to critically analyse overbooking in Hotel industry generally, and then by applying overbooking strategy of Marriot International revenue management evaluate the effectiveness of the strategy in the financial sustainable decisions.
Overbooking in tourism industry
Overbooking is a reaction to the fact that at the time the product is available, but consumers who order product for future delivery frequently fail to show up to collect. The ‘’no show’’ is often the choice of the consumer, however, ‘no shows’ may also be the product of operational factors: in the airline business, when a flight arrives late, passengers on that flight may ‘‘no show’’ for their connecting flights (Bell, 2010). Without overbooking, a large proportion of the sales for businesses that carry the risk of cancellations and no-shows will be decreased and their competitiveness will be adversely affected. Overbooking may result in overselling, however, means that the number of overbooked units may surpass the actual number of cancellations or no-shows (Rothstein, 1985). The possible benefits of overbooking strategies are subject to proper use, implementation and interpretation of the underlying overbooking models (Krawczyk et al., 2016). Despite the overbooking is a good strategy for revenue managers to maintain the upcoming no-shows or cancellations, companies could loose even more money if it’s not properly organised and forecasted. If the company’s revenue managers decide to overbook, there are four practiced approaches of overbooking they can choose from (Philips, 2005):

Get Help With Your Essay
If you need assistance with writing your essay, our professional essay writing service is here to help!
Essay Writing Service

Deterministic heuristic: Revenue manager often determines a booking limit (i.e the maximum number of rooms to be sold) based on the total capacity and estimated no-show levels. Specifically, it measures the booking by dividing the power by the historical display average. Accordingly, this approach is the easiest tool amongst the other approaches, which gives revenue manager a simple way to identify the limit for overbooking based on past experience and historical data (Philips, 2005).
Risk-based policy: seeks to reconcile the anticipated refused service costs with the revenue by accepting further reservations. This refused service expense is denoted and usually is higher than the resource price (Sulistio, Kim and Buyya, 2020). In other words, Revenue manager sets booking limit so that total revenue is maximised after deducting the overbooking expenses (gross refused service costs). This approach takes the overbooking costs into account, which makes it the best overbooking method in terms of revenue management. It also may take into account loss of customer trust in terms of the cost of alternative accommodation at a nearby property in case of overselling. Nevertheless, to take into account intangible costs like the loss of trust and satisfaction, the company should use complex mathematical patterns to measure each types of the costs.
Service-level policy: While policy above enhances the estimated revenue of the resource, customers who have been denied at the time of service appear to submit their jobs to other resources in the future and may lose some of these customers in the long term by using this policy. This approach may also add to the negative impact of overbooking on user satisfaction (Sulistio, Kim and Buyya, 2020). Revenue manager resolves the highest booking limit so incidents involving refused service will surpass the managerial perceptions. This approach is handful if the provider of a service wants to minimise the impact of refused service on the image of the client trust and the company.
Hybrid policy: This is the approach in which the risk-based limits are determined, but service-level requirements are limited (Dietz, Osborn and Sanli, 2012). In other words, the revenue manager determines the ideal overbooking cap for both the risk-based and service-level policies to optimise total revenue after cost deduction while reducing the maximum overbooking at the defined level of service. This approach would be the best for revenue management income, but challenging in customer’s satisfaction and trust when they are walked from the hotel. For instance, Marriott Bonvoy provides loyalty program for their members, and in case they are not accommodated they can be compensated for 100-200$ and around 90,000 their loyalty points and free stayover in the nearest hotel (Marriott International, 2019).
Hotel overbooking in practical terms
Unlike the tangible goods, hotel rooms are perishable, which means when a booking is unexpectedly cancelled, it is mostly impossible to get the client to occupy the hotel room in time. According to Dong and Ling (2015), if the hotel room gets unoccupied, hoteliers lose revenue. This distinctive aspect results in a zero salvage value for hotel rooms, since unoccupied rooms do not produce any income for hoteliers and can not be expressed for potential use in the inventory. Managing hotel rooms, on the other hand, involves high fixed costs but relatively low variable costs, leading to significant operating income for hoteliers. In particular, hotels suffer from last-minute cancellations and customer no-shows a significant loss of revenue. Hotels usually reduce this loss through implementing overbooking strategy to ensure full room occupancy and maintain a high occupancy rate. On the other hand, e-commerce increases the risk of cancellation, because consumers can easily compare hotel conditions and cancel their reservation (Koide and Ishii, 2004). In this case, taking overbooking effectively is significant for the hotel industries. Overbooking creates as much as 20% of the total revenue potential for hotels as a functional part of the revenue management framework.
However, if the no-shows are less than forecasted, hotels usually do some compensations to the affected customers, such as upgrade customers to a better room category (e.g. from a double room to suite). If the upgrades are exhausted, overbooking usually leads to walks (hotel refusing to accept the customers and moving them to different hotels) (DeKay et al., 2004). Most hotels use a suitable hierarchy of desirability concerning which guests can be walked. Members of frequent traveller services, regular business visitors, organisation members, conference planners and conventioners, regular airline parties, multi-night households, unaccompanied minors and single women are seldom walked. By comparison, visitors with single-night bookings and families on leisure travel tend to be targeted for walks, and first-come-first-served accommodation for late arrivals. According to Salomon, 2000, Although the traditional “best practise” used by hotel organisations is to provide the walked guest with free accommodation in a comparable hotel, transportation and telephone call, there is evidence that hotels will go beyond this standard and offer additional compensation in an attempt to placate the dissatisfied guest such as free night on a future stay, bonus reward program points or a cash amount.
Marriott International overbooking method
Marriot International hotels company has established a revenue management system that divides the various market segments (e.g. leisure or business travellers) using the scope of Marriott properties and offers a detailed understanding of the reservation behaviour, price sensitivity and stay trends of those segments mentioned above. This allows Marriott to optimise the room revenue by expanding sales of rooms. The system uses a security software, which is deeply integrated with the reservation programme of Marriott. It provides sophisticated recommendations for limiting inventory and also creates forecasts of arrival (demand). Revenue management programme offers all distribution outlets the same stock and rate information, so customers can find no diversity between the channels. It also makes suggestions for overbooking for each property by factoring in cancellations, stayovers and early departures. Eventually, the system enables revenue managers to maximise the revenue of each room sold on an adjusted price by demand, and enables Marriott to be able to cross sell its properties to some of its customers by allowing sales representatives on any Marriott property to access occupancy and room rate information (Siguaw and Enz, 1999).
The industry standard solution is to first set ideal booking limits for every stayover night in the planning stage, where the demand is combined across price categories covering every night stay. It is done through stochastic dynamic backward resource programming which is in phases the nights before stayover and whose states are present aggregate booking divided by rate class–duration–date of arrival. The element of decision for a stayover night is the upper limit on reservations. The stochastic components usually represent aggregate demand from one stage to another based on a forecast, the probability to cancel from stage to stage and the likelihood of no-show in the final stage. However, for each night prior to stay in all class – duration – arrival combination dates, Marriott maintains estimates of cancellation probability and they also hold estimates of the probability of no-shows (Marriott.gcs-web.com, 2018). Therefore, the optimal booking limit is determined one night before the first stay by testing out all possible booking limits and estimating expected costs. The expected cost is the overbooking fee multiplied by the average nightly total rate in the case of overbookings, which is only the average total nightly rate for underbookings, or the opportunity cost for unaccommodated room. Once the ideal booking limits for every possible state are determined on the night before the arrival, the maximum overbooking limits must be estimated two nights before arrival for every possible state. Finally, a list of all potential booking limits is determined in each case, with the estimated cost computed through backward recursion, taking into account all combinations of demand probabilities in the next 24 hours from the estimation of the reservation profile forecast, together with the likelihood of cancellation. The recursion moves one phase at a time, until the best booking limit for the current state is determined in the current phase (Pimentel, Aizezikali and Baker, 2018). This strategy could be considered as Hybrid policy, because the revenue manager sets the optimal booking limit in all of the cases as what is the demand of the room, the cancellation probability and the no-show probability as it is for risk-based policy. This could be very beneficial for revenue income because it shows the best price possible for the room, but on the other hand, it could be challenging in case of customer’s loyalty because they may be refused by the hotel and as compensation due to the hotel’s failure to accommodate the guest have to be moved to different hotel, which is characteristic for service-level approach. If the similar room in the nearest hotel is more expensive than the customer bought, hotel will lose more money. To maintain the customer’s trust, the hotel has to give some additional incentive to the customer as loyalty points in the loyalty program which can be redeemed as a discount in the next booking.
Possibilities and ideas for Marriot hotels revenue management strategy
Marriot’s full year 2017 results reflect year by year growth in the moderate but steady economic growth and brand demand on many markets around the world. In 2017, RevPAR (revenue per available room) rose by 3.1% to $115.02, ADR (average daily rate) increased 1.2 percent constantly to $157.12 and occupancy rate increased by 1.4% to 73.3% in contrast with 2016 (Marriott International, 2017). Comparison, Marriott’s full-year 2018 results reflect an increase because of the advantageous demand of their brands in various markets worldwide, and generally favourable economic conditions. Comparable to that in 2018, the worldwide RevPAR rose by 2.6 percent to $117,37, ADR rose by 2.0 percent to 160,27 dollars, while the occupancy of the rooms decreased by 0.1 percent points to 73.2 percent, compared to 2017 (Marriott.gcs-web.com, 2018). In the third quarter of 2019, their RevPAR increased by 1.5 percent, while their global RevPAR index rose by 210 basis points (McManus, 2019). As can be seen from above, the occupancy rate slightly decreased between the 2017 and 2018, which could have been caused by overbooking and refusing guests in the particular hotels, and then continuously loosing their trust and loyalty. For instance, Marriott should focus more on customer satisfaction by giving the guests more incentives like higher discounts in their future reservation in the hotel in case they are walked to a different hotel, which can generate more bookings in the future as well. The overbooking strategy is not very suitable for future financial decision making, as the company plans to increase the number of new properties and rooms in the Marriott hotels over the next years by around 280,000 by 2021 (Kim, 2019). They should also focus more on a past experience and historical data in case of overbooking, which would help the revenue manager to easily determine a booking limit of each stage.
Conclusion
The example above has underlined the fact, that when Marriott plans to expand their properties globally in the next couple years, there would be more overbooking increase and due to that, more guests will have to be walked and accommodated in different hotels and the company would lose more revenue. The overbooking strategy also have a lack of some important elements as dynamic demand forecasting which includes the short term and long term forecast. Therefore, the past cancellation rates and no-shows would be beneficial for revenue manager to simply set the booking limit and have also opportunity to forecast the probability of no-shows and cancellations in the future, which is characteristic for the deterministic heuristic policy.
References:

Bell, P. (2010). The concepts of revenue management: a tutorial. [ebook] London: Richard Ivey School of Business. Available at: http://www.ebscohost.com [Accessed 15 Jan. 2020].
DeKay, F., Yates, B. and Toh, R.S. (2004), “Non‐performance penalties in the hotel industry”, International Journal of Hospitality Management, Vol. 23 No. 3, pp. 273‐86.
Dietz, A., Osborn, N. and Sanli, T. (2012). Factoring Upgrades into Overbooking Decisions for Hotels and Casinos. [online] Support.sas.com. Available at: https://support.sas.com/resources/papers/proceedings12/384-2012.pdf [Accessed 19 Jan. 2020].
Dong, Y. and Ling, L. (2015). Hotel Overbooking and Cooperation with Third-Party Websites. Hefei: University of Science & Technology of China.
Koide, T. and Ishii, H. (2004). The hotel yield management with two types of room prices, overbooking and cancellations. [online] www.sciencedirect.com. Available at: https://www-sciencedirect-com.ezproxy.mdx.ac.uk/science/article/pii/S0925527304002609 [Accessed 15 Jan. 2020].
Marriott.gcs-web.com. (2018). Marriott International, Inc. 2018 Annual Report. [online] Available at: https://marriott.gcs-web.com/static-files/8799734e-b9e0-4e53-b194-7bd24a381118 [Accessed 22 Jan. 2020].
Marriott International. (2019). [online] Available at: https://www.marriott.com/loyalty/member-benefits/guarantee.mi [Accessed 23 Jan. 2020].
McManus, B. (2019). Marriott International Reports Third Quarter 2019 Results. [online] Marriott International Newscenter (US). Available at: https://news.marriott.com/news/2019/11/04/marriott-international-reports-third-quarter-2019-results [Accessed 23 Jan. 2020].
Milla, S., and S. Shoemaker 2007 Three Decades of Revenue Management: What’s Next? Journal of Revenue and Pricing Management 7:110-114
Noone, B. and Lee, C. (2010). Hotel overbooking: the effect of overcompensation on customers’ reactions to denied service. [online] https://journals.sagepub.com/. Available at: https://journals-sagepub-com.ezproxy.mdx.ac.uk/doi/abs/10.1177/1096348010382238 [Accessed 14 Jan. 2020].
Pimentel, V., Aizezikali, A. and Baker, T. (2018). An evaluation of the bid price and nested network revenue management allocation methods. Computers & Industrial Engineering, [online] 115, pp.100-108. Available at: https://www.sciencedirect.com/science/article/abs/pii/S0360835217305119 [Accessed 21 Jan. 2020].
Rothstein, M. (1985). Or and the airline overbooking problem. Estados Unidos: Institute for Operations Research and the Management Sciences.
Salomon, A. (2000). Overbooking: Hotels talk the talk to avoid the walk.. [online] Hotelinteractive.com. Available at: http://www.hotelinteractive.com/article.aspx?articleID=626 [Accessed 13 Jan. 2020].
Siguaw, J. and Enz, C. (1999). Best Practices in Information Technology. [online] www.sagepub.com. Available at: https://journals-sagepub-com.ezproxy.mdx.ac.uk/doi/abs/10.1177/001088049904000510 [Accessed 21 Jan. 2020].
Smith, B., J. Leimkuhler and R. Darrow 1992 Yield Management at American Airlines. Interfaces 22:8-31.
Sulistio, A., Kim, K. and Buyya, R. (2020). Managing Cancellations and No-shows of Reservations with Overbooking to Increase Resource Revenue. [ebook] Melbourne: The University of Melbourne, Australia. Available at: http://buyya.com [Accessed 18 Jan. 2020].
Uysal, M., Schwartz, Z. and Sirakaya-Turk, E. (n.d.). Management science in hospitality and tourism: Theory, Practice and Applications. Oakville, Canada: Apple Academic Press, pp.251-253.