The Uniqueness of Revenue Management Approaches in Nontraditional Settings: The Case of the Golf Industry

Abstract

This study focuses on two demand and supply characteristics that may affect the transferability of revenue management (RM) practices from traditional (e.g., hotels) to nontraditional (e.g., golf, restaurants, entertainment venues) RM settings. Consumption within many nontraditional RM settings is largely discretionary in nature, with the potential to affect how demand and price should be managed across the booking horizon. Equally, operators are often challenged with a high degree of time-based inventory complexity, which may require that price and inventoried demand are managed at a greater level of granularity than traditional RM applications dictate. Using longitudinal golf reservations data, we found that superior revenue performance was associated with capturing a higher proportion of demand early in the booking horizon, rather than protecting inventory at higher prices for late bookers. Competitive price positioning in which price was higher than the competition during within-day peak-demand tee times also shaped revenue gains. Similarly, conversion management was found to be most critical during within-day peak demand periods. These findings suggest that traditional RM strategies may not apply in nontraditional RM settings where one or both of the demand and supply characteristics of interest is present. The implications of these findings for practitioners are explored.

Keywords

revenue management discretionary purchase inventory complexity booking horizon price positioning inventoried demand management

Revenue management (RM) practice has its roots in the airline, hotel, and car rental industries. Evidence suggests that successful application of RM principles in these contexts results in revenue increases of 2% to 5% (Talluri & van Ryzin, 2004). The demonstrated performance effects of RM principles in these “traditional” settings have motivated academics to encourage the adoption of similar RM practices within other tourism-related environments, including restaurants, casinos, spas, parks, entertainment venues, cruise lines, and golf courses (see, e.g., Ayvaz-Cavdaroglu, Gauri, & Webster, 2017; Boo & Kim, 2010; Chen, Tsai, & Chen McCain, 2012; Kimes, Chase, Choi, Lee, & Ngonzi, 1998; Noone, Kimes, Mattila, & Wirtz, 2009; Schwartz, Stewart, & Backlund, 2012). Researchers have argued for expanding the application of RM to these environments—hereafter referred to as nontraditional RM settings—based on the characteristics of relatively fixed capacity, perishable inventory, reservations made in advance, variable demand, segmentable markets, and a relatively high fixed–low variable cost structure (Kimes & Schruben, 2002; Li, Miao, & Wang, 2014). However, many of these nontraditional settings also exhibit distinct demand and supply characteristics that are not present in the traditional RM environment.

From a demand perspective, traditional RM applications are designed to manage both discretionary and nondiscretionary demand. Nondiscretionary demand consumers typically cannot commit to a purchase early in the booking horizon. For example, business travelers may be uncertain about the timing of a business meeting, or a need may arise to attend a meeting at short notice. Thus, they are willing to pay a higher price to book their airline seat or hotel later in the booking horizon. In contrast, discretionary demand consumers tend to purchase earlier in the booking horizon. Price-sensitive leisure travelers, for example, are willing to commit to an early hotel or airline purchase to secure a lower price. Thus, the goal in the traditional RM setting is to manage demand such that the firm builds a base from early booking, price-sensitive consumers, and sets aside sufficient inventory for late booking, nondiscretionary demand consumers. In contrast, in many tourism-related nontraditional RM settings demand is almost exclusively discretionary in nature (e.g., golf courses, performing arts, and spectator sports), and therefore may require that operators manage demand and price across the booking horizon in a manner that does not fit with traditional RM practice. On the supply side, operators in tourism-related nontraditional RM settings are often challenged with a degree of time-based inventory complexity that is not encountered in traditional RM contexts. For example, hotel room and rental car inventory is typically sold for a 24-hour period. In contrast, nontraditional RM setting such as golf courses and restaurants sell inventory at significantly smaller time intervals within a given day. Golf tee times, for instance, are often sold at 15-minute intervals. This degree of time-based supply complexity suggests that, to optimize revenue, operators need to be able to identify within-day peak and off-peak demand patterns such that pricing and capacity management strategies can be deployed across short time intervals throughout the day.

To better understand how the distinct demand and supply characteristics of nontraditional RM settings may influence RM practice, and ultimately revenue performance, this study empirically examines, in the context of the golf industry, the effects of three key aspects of RM practice on revenue performance: (1) demand management across the booking horizon, (2) competitive price positioning, and (3) inventoried demand management. Since the golf industry represents a nontraditional RM setting in which both the demand and supply characteristics of interest in this study are exhibited, it provides the opportunity to fully explore their potential impact on RM performance. In pursing this research, the overarching goal is to move the literature beyond a prescriptive approach to RM that presumes universal RM dynamics across tourism sectors. In addition, we seek to provide practitioners with insights to guide RM efforts in tourism settings where the demand and/or supply characteristics of interest prevail.

Literature Review

RM in Nontraditional RM Settings

Prior research in the domain of RM has been dominated by a focus on traditional RM applications (i.e., airlines, hotels, and car rental) in terms of systems development, performance measurement, and consumers’ responses to RM practices (see, e.g., Dolasinski, Roberts, & Zheng, 2019; Karande & Magnini, 2011; Lindenmeier & Tscheulin, 2008; Noone & Lee, 2011; Pekgün et al., 2013; Schwartz & Chen, 2012; Tanford, Erdem, & Baloglu, 2011; Wittman & Belobaba, 2017). However, there has been a growing interest in the application of RM across a variety of tourism-related settings, including restaurants, theme parks, golf courses, spas, cruise lines, and entertainment venues (e.g., Li et al., 2014; Noone, Enz, & Glassmire, 2017; Noone et al., 2009). The application of RM in these nontraditional RM settings is predicated on the presence of a number of characteristics that are present within traditional RM environments: a relatively fixed capacity, perishable inventory, reservations made in advance, variable demand, segmentable markets, and a relatively high fixed–low variable cost structure (Kimes & Schruben, 2002). However, nontraditional RM settings can also exhibit characteristics that are not present in traditional RM settings but must be addressed for effective RM. Differences in the manner in which time is sold across traditional and nontraditional RM settings have arguably received the most attention in the literature. While services are sold to the consumer for a fixed amount of time in traditional RM settings (e.g., a one-night stay at a hotel), time is sold as an event in some nontraditional RM environments (e.g., a restaurant meal or a round of golf). In the latter instance, the customer largely controls the duration of the service experience, prompting researchers to investigate the implications of such circumstances for RM practice. For example, Noone et al. (2009) examined the role of pace in restaurant RM, while Tiger and Howard (2007) proposed a simulation-based decision support system to increase the pace of play on the golf course.

To our knowledge, two fundamental characteristics of many nontraditional RM settings—the largely discretionary nature of demand and a high degree of time-based inventory complexity—have received little attention in the literature. Yet these demand and supply characteristics, whether present individually or together in a given environment, may require a more nuanced approach to RM application than traditional RM dictates. To better understand how these two characteristics may influence RM practice, and associated revenue performance, this research focuses on their potential impact on three key aspects of RM practice: demand management across the booking horizon, price positioning, and inventoried demand management.

Discretionary Purchase and Demand Management Across the Booking Horizon

Consumption across many nontraditional RM settings can be regarded as largely discretionary in nature (e.g., golf, performing arts, and spectator events). In this context, consumer demand can be characterized in terms of both inventory and price sensitivity.

Inventory sensitivity refers to consumers’ sensitivity to inventory timing or location. In golf, for example, inventory sensitivity refers to consumers’ sensitivity to starting tee times (e.g., a 7 a.m. tee time may be significantly more desirable to a tee time later in the morning). In the case of spectator sports, on the other hand, inventory sensitivity refers to consumers’ sensitivity to seat location (e.g., a seat at the 50-yard line at an American football stadium is typically considered as more desirable than a seat in the end zone). Thus, inventory-sensitive consumers are those who highly value access to the “best” inventory, and, as such, they are willing to commit to a reservation early in the booking horizon to guarantee access to that inventory. By virtue of their inventory sensitivity, inventory-sensitive consumers can be expected to be somewhat price insensitive, willing to pay a higher price to secure their preferred inventory. Consumers who are less inventory sensitive, on the other hand, do not value prime inventory as much as those in the inventory-sensitive segment. As a result, it can be expected that they will not be as motivated as inventory-sensitive consumers to book early. Rather, given the discretionary nature of the purchase, they are more likely to display a higher degree of price sensitivity than inventory sensitivity, and are likely to purchase later in the booking horizon by virtue of that inventory insensitivity. Indeed, there is support for this type of booking behavior in the venue context, where it has been demonstrated that high-priced event tickets (attached to the “best” seating locations within a given event venue) are purchased early in the booking horizon by inventory-sensitive customers who are willing to pay a premium for their preferred seating location. Low-priced tickets, associated with less desirable seating locations, tend to sell in greater volume as the performance approaches (Moe & Fader, 2009). This pattern in booking behavior resembles that for high-fashion clothing, where early adopters who have a greater desire for, and by extension place a higher value on, high-fashion clothing than later adopters, are typically willing to pay a premium to get access to the clothing when it first comes onto the market (Desiraju & Shugan, 1999).

A pattern wherein the “first” customers are willing to pay the highest price is not characteristic of the traditional RM setting where prices tend to peak close to the time of use of a given service (Anderson & Wilson, 2003). This traditional RM approach to pricing pattern accommodates nondiscretionary demand consumers who cannot commit to a purchase early in the booking horizon, and thus are willing to pay a higher price to book their airline seat or hotel later in the booking horizon (e.g., business travelers). In contrast, discretionary demand consumers (e.g., price-sensitive leisure travelers) are willing to commit to an early hotel or airline purchase to secure a lower price. Thus, the goal in the traditional RM setting is to build a base from early booking, price-sensitive consumers, and set aside sufficient inventory for higher-paying, nondiscretionary demand consumers later in the booking horizon.

Here, we argue that when demand is largely discretionary in nature, operators in nontraditional RM environments should not manage demand across the booking horizon in the same manner as conventional RM practice dictates. Rather, revenue managers should drive as much booking volume as possible early in the booking horizon to capture demand from inventory-sensitive, high-valuation consumers. Then, later in the booking horizon, the need may arise to lower price to penetrate the market and induce demand from less inventory-sensitive consumers (Desiraju & Shugan, 1999). Thus, we suggest that an ability to capture as many reservations as possible early in the booking horizon will yield superior revenue performance. Capturing as much demand as possible early in the booking horizon, rather than holding inventory for late bookers as traditional RM dictates, will allow firms to serve inventory-sensitive consumers who place value on inventory timing or location, and will pay more for that value. Hence, we hypothesize the following:

Hypothesis 1₀: Operators whose largest proportion of bookings are made early in the booking horizon will not yield the strongest revenue performance.

Hypothesis 1_a: Operators whose largest proportion of bookings are made early in the booking horizon will yield the strongest revenue performance.

Time-Based Supply Complexity and Price Positioning

A key dimension of strategic pricing is positioning, in which a firm determines their desired long-term pricing vis à vis their competitors, taking into account multiple factors, including product and service differentiation, brand reputation, and market share targets (Enz, Canina, & van der Rest, 2015). Tactical pricing decisions are subsequently guided by strategic price positioning goals (Abrate & Viglia, 2016).

Prior research in the domain of traditional RM suggests that maintaining a price position above that of the competition will yield greater revenue results. For example, a series of studies in the U.S. and European hotel markets demonstrated that pricing above direct competitors resulted in lower comparative sales volume (i.e., room occupancies) but higher relative revenue per available room (see, e.g., Enz & Canina, 2010; Enz et al., 2015; Noone, Canina, & Enz, 2013). Within the hotel context, and indeed within the car rental environment, competitive positioning strategies relate to daily price positioning, while, for airlines, competitive price positioning is typically considered at the route level. In nontraditional RM settings such as restaurants, spas, and golf courses, competitive price positioning is arguably more complex, with within-day demand fluctuations suggesting that operators need to consider the revenue effects of price positioning across distinct within-day peak, and off-peak demand periods.

Prior research suggests that the price elasticity of demand varies for peak and off-peak demand periods. For example, research in the domain of transportation services has shown that consumers demonstrate lower price elasticities during peak demand periods than during off-peak periods. Litman (2004), for instance, found that off-peak period elasticity values were about twice the peak period values. Research also suggests that leveraging these differences in price elasticities by charging a higher price in the peak period and a lower price in the off-peak period will result in increases in both revenue and customer throughput rate (Af Rantzien & Rude, 2014). Thus, in keeping with sound RM practice, we expect that operators who recognize within-day peak demand patterns could raise prices for these time periods to maximize revenue. In addition, managers could build capacity utilization during off-peak periods through the use of intelligent price discounting and value-add offers.

Drawing on previous work in the domain of competitive price positioning (Enz & Canina, 2010; Noone et al., 2013), we suggest that, for nontraditional RM settings where within-day fluctuations in demand exist (e.g., golf, restaurants, spas), operators who are able to manage their within-day price positioning such that they achieve a comparative price advantage over the competition during both within-day peak, and within-day nonpeak demand periods will yield the strongest revenue performance. However, we expect that the greatest revenue opportunities will be during within-day peak demand periods due to inventory- and price-insensitive consumers’ preferences for these times of the day. Therefore, we hypothesize as follows:

Hypothesis 2a₀: Operators who offer higher prices relative to their competition during the peak and off-peak time intervals of the day will not yield the strongest revenue performance.

Hypothesis 2a_a: Operators who offer higher prices relative to their competition during the peak and off-peak time intervals of the day will yield the strongest revenue performance.

Hypothesis 2b₀: The positive revenue effects of a price position above the competition will not be most pronounced during peak (vs. off-peak) times of the day.

Hypothesis 2b_a: The positive revenue effects of a price position above the competition will be most pronounced during peak (vs. off-peak) times of the day.

Inventoried Demand Management

Firms within both traditional and nontraditional RM settings cannot inventory supply but they can inventory demand using a reservation system. While there are a number of advantages to employing an advance reservation system (e.g., consumers are provided the certainty that they will receive their desired service at the promised time; the firm can intelligently leverage reservations data to shift demand from peak to off-peak periods, and anticipate how best to drive more demand to the facility through marketing and promotions), revenue is lost when reservations do not materialize (i.e., no-shows).

Operations-centric internal strategies exist in RM to better handle inventoried demand and mitigate the risk of customer exit without notice (Noone & Lee, 2011). For example, within traditional RM settings, operators often oversell on capacity (i.e., overbook) to counteract the potential revenue loss associated with no-show activity. Operators in nontraditional settings, on the other hand, can be reluctant to overbook due the problems that can ensue if more consumers arrive than expected. A lack of alternatives with which the consumer can be provided can discourage operators from overbooking (e.g., at a one-time concert or sporting event). Equally, operators often want to avoid the potential consumer dissatisfaction arising from long wait times, and associated pace and flow problems (e.g., for golf), when arrivals are higher than expected.

External approaches to managing inventoried demand encompass the design of procedures that discourage consumers from not honoring their reservation including, for example, cancellation fees, credit card guarantees, and late arrival policies (Kimes, 2000). The goal is to assure high conversion rates (i.e., actualized reservations as a percentage of total reservations made), with conversion reflecting a firm’s ability to successfully leverage customer-focused external approaches to managing inventoried demand. The revenue impact of conversion rate has received little attention in the literature. Here, we posit that the conversion challenge is much greater in the nontraditional RM setting than in most traditional RM environments due to the largely discretionary nature of demand (vs. a strong nondiscretionary component of demand in traditional RM settings). For example, the hotel industry average no-show rate has been reported as 1% to 2% (Visa U.S.A. Inc., 2007), while average no-show rates are as high as 20% in the restaurant industry (Fitzgerald, 2016). Nontraditional RM operators may be reluctant to enforce cancellation and no-show policies, particularly during low demand periods, in case enforcement results in the loss of market share to a more flexible competitor. However, failure to enforce cancellation and no-show policies may result in strategic consumer behavior where consumers “game” the reservation system. For example, if the price for a round of golf falls closer to the time of play, a lack of proactive inventoried demand management could easily incentivize knowledgeable players to make reservations far out at high prices, cancel at the last minute, and then rebook at lower prices just before play securing the optimal tee time in the process.

We expect that, in keeping with good RM practice, nontraditional RM operators who actively focus on managing inventoried demand to maximize conversion rate, will drive the strongest revenue performance. Furthermore, we argue that, where within-day fluctuations in demand are present, the positive revenue effects of conversion management will be most pronounced during within-day peak (vs. off-peak) demand period by virtue of the ability to maintain higher average prices during these demand periods. We hypothesize the following:

Hypothesis 3a₀: Operators who are best able to convert inventoried demand during the peak and off-peak time intervals of the day will not yield the strongest revenue performance.

Hypothesis 3a_a: Operators who are best able to convert inventoried demand during the peak and off-peak time intervals of the day will yield the strongest revenue performance.

Hypothesis 3b₀: The positive revenue effects of conversion will not be most pronounced during peak (vs. off-peak) times of the day.

Hypothesis 3b_a: The positive revenue effects of conversion will be most pronounced during peak (vs. off-peak) times of the day.

Method

Sample

This study is based on reservations data for all public golf courses, 80 in total, at a popular U.S. golf destination over a 2-year period. Private country clubs were excluded from the study as their pricing is based on initiation fees and dues, not rounds of golf played. The reservations data set consisted of 5,175,790 rounds of booked golf for the 80 courses under investigation over the period 2013-2014.

Reservations data were obtained from the destination’s shared tee time reservation system with the permission of the course owners, and in cooperation with the local golf course owners’ association. Because the focus of this study was daily gross revenue performance, we excluded complimentary rounds, and replay rounds of golf (i.e., playing the course again on the same day) for all courses. The individual reservation data included (1) the date a reservation was made, and cancelled, no-show, or played; (2) the reserved tee time; and (3) the total price paid by the golfer.

In addition to reservation data, data relating to each golf course’s competitive set classification were collected. The competitive sets, known as tiers, were established by the golf course owners’ association in cooperation with all of the course owners in the market as well as experts in the golf industry. Golf course characteristics including the quality of the course, unique features and design of the course, and the slope and length of the course were used to segment the courses into four competitive tiers, from high-end Tier 4 courses to economy Tier 1 courses, with an approximately equal distribution of golf courses in each tier. This classification system is analogous to the classification system for hotels (i.e., luxury, upscale, midscale, and economy), and given that it was endorsed by all of the golf courses in the market, it was used to determine relative competitive price positioning in this study.

The golf owners’ association also provided the research team with three weather-related indicator variables by date of play; frost, hot, and rain days.¹

Measures

Daily gross revenue for each golf course was calculated as the sum of the total fees paid by golfers for a round of golf at the course on a given date. The total fee included both greens and cart fees because these two elements of a round of golf were often packaged together making it difficult to examine the two fee types separately.

Booking window was measured by computing the proportion of bookings made at different points in booking horizon. We selected the booking window intervals, 7 in total, to enable us to hone in on the booking behavior of two key types of golf players, time-of-play–sensitive players and price-sensitive players, when probing the revenue–booking window relationship. First, to reflect the high volume of bookings from price-sensitive players near to the time of play, we identified 5 booking window intervals: 0 days before play (i.e., same day as play; 7% of bookings), 1 day before play (18% of bookings), 2 days before play (14% of bookings), 3 to 7 days before play (17% of bookings), and 8 to 14 days before play (5% of bookings). We then created the two remaining booking window intervals to reflect the booking behavior of more time-of-play–sensitive players: 15 to 90 days before play (17% of bookings), and 91+ days before play (22% of bookings).²

Relative price position was measured for both peak and off-peak times of each day, based on the price difference between a given course and the other courses in its tier. Demand trends in the market, based on the more than 5 million rounds of golf played in the golf destination for the 2013-2014 time period under investigation, were analyzed to identify within-day peak and off-peak tee times. Within-day peak demand periods were found to historically occur between the hours of 7.30 a.m. and 9.30 a.m. and between 12 p.m. and 2 p.m.; and off-peak hours represented all other times of the day. In a similar manner, day of week (Monday through Sunday) was distinguished, and seasons were classified as peak (March, April, May, and October) or off-peak (all other months). The seasonality of golf in this destination is well known by golf players. Golf course owners were consulted to assure that the within-day peak demand periods and seasons that were derived from the data were consistent with local industry leaders’ deep knowledge of historical demand.

Relative price position was then determined by time of day, day of week, and season as represented in the following equation:

\begin{array}{l} % Price difference from competitors = \\ \frac{[\begin{array}{l} Average price per round (i, d, h, s) \\ - Average price per round (c, d, h, s) \end{array}] * 100.00}{Average price per round (c, d, h, s)} \end{array}

where i represents the golf course, c represents the competitive tier that the golf course belongs to, d represents the day of the week (Sunday, Monday, Tuesday, etc.), h represents the time of the day (peak or off-peak hour), and s represents the season (peak or off-peak season).

For example, to calculate the relative price position of a golf course for peak hours (time of day) on Mondays (day of week) during the peak season, the average price per round for the golf course and the other golf courses in its tier were calculated using data from 7.30 to 9.30 a.m. and 12.00 p.m. to 2.00 p.m. on Mondays in March, April, May, and October.

Conversion rate captures the ability to convert reserved tee times to played rounds of golf. In this study, we distinguish conversion during peak times of the day and off-peak times of the day. The two measures of conversion rate were determined by calculating actualized reservations as a percentage of total reservations for the given peak and off-peak times of the day.

Analysis

The data sample consisted of both cross-sectional and time series data, suggesting the need to use a two-way fixed effects model in the analysis (Greene, 1990). The application of a two-way fixed effects methodology enabled us to capture numerous golf course attributes that are constant across time (e.g., quality of the course), in addition to time-related characteristics that are constant across courses and are not captured by the independent variables included in the model (e.g., economic conditions and the weather). Failure to capture these effects could lead to model misspecification and biased model coefficients resulting in errors in understanding course performance.

Daily gross revenue was entered as the dependent variable in our model, with the independent variables of booking window (91+ days before play as the reference group), two relative price position measures (one for peak and one for off-peak times of the day), and two conversion rate measures (one for peak and one for off-peak times of the day).

Results

The means and standard deviations for daily gross revenue, and the independent variables of interest in this study, are provided in Table 1. The mean daily gross revenue across the entire sample was $4,329. An average of 22% of golf bookings were made 91+ days before play, with another 17% made 15 to 90 days before play. A similar proportion of bookings, 39% in total, were made between 2 and 0 days before play. The smallest proportion of bookings were made 8 to 14 days in advance of play. The negative values for the mean relative price position (peak times of the day: −2.07; off-peak times of the day: −1.51) indicate that, overall, there was a strong tendency toward a price position below other courses in the competitive tier. The standard deviations suggest substantial variation in relative price position across competitive courses. The mean conversion rate percentages (peak times of the day: 67%; off-peak: 68%) indicate that the golf courses in the sample experienced a relatively high cancellation/no-show rate (i.e., 33% peak times and 32% off-peak times) across all times of the day. The standard deviations reveal substantial variation in cancellation rates across golf courses.

Table 1

Means and Standard Deviations

Variables	M	SD	Minimum	Maximum
Daily gross revenue ($)	4329.00	4119.00	0	43541.00
Booking window
0 days before play	0.07	0.11	0	1
1 day before play	0.18	0.15	0	1
2 days before play	0.14	0.12	0	1
3-7 days before play	0.17	0.15	0	1
8-14 days before play	0.05	0.10	0	1
15-90 days before play	0.17	0.16	0	1
91+ days before play	0.22	0.23	0	1
Relative price position
Off-peak period	−1.51	38.44	−97.80	175.91
Peak period	−2.07	35.87	−99.02	196.32
Conversion rate
Off-peak period	0.68	0.30	0	1
Peak period	0.67	0.25	0	1

The Pearson product–moment correlations are provided in Table 2. The correlation coefficients indicate that, of all booking window categories, bookings made 91+ before day of play were most positively correlated with revenue on the day of play (r = 0.39, p < .0001). In contrast, bookings made 1 day before play had the greatest negative correlation with daily gross revenue (r = −0.32, p < .0001). The negative correlations between the conversion rate measures and the reservations booked farthest from the day of play (i.e., 91+ days) suggest it is hardest to convert early bookers to players (peak hours: r = −0.23, p < .0001; off-peak hours: r = −0.18, p < .0001). Cancellations and no-shows are thus a serious challenge when securing reservations far in advance of play. Conversion of reservations in contrast is strongest for both peak and off-peak tee time periods when booking the day before play (peak hours: r = 0.29, p < .0001; off-peak hours: r = 0.23, p < .0001). The off-peak hours conversion rate is positively and significantly correlated with higher relative competitive price position during both peak and off-peak times of the day (r = 0.06, p < .0001 in both cases). Similarly, the peak hours conversion rate is positively correlated with higher relative competitive price position during both peak and off-peak times of the day, however, to a lesser extent during off-peak hours (peak hours: r = 0.02, p < .0001; off-peak hours: r = 0.01, p > .05). This suggests that higher-priced competitors either may devise more hurdles to cancellation, or are in greater demand, or both.

Table 2

Pearson Correlation Coefficients

		1	2	3	4	5	6	7	8	9	10	11	12
1	Daily gross revenue	1.00
	Conversion rate
2	Off-peak hours	−0.05**	1.00
3	Peak hours	−0.07**	0.39**	1.00
	Relative price position
4	Off-peak hours	0.28**	0.06**	0.01	1.00
5	Peak hours	0.27**	0.06**	0.02**	0.09**	1.00
	Booking window
6	0 days before play	−0.15**	0.06**	0.02**	−0.02**	−0.02**	1.00
7	1 day before play	−0.32**	0.23**	0.29**	0.12**	0.15**	0.02**	1.00
8	2 days before play	−0.16**	0.12**	0.16**	0.15**	0.14**	−0.08**	0.19**	1.00
9	3-7 days before play	−0.26**	−0.01*	0.04**	−0.17**	−0.15**	0.01*	−0.02**	−0.04**	1.00
10	8-14 days before play	−0.12**	0.02**	0.02**	−0.15**	−0.17**	−0.12**	−0.05**	−0.08**	−0.06**	1.00
11	15-90 days before play	0.27**	−0.09**	−0.13**	0.06**	0.04**	−0.21**	−0.32**	−0.22**	−0.28**	−0.09**	1.00
12	91+ days before play	0.39**	−0.18**	−0.23**	−0.01*	−0.01*	−0.26**	−0.49**	−0.38**	−0.40**	−0.16**	−0.04**	1.00

p < .05. **p < .0001.

Table 3 shows the results from the two-way fixed effects model estimating the impact of the booking window, competitive price position, and conversation rate on daily gross revenues. The model was significant in explaining gross revenue performance (F = 64.05, p < .0001, R² = 0.72). A detailed examination of the results is presented in the sections that follow.

Table 3

Fixed Effects Regression Model for Daily Gross Revenues

Independent Variables	Dependent Variable: Daily Gross Revenue
Independent Variables	Coefficient	t Value
Intercept	3779.41	9.86**
Booking window^a
0 days before play	−4006.00	−23.87**
1 day before play	−3998.02	−32.31**
2 days before play	−4795.72	−35.75**
3-7 days before play	−3373.22	−29.92**
8-14 days before play	−2307.58	−14.21**
15-90 days before play	−1465.12	−14.20**
Relative price position
Off-peak hours	3.36	3.09*
Peak hours	9.70	7.79**
Conversion rate
Off-peak hours	595.10	11.71**
Peak hours	1550.32	25.13**
F	64.05**
R ²	0.72
F test for fixed effects	58.47**
Wald chi-square
Relative price position off-peak hours = peak hours	9.02*
Conversion rate off-peak hours = peak hours	113.9**

91+ days before play was used as the reference group in the analysis.

p < .005. **p < .0001.

Booking Window and Revenue Performance

Booking window had a significant effect on revenue performance (p < .0001; see Table 3). Using the earliest booking window of 91+ day before play as the reference group, all other booking window categories had a negative effect on gross revenue, particularly between 2 days before play and 0 days before play (2 days before play: β = −4795.72, p < .0001; 1 day before play: β = −3998.02, p < .0001; 0 days before play: β = −4006, p < .0001). In other words, operators with the largest proportion of reservations made early in the booking window yielded the greatest daily gross revenues. The further in advance of play that reservations were made, the higher the daily gross revenues. Bookings on the day of play yielded the largest negative effect on gross revenue. These findings provide support for Hypothesis 1_a.

To highlight the revenue impact of building bookings early in the booking horizon, we conducted a follow-up analysis using Tier 3 golf courses. We selected the courses in Tier 3 because they represent the top revenue performers in the entire golf market under investigation with the average total gross revenue per course exceeding $3.6 million for the 2013-2014 period, outperforming even the premier Tier 4 courses in the market (see Table 4). Tier 3 also had sufficient variation across courses to merit dividing competitors into subcategories according to their revenue performance. We compared price and booking patterns for courses at the highest and lowest performance levels of Tier 3, where the highest and lowest performers were categorized by forming quartiles based on total revenue over the 2-year period. The lowest performers were in the first quartile and the highest performers were in the fourth quartile. While all Tier 3 courses were strong competitors relative to the rest of the market, a more granular comparison of high- and low- performing courses in Tier 3 allowed us to deeply examine the linkage between building demand early in the booking horizon and price per round.

Table 4

Summary Golf Course Gross Revenue Statistics by Competitive Tier (2013-2014)

Competitive Tier	Total Gross Revenue, $ (2013-2014)
Competitive Tier	M	SD
1 (Lowest)	1,543,956.44	975,987.28
2	1,653,614.95	950,247.44
3	3,608,715.90	2,260,966.00
4 (Highest)	2,010,929.27	750,390.28

Figure 1 shows the average aggregate price per round and the percentage of bookings throughout the booking window for the high-performing (see Figure 1A) and low-performing (see Figure 1B) courses in Tier 3. When looking at high-performing courses, the average aggregate price per golf round was at its highest early in the booking horizon (91+ days before play: $78.68). Price decreased over the booking horizon until the last week before play when the average price per round remained at approximately $42 (Figure 1A). In terms of booking activity, the greatest proportion of bookings occurred 91+ days before play (36%), with a further 28.9% of bookings made between 15 and 90 days before play (see Figure 1A). The patterns in price and booking activity for the low performing courses in Tier 3 are shown in Figure 1B. The average price of a round of golf was at its highest for the 91+ days before play booking window, although lower than the average for the high-performing courses in the tier ($63.63 vs. $78.68). Also, rather than leveling off in the last week before play (as with price for high-performing courses in the competitive tier), price increased slightly 2 days before play, from an average of $28.67 three to seven days before play to $34.46 two days before play, and continued to increase until the day before play (see Figure 1B). The greatest proportion of bookings occurred 91+ days before play, but this proportion was much lower than it was for the high-performing courses (25.5% vs. 36%). In the week before play, the proportion of bookings was much higher than it was for high-performing courses (34.31% of all bookings were made 2 days or less before play vs. 17.81% for courses in the highest performing category of Tier 3 courses). The upturn in price late in the booking horizon for the low-performing courses, coupled with a relatively higher proportion of booking activity nearer the day of play (vs. high-performing courses), was not sufficient to offset the lower proportion of bookings, and lower average price, relative to the high-performing courses, early in the booking horizon. Consequently, overall revenue performance relative to high-performing courses suffered. These findings lend additional support for Hypothesis 1_a and underscore the importance of driving booking activity early in the booking horizon.

Figure 1

Price and Booking Activity Across the Booking Horizon. (A) High-Revenue Performers in Tier 3; (B) Low-Revenue Performers in Tier 3

Competitive Price Positioning

The results show a significant positive relationship between relative price position and gross revenue for both peak and off-peak times of the day (peak times of the day: β = 9.70, p < .0001; off-peak times of the day: β = 3.36, p < .005; see Table 3). This finding provides support for Hypothesis 2a_a that argues that operators who charge higher prices than their competitors during peak and off-peak times of the day will have the strongest revenue performance. The difference in effect size for relative price position between peak and off-peak times of the day indicates that the positive revenue effect of sustaining a price position above that of the competition is most pronounced for within-day peak demand periods (χ² = 9.02, p < .005). This finding supports Hypothesis 2b_a.

From Booking to Play: Conversion Rates

In Hypothesis 3a_a, we argued that operators who convert inventoried demand to play during peak and off-peak time intervals of the day would yield the strongest revenue performance in keeping with traditional arguments within the RM literature. The results support this hypothesis, suggesting that conversion rate, both during peak and off-peak times of the day, had a significant positive effect on revenue performance (peak times of the day: β = 1550.32, p < .0001; off-peak times of the day: β = 595.10; p < .0001) as shown in Table 3. Furthermore, the effect of conversion rate on revenue performance was most pronounced for within-day peak demand periods (χ² = 113.90; p < .0001), providing support for Hypothesis 3b_a.

Supplemental analysis was conducted removing rain days from the data set to assure that cancellation was not unduly influenced by these external factors. The supplemental findings with rain days removed provided a similar finding on the impact of conversation rate.³

Discussion

Since the late 1980s, various aspects of RM, from the development of optimization and overbooking solutions for the RM problem, to the study of consumer reaction to RM practices and performance metrics, have been studied predominantly in the context of traditional RM applications (see, e.g., Baker & Collier, 1999; Dolasinski et al, 2019; Kimes, 1994; Lindenmeier & Tscheulin, 2008; Pekgün et al., 2013; Wittman & Belobaba, 2017). While researchers have advocated for the application of RM in nontraditional tourism settings, the potential impact of the unique characteristics associated with these nontraditional settings on RM practice and performance has been underresearched. In this study, we sought to contribute to the literature by empirically examining the revenue effects of RM practice in the golf environment, a nontraditional RM setting wherein the demand and supply characteristics of interest in this research are present. Specifically, golf represents an environment where demand is largely discretionary in nature, and, from a supply perspective, it exhibits a degree of time-based inventory complexity that is not encountered in traditional RM contexts.

Our study’s findings contribute to the literature in a number of ways. First, our results support the notion that, for nontraditional RM environments where demand is largely discretionary in nature, a fundamentally different approach to demand management to that used in traditional RM applications may be required. Within the traditional RM setting, the focus is on protecting inventory for late-booking, high-paying nondiscretionary demand, while the nontraditional RM setting with discretionary demand may require a focus on building demand early in the booking horizon from inventory-sensitive consumers who are willing to book early and pay a premium to secure their most desirable inventory location or time. In this study, we observed a negative relationship between close-to-play booking windows and revenue performance whereby the greater the proportion of bookings made early in the booking horizon (e.g., 91+ days), the greater the revenue lift. This is in sharp contrast to the traditional RM setting whereby a positive relationship between the proportion of reservations made late in the booking horizon and revenue performance can be expected. Much of the RM literature has focused on the development of optimization solutions that assume that prices increase over time, and as a result are designed to limit the amount of inventory sold early in the booking horizon such that sufficient inventory is available for higher-priced, late-booking demand (e.g., Belobaba, 1992). The results of this study suggest that these traditional optimization solutions are not applicable to nontraditional RM environments where demand is largely discretionary in nature. Rather, the goal should be to focus on driving as much high-priced, early bookings as possible to maximize revenue.

Second, while prior studies have broadly considered the revenue effects of competitive price positioning (Enz & Canina, 2010; Enz et al., 2015; Noone et al., 2013), this study extends the literature by demonstrating that, for nontraditional RM settings with a high level of inventory complexity, within-day demand patterns can have a significant impact on the price positioning–revenue relationship. Specifically, our findings suggest that maintaining a price position above the competition during within-day peak demand periods will garner the greatest revenue boost (vs. within-day off-peak periods). This can be attributed to lower price elasticities, and associated prices, during within-day peak demand periods (Af Rantzien & Rude, 2014; Litman, 2004).

Third, we extend the literature by empirically examining the relationship between conversion rate and performance, and the impact of within-day demand patterns on the nature of this relationship. Much of the existing research on managing inventoried demand has focused on the development of internal and external approaches to counter the negative revenue effects of cancellations and no-shows (Kimes, 2000). Our findings suggest significant within-day demand effects of conversion management. Because of higher average prices during within-day peak (vs. off-peak) demand periods, the positive revenue effects of conversion management are most pronounced during these within-day high-demand periods.

Recent research suggests that operators within nontraditional settings are moving toward a structured approach to RM implementation (Noone et al., 2017). The findings of this study provide insights to guide their efforts. First, our findings suggest that, in order to maximize revenue, operators in nontraditional RM settings where demand is largely discretionary in nature need to focus on proactively building demand early in the booking horizon where the potential exists to drive greater revenue per reservation from high-paying, inventory-sensitive consumers. Capacity constraints limit the number of consumers who can secure the most valuable inventory (e.g., tee time for a golf course, best seat location for a sports or arts event), so operators should consider developing high-value, premium-priced packages to encourage the purchase of additional, less attractive inventory early in the booking horizon. Unique packages should be developed to appeal to different market segments. For example, a golf and food and beverage package, whether the food and beverage component is offered at on-site facility or through a strategic alliance with a local restaurant, may appeal to golfing groups, while a golf and pro-lesson package may appeal to the rookie golfer.

Internal approaches to managing cancellations and no-shows (e.g., overbooking) cannot be readily applied in many nontraditional RM environments. This is a function of several factors, from a lack of alternatives with which the consumer can be provided if more demand materializes than expected (e.g., at a one-time concert or sporting event), to the negative impact that excess consumer arrivals can have on the consumer experience (e.g., long wait times, pacing problems). Thus, the application of external approaches to managing inventoried demand is essential. A strict cancellation penalty can be applied to reservations made early in the booking horizon, particularly for prime inventory (times or location), to prevent consumers from gaming the reservation system by canceling their reservations late in the booking horizon, and rebooking them at last-minute discounted prices.

While last-minute discounting can encourage early bookers to engage in undesirable cancellation and rebooking behavior, it can also encourage deal-seeking consumers to defer purchase in the belief that a lower, last-minute price will become available (Anderson & Wilson, 2003). Thus, rather than relying solely on discounting to manage distressed inventory and drive last-minute demand, operators could offer more value (e.g., adding amenities) at regular prices. Additionally, a good forecast of demand can aid the detection of low-demand periods where operators may be able to find alternative uses for their facilities rather than selling inventory at deep discounts. Weddings and family gatherings, for example, are becoming increasingly popular on golf courses, and at performing arts and convention facilities.

This study’s findings suggest that it behooves operators in nontraditional RM environments that experience within-day fluctuations in demand to maintain a price position above the competition, particularly during within-day peak hours. However, a strategy of pricing above the competition will only be sustainable if consumers perceive that the quality of the service experience is commensurate with the higher price. Thus, it is important that operators understand how their facilities are positioned against the competition in terms of quality. Online consumer reviews on, for example, Yelp.com, Tripadvisor.com, and golfadvisor.com, can provide insights into consumers’ quality perceptions of competing facilities. These insights can then be leveraged to determine the unique selling propositions that the operator can exploit to justify higher relative prices. While physical facilities can provide opportunities for differentiation, they are costly to alter in the short term. However, lower cost opportunities for differentiation exist, such as enhancing the quality of the service experience provided to consumers, in addition to improving the quality and range of products available to consumers (e.g., food and beverage, and merchandise offerings).

In terms of data management, operators need to track reservations data in such a way that detailed forecasts of demand can be developed to capture within-day fluctuations in demand as well as broader, more conventional seasonal demand patterns. Such demand forecasts will enable management to develop sophisticated within-day pricing and conversion strategies. Arguably, current reservations and distribution systems cannot support such sophistication, nor may the consumer be ready for such RM practices. However, just as RM practices have become increasingly sophisticated, and more acceptable to consumers, in traditional RM domains (Wirtz & Kimes, 2007), the potential to fully leverage RM practices in nontraditional RM settings is significant. Furthermore, as operators become more sophisticated at understanding time-based demand, there is a need to develop appropriate benchmarks and more refined metrics of performance for RM.

Limitations and Further Research

As with any study, the present study has a number of limitations that could serve as the basis for future research. First, this empirical investigation was limited to one context, the golf industry. Future research should examine other nontraditional RM settings with similar demand and supply characteristics—whether jointly or individually present—to determine the robustness of this study’s findings. Second, this study focused on three aspects of RM practice: demand management across the booking horizon, price positioning, and inventoried demand management. The revenue impacts of other aspects of RM practice also warrant empirical investigation. For example, pay-for-play models are being introduced in the golf industry, with resorts such as the Lodge of Four Seasons, Lake Ozark, featuring dynamic pricing during the golf season (Beall, 2017). Rather than pay a traditional flat greens fee, golfers pay for the time spent on the course. Arguably, this novel approach to the sale of time on the golf course will encourage a faster pace of play, which could increase throughput rate, associated revenues, and possibly segment the market further for expert golfers. Investigation of the potential revenue impact of this type of dynamic pricing, and consumer reaction to such practices is merited.

Concluding Summary

Many nontraditional RM settings exhibit distinct demand and supply characteristics that are not present in traditional RM environments. On the demand side, demand is often largely discretionary in nature, as opposed to comprising a mix of discretionary and nondiscretionary demand as in traditional RM. From a supply perspective, some nontraditional RM settings experience a high degree of time-based supply complexity not present in traditional RM environments. The findings of this study suggest that, when either, or both, of these demand and supply characteristics are present, a nuanced approach to RM implementation is required. Largely discretionary demand requires a focus on driving as many high-priced bookings as possible early in the booking horizon, rather than the traditional RM approach to limiting early bookings to protect inventory for late-booking demand. When a high degree of time-based supply complexity is present, operators need to be able to identify within-day peak and off-peak demand patterns such that pricing and capacity management strategies can be deployed across short time intervals throughout the day.

Footnotes

Notes

Breffni M. Noone, PhD (e-mail: bmn2@psu.edu), is an associate professor at the School of Hospitality Management, Pennsylvania State University, University Park, PA. Cathy A. Enz, PhD (e-mail: cae4@cornell.edu), is a professor of strategy at the School of Hotel Administration, Cornell University, Ithaca, NY. Linda Canina, PhD (e-mail: lc29@cornell.edu), is a professor of finance at the School of Hotel Administration, Cornell University, Ithaca, NY.

References

Abrate

Viglia

(2016). Strategic and tactical price decisions in hotel revenue management. Tourism Management, 55, 123-132.

Af Rantzien

V. H.

Rude

. (2014). Peak-load pricing in public transport: A case study of Stockholm. Journal of Transport Literature, 8(1), 52-94.

Anderson

C. K.

Wilson

J. G.

(2003). Wait or buy? The strategic consumer: Pricing and profit implications. Journal of the Operational Research Society, 54, 299-306.

Ayvaz-Cavdaroglu

Gauri

D. K.

Webster

(2017). Empirical evidence of revenue management in the cruise line industry. Journal of Travel Research, 58(1). doi:10.1177/0047287517737178

Baker

T. K.

Collier

D. A.

(1999). A comparative revenue analysis of hotel yield management heuristics. Decision Sciences, 30, 239-263.

Beall

(2017). A pay-by-the-hour greens fee is genius. And slightly diabolical. Golf Digest. Retrieved from https://www.golfdigest.com/story/a-pay-by-the-hour-greens-fee-is-genius-and-slightly-diabolical

Belobaba

P. P.

(1992). Optimization vs. heuristic methods for nested seat allocation. Transportation Science, 21, 63-73.

Boo

Kim

(2010). The influence of convention center performance on hotel room nights. Journal of Travel Research, 49, 297-309.

Chen

Tsai

Chen McCain

S. L.

(2012). A revenue management model for casino table games. Cornell Hospitality Quarterly, 53, 144-153.

10.

Desiraju

Shugan

S. M.

(1999). Strategic service pricing and yield management. Journal of Marketing, 63, 44-56.

11.

Dolasinski

M. J.

Roberts

Zheng

(2019). Measuring hotel channel mix: A Dea-BSC model. Journal of Hospitality & Tourism Research, 43, 188-209. doi:10.1177/1096348018789741

12.

Enz

C. A.

Canina

(2010). Competitive pricing in European hotels. In Chen

(Ed.), Advances in hospitality and leisure (pp. 3-25). Bingley, England: Emerald Group.

13.

Enz

C. A.

Canina

van der Rest

J. P.

(2015). Competitive hotel pricing in Europe: An exploration of strategic positioning. Cornell Hospitality Report, 15(2), 6-16.

14.

Fitzgerald

(2016). 7 sobering stats on no-shows in pubs and restaurants. Retrieved from http://liveresblog.zonal.co.uk/7-sobering-stats-on-no-shows-in-pubs-and-restaurants

15.

Greene

W. H.

(1990). Econometric analysis. New York, NY: Macmillan.

16.

Karande

Magnini

V. P.

(2011). The relative use of contextual and temporal reference price components in hotel and airline purchases. Journal of Hospitality & Tourism Research, 35, 119-141.

17.

Kimes

S. E.

(1994). Perceived fairness of yield management. Cornell Hotel and Restaurant Administration Quarterly, 35(1), 22-29.

18.

Kimes

S. E.

(2000). Revenue management on the links: Applying yield management to the golf-course industry. Cornell Hotel and Restaurant Administration Quarterly, 41(1), 120-127.

19.

Kimes

S. E.

Chase

R. B.

Choi

Lee

P. Y.

Ngonzi

E. N.

(1998). Restaurant revenue management: Applying yield management to the restaurant industry. Cornell Hotel and Restaurant Administration Quarterly, 39(3), 32-39.

20.

Kimes

S. E.

Schruben

L. W.

(2002). Golf course revenue management: A study of tee time intervals. Journal of Revenue and Pricing Management, 1, 111-120.

21.

Miao

Wang

B. X.

(2014). Modeling a cruise line revenue management problem. Journal of Revenue and Pricing Management, 13, 247-260.

22.

Lindenmeier

Tscheulin

D. K.

(2008). The effects of inventory control and denied boarding on customer satisfaction: The case of capacity-based airline revenue management. Tourism Management, 29, 32-43.

23.

Litman

(2004). Transit price elasticities and cross-elasticities. Journal of Public Transportation, 7(2), 37-58.

24.

Moe

W. W.

Fader

P. S.

(2009). The role of price tiers in advance purchasing of event tickets. Journal of Service Research, 12(1), 73-86.

25.

Noone

B. M.

Canina

Enz

C. A.

(2013). Strategic price positioning for revenue management: The effects of relative price position and fluctuation on performance. Journal of Revenue and Pricing Management, 12, 207-220.

26.

Noone

B. M.

Enz

C. A.

Glassmire

(2017). Total hotel revenue management: A strategic profit perspective. Cornell Center for Hospitality Research Report, 17(8), 3-15.

27.

Noone

B. M.

Kimes

S. E.

Mattila

A. S.

Wirtz

(2009). Perceived service encounter pace and customer satisfaction: An empirical study of restaurant experiences. Journal of Service Management, 20, 380-403.

28.

Noone

Lee

C. H.

(2011). Hotel overbooking: The effect of overcompensation on customers’ reactions to denied service. Journal of Hospitality and Tourism Research, 35, 334-357.

29.

Pekgün

Menich

R. P.

Acharya

Finch

P. G.

Deschamps

Mallery

. . . Fuller

(2013). Carlson Rezidor hotel group maximizes revenue through improved demand management and price optimization. Interfaces, 43, 21-36.

30.

Schwartz

Chen

C. C.

(2012). Hedonic motivations and the effectiveness of risk perceptions-oriented revenue management policies. Journal of Hospitality & Tourism Research, 36, 232-250.

31.

Schwartz

Stewart

Backlund

E. A.

(2012). Visitation at capacity-constrained tourism destinations: Exploring revenue management at a national park. Tourism Management, 33, 500-508.

32.

Talluri

K. T.

van Ryzin

G. J.

(2004). The theory and practice of revenue management (International Series in Operations Research and Management Science, Vol. 68). New York, NY: Springer.

33.

Tanford

Erdem

Baloglu

(2011). Price transparency of bundled vacation packages. Journal of Hospitality & Tourism Research, 35, 213-234.

34.

Tiger

A. A.

Howard

R. W.

(2007). Managing golf course revenue using pace of play modeling. Journal of Business and Entrepreneurship, 19(2), 79-87.

35.

Visa U.S.A. Inc. (2007). Prevent and handle no-shows: A guide for hotel owners and managers. Retrieved from https://www.belizebank.com/wp-content/uploads/2016/05/Prevent-and-Handle-No-Shows-A-guide-for-Hotel-Owners-and-Managers.pdf

36.

Wirtz

Kimes

S. E.

(2007). The moderating role of familiarity in fairness perceptions of revenue management pricing. Journal of Service Research, 9, 229-240.

37.

Wittman

M. D.

Belobaba

P. P.

(2017). Personalization in airline revenue management–heuristics for real-time adjustment of availability and fares. Journal of Revenue and Pricing Management, 16, 376-396.