Dynamic pricing in different Spanish resort hotels

Abstract

The present study uses data on seven 4-star hotels belonging to the same multinational hotel chain located in different Spanish regions. The objective is to estimate the dynamic prices that allow the hotel revenue maximization during high season. The study includes the demand functions of seven resort hotels and implements a dynamic pricing deterministic model to estimate the prices that will maximize the hotel revenue for each date of stay. The results point out general revenue management implications, mainly that hotels located in the same destination should follow individualized pricing policies, more focused in the specific hotel and tourists’ characteristics; while in practice, hotel companies apply similar pricing policies to hotels located in the same destination. Furthermore, the deterministic model performs well with the data available on seven different hotels with different customer profiles and hotel characteristics.

Keywords

demand function deterministic model dynamic pricing resort hotels

Introduction and literature review

The demand behavior in terms of price elasticity of demand and the effect of time across the booking horizon represent key requirements for the dynamic pricing process that allows the revenue maximization in the hotel resort sector. The reservation time is one of the most common ways of practicing price discrimination in revenue management (RM): the limited capacity of a hotel establishment and the expectations of selling the same room at a higher price at a later moment are the reasons for limiting room sales across the booking horizon (Aziz et al., 2011; Bandinelli, 2000), and this is defined as dynamic pricing.

Most of the literature about dynamic pricing models that allows the hotel revenue maximization are theoretical models and/or use data simulations (Anderson and Xie, 2010; Chatwin, 2000; Gallego and Van Ryzin, 1994; Levin et al., 2008; Zhang and Weatherford, 2016), that is, they do not test the models’ performance with real information and are unable to compare the results coming from different hotels (see Table 1). Other studies include real data but do not estimate any demand behavior variation across the booking horizon (Aziz et al., 2011; Bayoumi et al., 2013), the typical way of customer segmentation. Finally, other empirical papers confirm the use of dynamic pricing practices in the sector studying the information of hundreds of hotels located in different destinations (Abrate et al., 2012, 2019; Melis and Piga, 2017), most of them use the hedonic pricing model to assess the variable that affect prices.

Table 1.

Dynamic pricing literature in the hotel sector.

Author	Model/methodology
Abrate et al. (2012)	Empirical analysis studying the price variability of hotels located in European capital cities
Abrate et al. (2019)	Empirical analysis proving that the higher dynamic price variability lead to higher hotel revenues.
Anderson and Xie (2010)	Theoretical model/simulation that links the dynamic pricing with the MNL models by considering the different hotel attributes.
Aziz et al. (2011)	Empirical model that uses an algorithm to calibrate the elasticity value in the revenue maximization.
Bandinelli (2000)	Theoretical model/simulation which introduces a Poisson process.
Bayoumi et al. (2013)	Empirical model where the demand is estimated using a probit function.
Chatwin (2000)	Theoretical model/simulation where the demand is estimated using a probit function.
Chen and Schwartz (2013)	Experimental model studying the last minute booking dynamics with a binomial logistic regression.
Gallego and Van Ryzin (1994)	Theoretical model/simulation where the demand is estimated using a probit function.
Guadix et al. (2010)	Theoretical model/simulation which forecasts the individual and group demands.
Lee (2011)	Empirical model that uses a linear demand model.
Talluri and Van Ryzin (2004)	Theoretical model/simulation that combine capacity control and dynamic pricing models following a stochastic process.
Vives et. al. (2019)	Empirical model that uses a Cobb–Douglas demand function with two differentiated booking periods.
Vives and Jacob (2020)	Empirical model that uses a Cobb–Douglas demand function and a dynamic pricing deterministic model.
Vives (2018)	Empirical model that uses a Cobb–Douglas demand function with two differentiated booking periods.

Source: Own elaboration.

Note: MNL: multinomial logit.

Shy (2008) places the demand function as a direct and common way to reflect the consumers’ willingness to pay (consumer’s behavior). Lee (2011) points out that the demand function is a key tool in the hotel revenue maximization, as it is able to measure demand changes under different market conditions and pricing policies.

In that sense, the deterministic dynamic models consider the customer behavior across the booking horizon and allow the hotel revenue maximization (Talluri and Van Ryzin, 2004; Vives and Jacob, 2020; Zhang and Weatherford, 2016) point out that the deterministic models are suitable when the sensitivity of customer varies across the booking horizon. These models allocate the capacity according to the average demand across the booking horizon (Talluri and Van Ryzin, 2004) and try to predict the price elasticity of demand segments, which allow the revenue maximization at the date of stay (Guadix et al., 2010). Vives et al. (2019) highlight that most of the hotel elasticities present in the literature are static and usually inelastic, while Vives (2018) finds that the high season elasticities are mostly inelastic. Vives (2018) identifies some hotel/demand factors fostering inelastic demand. Finally, Vives and Jacob (2020) find that the elasticities distribution across the booking horizon cannot explain the variation of prices and bookings.

The present article uses the same data and the demand functions specified by Vives (2018) and implements the dynamic pricing deterministic model used by Vives and Jacob (2020). In fact, Vives and Jacob (2020) demonstrate that the deterministic model performs well when the average data match with the values to be estimated, while Guadix et al. (2010) obtain higher revenue estimations when they are compared with those obtained with stochastic models. The deterministic model does not only allow the revenue maximization, while considering the demand behavior and, hence, leads to correct segmentation maximization across the booking horizon, but also performs well with the data available and collected by most of the hotels. The objective of the study is to estimate the high season dynamic prices of seven resort hotels located in different Spanish regions, which allow the hotel revenue maximization across the booking horizon, and compare them with the observed prices and room bookings. The aim is to show then that the deterministic model behaves well with the data available on seven different hotels as well as with the behavior of customers when they book rooms. The empirical comparison of the deterministic model estimations, which identify the dynamic prices, allows the analysis of the RM practices in these hotels and the detection of erroneous policies. The main paper’s contribution is not only managing implications for each of the seven hotels but also the possibility to analyze and compare the price elasticities of demand and dynamic prices of seven hotels which can share some working methods and pricing strategies—as they belong to the same hotel chain and have the same hotel star rating and coastline location—but they present some significant differences, such as the destination, hotel size, specific facilities and services, and customer profile. As we pointed out, the main gap in the literature is the lack of empirical dynamic pricing studies in the hotel sector as well as the lack of studies describing the implications for the RM department of hotels.

Data and methodology

The present study uses the same data as in Vives (2018), data on seven 4-star hotels belonging to the same multinational hotel chain, two of which are located in Majorca, two in Tenerife, and the last three are located in Málaga, Cádiz, and Huelva (Andalusia region). Apart from the hotel location, the main differences that present the seven hotels are the hotels’ size, recent hotel renovation, the supply of additional facilities and services, the customer segments in terms of type of room and board, and the tourist nationality. The RM department of the hotel chain provided the data on the online transient reservations and prices across the booking horizon of a typical Spanish high tourist season, which goes from the end of June to the beginning of September of 2017.

Regarding the methodology used for measuring the different demand behaviors across the different seasons, booking horizons and hotels, a demand function using a Cobb–Douglas formulation is estimated, and this function is then linearized with the application of natural logarithms (as in Vives, 2018; Vives et al., 2019,¹ which prove that the demand function performs well in different resort hotels with different tourists’ profiles and hotel characteristics)

ln q_{t^{'}}^{d} = β_{0} + D u m m i e s_{y, d, b p} + β_{p} \cdot ln p_{t ′}^{d} + β_{r} \cdot ln r_{t ′}^{d}

D u m m i e s_{y, d, b} = \sum_{y = 1, 2,, \dots, Y} β_{y} D y + \sum_{d = 1, 2, \dots, D} β_{d} D d + \sum_{b = 1, 2, \dots, B} β_{b} D b

where q represents the average number of daily room reservations for each period of time when prices (p) are not changed by the revenue manager over the booking horizon, r represents the distance between the booking time and the date of stay, t′ represents the reservation dates over the booking horizon, that is, the period of time where p remains constant, and the Dummies include variables such as the year when the observation takes place (y), the date of stay (d), and the booking period (b).

In each group of dates of stay (seasons), two different demand functions are defined and estimated (as in Vives and Jacob, 2020; Vives et al., 2019), and both papers show two completely different behaviors across the booking horizon in the resort hotels: the first period, the farthest dates across the booking horizon before the date of stay, usually presents low levels of bookings and lower prices (early booking discounts) and the second period, the closest dates to the date of stay, the prices and booking generally increase.

Then, the two demand functions are introduced in the price optimization deterministic model presented by Vives and Jacob (2020) to estimate the prices that will maximize the hotel revenue for each date of stay (Aziz et al., 2011; Bandinelli, 2000; Guadix et al., 2010; Lee, 2011)

R (t^{'}) = Max \sum_{t^{′}} Q_{t ′} \cdot p_{t ′}

When we multiply the number of online transient reservations (Q_t′) by the price (p_t′), we can obtain the number of average daily room reservations (q_t′) that maximizes the revenue obtained along the booking horizon t′.

More specifically, a Lagrange multiplier method is used to estimate the prices and bookings along the booking horizon that maximize the revenue. Lagrange method allows the consideration of the hotel capacity and the booking moment factors in the revenue maximization process (Chatwin, 2000; Gallego and Van Ryzin, 1994)

M a x R (t^{'}) = \sum_{1, t^{'}} Q_{1, t ′} \cdot f (q_{1, t ′}, r_{1, t ′}) + \sum_{2, t^{'}} Q_{2, t ′} \cdot f (q_{2, t ′}, r_{2, t ′}) s . t . (1) λ_{1} : \sum_{1, t^{′}} Q_{1, t ′} + \sum_{2, t^{′}} Q_{2, t ′} \leq r

which is subject to the number of rooms r the revenue manager is willing to sell (s.t. (1)); the maximum number of rooms that can be sold is the hotel capacity.

Results

This section summarizes the dynamic prices and bookings estimations of the seven hotels shows a comparison with the observed prices, that is, average daily rate (ADR)—ADR from 2017 and ADR average for the period 2014–2017—and reservations—observed bookings from 2017 and observed bookings 2014–2017—and also presents the price elasticity of demand values for all hotels. Finally, some discussion of main results is included.

The hotel 1 is located at the North area of Majorca Island, it has 360 rooms, focuses on the couples and half board segments, and the majority of tourists come from Germany. Results on hotel 1 show (Figure 1) that the main recommendation is to reduce prices during period I and to increase them in period II. During the peak season (from July 15 to August 26), when seeking revenue maximization, the estimations obtained lead to recommend to slightly increase the bookings in period I (Per I) and maintain reservations in period II (Per II). In general, high season demands are slightly elastic, while there exist small differences between the elasticity values in period I and period II, but the period II demands are usually more elastic. In parallel with the elasticity values, the prices do not significantly vary among the seasonal periods included in the study. Finally, no large differences among estimated Prices, ADR from 2017, and the ADR from 2014 to 2017 are observed.

Figure 1.

Hotel 1 seasonal comparison: optimal prices and bookings, observed prices and bookings, and elasticities. Source: Own elaboration and Vives (2018).

The hotel 2 is located at the Southeast area of Majorca Island, it is the largest hotel of the sample with 619 rooms, it focuses on the family segment and the majority of tourists book the all-inclusive board and come from the United Kingdom. Figure 2 shows that, excluding the second half of July, the general tendency is the price convergence between period I and period II to reduce the level of bookings in period I and significantly increase reservations in period II. Regarding the second half of July, it is recommended to reduce prices and restrict the bookings in period I, while to increase prices and reservations during period II. In general, the estimations obtained lead to recommend to moderate prices during the most elastic seasonal periods, as the elasticity level increases, the price levels decrease compared with ADR 2017 (July 15–August 3 vs. August 4–August 26 periods). Finally, the price tendency of the last seasons has been a positive growth (increasing prices), as the ADR from 2014 to 2017 are significantly lower compared with ADR from 2017, which are quite similar to estimated Prices.

Figure 2.

Hotel 2 seasonal comparison: optimal prices and bookings, observed prices and bookings, and elasticities. Source: Own elaboration and Vives (2018).

The hotel 3 is located at the South area of Tenerife Island, it has 404 rooms, focuses more on the couple segment, all-inclusive board is chosen in almost all the reservations, and almost half of tourists come from the United Kingdom. In general, Figure 3 exhibits that in period I the estimations recommend to contain prices, while in period II the suggestion is to slightly increase prices. Nevertheless, period II demand is usually more elastic than the first period, while the general recommendation is to increase prices during period II, this result may be explained by the fact that the weight of the booking moment is higher than the elasticity effect in the demand behavior. In this specific case, as the differences in elasticity values between period I and period II increase, the closer are the estimated prices and ADR from 2017, so the inelastic demand makes the estimations differ from the observed prices. Finally, the price tendency of the last seasons has been an increase in prices, as the ADR from 201 to 2017 are significantly lower compared with the ADR from 2017, which are closer to the estimated prices.

Figure 3.

Hotel 3 seasonal comparison: optimal prices and bookings, observed prices and bookings, and elasticities. Source: Own elaboration and Vives (2018).

The hotel 4 is also located at the South area of Tenerife Island, it is the second largest hotel with 505 rooms, focuses on the couple and half board segments, and most of tourists come from the United Kingdom. Although the RM prices from 2017 significantly vary across the seasonal periods, the estimated prices suggest the price convergence across the booking horizon (Figure 4). Additionally, it is also recommended to maintain the same prices in period I and period II, except for the last seasonal period (August 21–September 11), where the suggestion is to increase prices in period II. The estimated prices are consistent with the elasticity values, during the first three seasonal periods (June 23–August 20), the general demands are more inelastic and period II is slightly more elastic compared with period I, while in the last seasonal period (August 21–September 10), the whole demand gets more elastic and period I elasticity increases over period II. Nevertheless, the estimated prices are significantly different to the observed prices (ADR 2017 and ADR 2014–17), while the observed prices are equivalent to the hotel 3 prices. This could indicate an erroneous pricing strategy followed in hotel 4, due to the fact that hotel chains frequently follow similar pricing strategies in hotels located in nearby areas. Finally, the tendency of the last seasons has been to increase prices, as ADR from 2017 are significantly higher compared with average ADR from 2014 to 2017.

Figure 4.

Hotel 4 seasonal comparison: optimal prices and bookings, observed prices and bookings, and elasticities. Source: Own elaboration and Vives (2018).

The hotel 5 is located in Chiclana (Cádiz), with a supply of 413 rooms, focuses more on the couple segment, shares a similar proportion of all-inclusive and half board reservations, and most of the tourists are Spanish. In general, the estimations (Figure 5) indicate that in period I the general recommendation is to keep or slightly lower prices and to increase bookings, while to increase prices and maintain bookings stable in period II. The peak season (August 4–August 19) is the most inelastic seasonal period and the best strategy is to raise prices; while in the most elastic seasonal periods, the results recommend to slow down the price increase. Finally, the general tendency has also been a price rise, as ADR 2017 are higher than ADR 2014–17.

Figure 5.

Hotel 5 seasonal comparison: optimal prices and bookings, observed prices and bookings, and elasticities. Source: Own elaboration and Vives (2018).

The hotel 6 is located in Torrox Costa (Málaga), it also has 413 rooms as hotel 5, focuses more on the family segment, shares a similar proportion of all-inclusive and half board reservations, and most of the tourists come from Germany. Unlike what happens with hotel 5 estimations, the tendency of hotel 6 for the main high/peak season (from July 14 to August 26) is to match prices among the different seasonal periods (Figure 6). During the inelastic and unit-elastic periods, the results recommend in period I to hold prices stable and to increase bookings, while in period II to significantly increase prices and reduce the room reservations. It is worth pointing out that the most elastic seasonal period (July 14–August 3) presents the lowest differences between the estimated prices and the ADR from 2017. Finally, the price tendency has been to slightly increase prices (ADR 2017 vs. ADR 2014–17).

Figure 6.

Hotel 6 seasonal comparison: optimal prices and bookings, observed prices and bookings, and elasticities. Source: Own elaboration and Vives (2018).

The hotel 7 is located in Ayamonte (Huelva), it is the smallest hotel with 300 rooms, it focuses more on the family and all-inclusive segments, and most of tourists are Spanish. Figure 7 indicates that in peak season (August 4–August 19) prices should be increased, especially by raising the period I prices over the period II prices, and at the same time maintaining the bookings constant in both periods. In the rest of seasonal periods, the recommendation is to reduce prices and to slightly increase bookings in period I, while to raise prices and moderate the reservation levels in period II. In general, hotel 7 demand is significantly inelastic in high season periods and the estimations recommend increasing prices because it will have a low impact on the booking levels. The price tendency is also increasing during the last seasons (ADR 2017 vs. ADR 2014–17).

Figure 7.

Hotel 7 seasonal comparison: optimal prices and bookings, observed prices and bookings, and elasticities. Source: Own elaboration and Vives (2018).

Discussion

In general, the results point out that the global price tendency has been to increase prices during the last years in Spain as well as most of the demand is elastic (as in Vives, 2018).

More specifically, the elasticity levels significantly affect the hotel dynamic pricing in high season, as Vives (2018) identify that hotel-specific characteristic can significantly affect the elasticity levels. Generally, inelastic season estimations recommend the price reduction in period I (the farthest dates across the booking horizon before the date of stay) and the price increase in period II (the closest dates to the date of stay), while the elastic seasons recommend the price convergence between period I and period II. However, the elasticity distribution across the booking horizon cannot completely explain the price and the booking variability (Vives and Jacob, 2020).

Regarding the implications for RM departments of resort hotels, it is observed that in practice hotels located in the same destination, which belong to the same hotel chain, follow similar pricing strategies. This can be explained by the fact that the same revenue manager deals with these hotels and applies similar pricing strategies. Results show that this can be an erroneous pricing strategy. Additionally, the present research note confirms the hypothesis made by Vives (2018) and Vives and Jacob (2020) in which the elasticity values and optimal pricing are affected by hotel-specific location, tourists’ profile, and hotel specific attributes. Therefore, the general advice is that the revenue managers must design individual pricing strategies according to the specific hotel characteristics, that is, their demand behavior. The demand function and deterministic model presented in the present study could help to improve the individual hotel pricing policy by exploiting the historical data collected by every hotel. In fact, the deterministic model behaves well with the data available on seven different hotels as well as with the actual behavior of customers when they book rooms.

Conclusions

The deterministic model allows the revenue maximization, while considering the demand behavior, which permits their correct segmentation maximization across the booking horizon. The article includes the estimations of prices and level of booking of seven Spanish resort hotels on high season. These dynamic pricing estimations not only have individual managing implications for the seven hotels but also point out general RM implications for improving hotel pricing, that is, we found that these resort hotels follow erroneous pricing strategies.

The hotel prices have tended to increase during the last 3 years, while the recommendation drawn from this analysis is to increase even more the prices in the three hotels located in Andalusia, while in the rest of destinations, the general recommendation is to vary the price distribution and room bookings across the booking horizon to maximize the revenue according to the different demand behaviors. The hotels located in the same destination follow similar pricing policies, due to the fact that in the hotel chains it is usually the same person the one who manages the nearby hotels. Meanwhile, the results show that hotels located in the same destination should follow individualized pricing policies, more focused in the specific hotel and tourists’ characteristics. Hence, results advice an adaptation of RM strategy taking into account hotel characteristics and tourists’ profiles as opposed to RM standardization which seems to be the norm at least in the resort hotel sector of a mature destination such as the Balearic Islands.

The main limitations of the study are that the estimations do not consider competitors’ prices, online hotel ratings, and do not allow the customers’ segmentation at the individual hotel level; in fact, these will be the future research directions to include this additional information in the demand function to improve the demand behavior estimations.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The work was financially supported by the Spanish Ministry for Science and Innovation through Project ECO2011-28999, by the Spanish Ministry for the Economy and Competitiveness through Project ECO2013-47301-R, and the Balearic Government through Research Fellowship FPI/1574/2013.

ORCID iD

Aldric Vives

Note

References

Abrate

Fraquelli

Viglia

(2012) Dynamic pricing strategies: evidence from European hotels. International Journal of Hospitality Management 31(1): 160–168.

Abrate

Nicolau

Viglia

(2019) The impact of dynamic price variability on revenue maximization. Tourism Management 74: 224–233.

Anderson

Xie

(2010) Improving hospitality industry sales: twenty-five years of revenue management. Cornell Hospitality Quarterly 51(1): 53–67.

Aziz

Saleh

Rasmy

, et al. (2011) Dynamic room pricing model for hotel revenue management systems. Egyptian Informatics Journal 12: 177–183.

Bandinelli

(2000) An optimal, dynamic policy for hotel yield management. European Journal of Operational Research 121(3): 476–503.

Bayoumi

AEM

Saleh

Atiya

, et al. (2013) Dynamic pricing for hotel revenue management using price multipliers. Journal of Revenue & Pricing Management 12(3): 271–285.

Chatwin

(2000) Optimal dynamic pricing of perishable products with stochastic demand and a finite set of prices. European Journal of Operational Research 125(1): 149–174.

Chen

Schwartz

(2013) On revenue management and last minute booking dynamics. International Journal of Contemporary Hospitality Management 25(1): 7–22. Available at: https://doi.org/10.1108/09596111311290192.

Gallego

Van Ryzin

(1994) Optimal dynamic pricing of inventories with stochastic demand over finite horizons. Management Science 40(8): 999–1020.

10.

Guadix

Cortés

Onieva

, et al. (2010) Technology revenue management system for customer groups in hotels. Journal of Business Research 63(5): 519–527.

11.

Lee

(2011) Study of Demand Models and Price Optimization Performance. Doctoral Thesis, Georgia Tech Theses and Dissertations, H. Milton Stewart School of Industrial and Systems Engineering, Available at: https://smartech.gatech.edu/bitstream/handle/1853/42914/lee_seonah_201112_phd.pdf (accessed 8 August 2016).

12.

Levin

McGill

Nediak

(2008) Risk in revenue management and dynamic pricing. Operations Research 56(2): 326–343.

13.

Melis

Piga

(2017) Are all online hotel prices created dynamic? An empirical assessment. International Journal of Hospitality Management 67: 163–173.

14.

Shy

(2008) How to Price: A Guide to Pricing Techniques and Yield Management. Cambridge University Press: Cambridge.

15.

Talluri

Van Ryzin

(2004) The Theory and Practice of Revenue Management. Boston: Kluwer Academic.

16.

Vives

(2018) Revenue Management Techniques in the Resort Hotel Sector: Segmentations, Demand Function Estimation and Optimal Pricing. Doctoral Thesis, Universitat de les Illes Balears, Tourism School.

17.

Vives

Jacob

(2019) Optimal pricing for online hotel demand: the case of resort hotels in Majorca. Journal of Vacation Marketing forthcoming.

18.

Vives

Jacob

Aguiló

(2019) Online hotel demand model and own-price elasticities: an empirical application to two resort hotels in a mature destination. Tourism Economics 25(5): 670–694.

19.

Zhang

Weatherford

(2016) Dynamic pricing for network revenue management: a new approach and application in the hotel industry. INFORMS Journal on Computing 29(1): 18–35.