國人國內旅遊RFM指標比較靜態分析之研究

Comparative Static Analysis of Domestic Travel RFM Indicators

楊依蓁(Yi-Jen Yang)

指導教授 ： 鄭聖時

長榮大學/管理學院/企業管理研究所/碩士(2007年)

https://doi.org/10.6833/CJCU.2007.00084

摘要

休閒觀光產業是二十一世紀最熱門的產業之一，旅遊市場不但具有高度的顧客導向，且是一個變化性極高的市場，產品之間的替代性非常強，不同層面的消費者需要的旅遊產品與追求的目標亦不盡相同，故不同遊客的需求、旅遊的動機與行為等互有差異。
為有效地瞭解國人國內旅遊活動之趨勢，本研究探討民國93年與當年第三季，國人國內旅遊總花費（M）、旅遊頻率（F）、最近一次旅遊至今時間（R）與娛樂費用之情況，實證瞭解人口變數、地理區位變數及旅遊行為變數對其之影響程度，其中將國內旅遊活動花費的數據，利用簡單線性迴歸模型（Simple Linear Regression Model）分析影響旅遊總費用的因素，或托比模型（Tobit Model）找出影響娛樂費用的指標；將國內旅遊活動頻率與最近一次旅遊至今時間的數據，利用階層羅吉特模型（Ordered Logit Model）建構預測指標，並以RFM指標探討國人國內旅遊情況，藉此尋找出貢獻度高的顧客群，且將旅遊市場做區隔、顧客金字塔及邊際效果的分析。
本研究的重要實證結果有：（一）人口統計變數與地理區位變數對RFM指標有不同的估計結果，而旅遊行為變數中，旅遊住宿方式為當日來回與旅遊主要目的為純觀光旅遊此兩變數對旅遊總花費、旅遊頻率、最近一次旅遊至今時間與娛樂費用預測指標中一致。（二）顧客金字塔層級劃分：1. 選定「年齡為12~19歲、居住地區為北部地區、旅遊住宿方式為當日來回、旅遊季節為第1季」組合為旅遊頻率的白金級顧客；2. 選定「職業為未就業或待業中、旅遊住宿方式為當日來回、旅遊主要目的為純觀光旅遊、旅遊費用來源為全部自費、旅遊季節為第2季」組合為最近一次旅遊至今時間的白金級顧客。（三）邊際效果分析：1. 在旅遊總花費方面：民國93年全年與第3季中，當訊息來源為其他轉變為訊息來源為觀光旅遊相關單位的話，業者能夠增加的收入最高。2. 在娛樂費用方面：民國93年全年與第3季的國內旅遊市場參與比率皆大於國內旅遊活動消費支出比率。3. 在旅遊頻率與最近一次旅遊至今時間方面：比較階層羅吉特（Ordered Logit）與階層普羅特（Ordered Probit）分析結果，唯有當日來回，能夠使民國93年全年與第3季旅遊參與次數增加；而當日來回與純觀光旅遊較能使民國93年全年與第3季從事旅遊的間隔時間縮短，可知兩者邊際效果分析結果相似。
上述實證結果將有助於觀光旅遊業者瞭解國人國內旅遊活動的型態與其影響因素，業者可藉由本研究發現，瞭解其特性及需求，透過資源的整合以建立一套經營管理策略來吸引遊客，方能創造出觀光旅遊業者獲利與符合消費者需求的雙贏局面。

關鍵字

旅遊總花費 ； 旅遊頻率 ； 最近一次旅遊至今時間 ； RFM指標 ； 娛樂費用 ； 托比模型 ； 階層羅吉特模型

並列摘要

Tourism industry is one of the most popular industries in the 21st century. Due to the facts that tourism is highly customer-oriented and ever-changing, firms should constantly provide various product packages to satisfy customers’ diversity needs. However, previous research mainly focuses on the amounts spent by travelers. Few discussed the frequencies and recencies of activities of travelers. This study tries to close this gap. In nature, this is an empirical study of domestic travel behaviors of Taiwan residents whose ages are over 12.
In order to deeply and effectively understand the behaviors of domestic travel activities in Taiwan, we first set up three forecast models for monetary amounts spent by travelers (M), frequency of travel (F) and time period up to the time of interview (or recency, R), or so-called the RFM indicators. Forecast variables are three parts: demographic variables, geographic variables and travel behavior variables. Data used in this study are collected from the 2004 Survey of Travel by R.O.C. Citizens, conducted by Tourism Bureau, MOTC. According to the nature of RFM dependent variables, ordinary least squares (OLS) regression is employed to the travel amounts and Ordered Logit regressions are employed to the frequency and recency of domestic travelers. For the sake of effective market segmentation, forward selection method is adopted to reveal the most important segmentation variables.
Regression coefficients are further used to construct customer pyramids and marginal effect analyses. We first calculate the means and variances of RFM indicators and then follow Bass and Zeithaml & Bitner(2000) suggestions. Three different customer pyramids are made up for M, F, and R indicators.
Major findings of this study are the followings: (1) Different demographic and geographic variables have different impacts on M, F, and R indicators. Travel behavior variables such as travel on the same day and pure sightseeing purpose have the same effects on M, F, and R indicators. (2) Customer pyramids for M, F, and R can be created. For example, the platinum target customers of frequency of travel are those age between 12 to 19 years, living in the northern part of Taiwan, travel lodgings way for same day, and travel in the first season (January to March). The platinum target customers are those who are currently unemployed, travel lodgings way for same day, the main goal of traveling being sightseeing, provide own traveling funds and travel in the second season (April to June). (3) Marginal effect analysis results show that raising the rate of market participation is much more important than raising the rate of conditional expenditure amounts.
Empirical results of this study can help tourism providers understand the behaviors of domestic travelers in Taiwan. In addition, tourism providers can have a sense of how to select their platinum target customers. In such, a win-win strategy for both tourism providers and their customers can be created.

並列關鍵字

Monetary Amounts Spent by Travelers (M) ； Frequency of Travel (F) ； Time Period up to The Time of Interview (R) ； RFM Index ； Entertainment Expense ； Tobit Model ； Ordered Logit Model

參考文獻

2. 李明宗(1994)。休閒、觀光、遊憩論文集。台北市：地景企業股份有限公司。

4. 林東封(2004)。旅遊電子商務經營管理。台北市：揚智文化事業股份有限公司。

1. 李銘輝(1991)。遊憩需求與遊客行為特性之探討。戶外遊憩研究，4(1)，17-33。

2. 沈立、林晏州(1999)。遊客選擇搭乘遊園公車行為之影響因素探討─以陽明山國家公園為例。戶外遊憩研究，12(1)，41-59。

5. 陳思倫、劉錦桂(1992)。影響旅遊目的地選擇之地點特性及市場區隔之研究。戶外遊憩研究，5(2)，39-70。