E-Branding in the Swiss Hotel Industry
Steeve Genaine

A dissertation submitted to the Ecole Hôtelière de Lausanne in partial fulfillment of the requirements for the degree of Bachelor of Science in International Hospitality Management

Lausanne
2006

Abstract

This research goes through the evolution of branding to e-branding and analyzes it for the Swiss lodging industry. From a database of 2'448 hotels in Switzerland, this research explores domain names used by Swiss Hotels. Based on many criteria such as: registration date, region (canton), language, touristy zones, general situation and the quality of the domain name, we tried to define what the general factors that influences the quality of a domain name in the Swiss lodging industry are.

Acknowledgements

A dissertation appears to be a personal work. But to complete a project of this magnitude it requires a lot of support.

I am especially thankful to:

Annemarie, Alain and Marie-France Genaine for their encouragements Roland Schegg for his guidance, advices, patience and help

Estelle Martin for her support with statistics

Elsa Berseth and Boris Mateev for their support in English

www.swisshotels.com for providing a database

Table of contents

Introduction 4

Literature review 5

Branding Backgrounds

Branding in the entire industry 5

Brand name 6

Branding in the lodging industry 7

Transformation to e-branding and its reasons 8

Domain names 8

The study

Methodology 10

Quality Coding 10

Database coding explanation 11

Results and discussion

Bivariate analysis for dependency between quality and the other variables 12

Dependenc y between the separate elements H, N, V 14

Multivariate studies 16

Limitations 18

Conclusions 19

Future research 20

References 21

Annexes

Table 1.1 - 1.13 25

Table 2.1 - 2.27 30

Table 3.1 - 3.4 35

Introduction

Branding is used in many businesses when talking about marketing. A brand is the leading flag of any enterprise. With the internet, e-branding has become a key issue for small and large brands. There are almost no barriers into buying a domain name, being online and selling any goods or services on the «market space». The lodging industry currently has a great opportunity to make its services more tangible through the use of this technology. By having an appealing domain name and an operational website, customers are now able to get the «feel» of a hotel through the World Wide Web.

As the Internet is also intangible, users are more aware and they usually rely on big brands to finalize the purchasing procedure. It seems hard for small businesses to have a great development through the Internet but potential buyers are millions.

Literature review Branding backgrounds

Business historians agree that branding itself is over 100 years old, with the majority of countries having trademark acts to establish the legality of a protected asset by 1890 (Rooney 1995; The Economist 1988). It was from 1800 through 1925 that was known as the richest period of name-giving (Rooney 1995; Hambleton, 1987). From these beginnings, branding has evolved as a major component of marketing strategy. Its uses and applications continue to grow and diversify. Although the focus of branding has shifted over the last two decades, its importance to the business community and the consumer has not diminished (Rooney 1995). In his book, Great American Brands, Cleary (1981) in Rooney (1995), writes that without trademark brands, there would be no trustworthy marketplace and no sure, simple way to know what to reach for and what to avoid.

Many feel that the development of new mega brands would be impossible in the future and money would be better spent on acquisitions than on research and development (Rooney 1995). The fact that 90-95% of all new products failed strengthened the argument that takeovers made more sense than trying to develop new successful brands (Dagnoli, 1990; The Economist, 1988; Rooney 1995).

Branding in the entire industry

A brand is a name, term, sign, symbol, design, or any combination of these concepts, used to identify the goods and services of a seller (Bennett, 1988; Turley, Moore, Patrick 1995). Despite the formal definition, the purpose of branding is essentially to build the product's image (Cleary, 1981; Rooney 1995). Ginden (1993) in Rooney (1995) underlines that the point of a name is to have consumers link it to quality. A brand is the commercial value of the trust between a company and a customer (Morgan and Pritchard, 2000; Nassar, Jones, Morgan), which is true in the sense that a brand name is a promise to the customer of a certain level of product quality and service execution (Muller 98). To many, a brand suggests the best choice (Ginden, 1993; Rooney 1995).

Brands introduce stability into businesses, help guard against competitive imitation, and allow consumers to shop with confidence in an increasingly complex world (Aaker, 1991; Tepeci 1999). Once customers have made a decision about a brand and its associations, they are often loyal to that brand, continue to buy it in the future, recommend it to friends, and choose the product over others, even those with better features or lower prices (Assael, 1991; Tepeci 1999).

Brand Name

Choosing a brand name for a consumer product or service is so critical that some writers argue it is one of the most important marketing management decisions (Landler, Schiller, Therrien 1991; Turley Moore, 1995). A brand name can provide a customer with a symbolic meaning, which assists in both the recognition of the product and the decision-making process (Herbig and Milewicz, 1993; Turley, Moore, 1995). A well-chosen brand name can produce a number of specific advantages including suggesting product benefits (McCarthy and Perault, 1990; Turley, Moore, 1995), contributing to brand identity, simplifying shopping, implying quality (McNeal and Zerren, 1981; Turley, Moore, 1995) evoking feelings of trust, confidence, security, strength, durability, speed, status and exclusivity (Shimp, 1993; Turley, Moore, 1995). There are even times, particularly when marketing homogeneous goods, where the brand name may be a product's only distinguishing characteristic (Skinner, 1990; Turley, Moore, 1995). Most introductory marketing textbooks recognize that a good brand name should also have several properties. A short crisp brand name is usually preferred over longer more complex names. It should suggest benefits or qualities associated with the product. A good brand name should be easy to spell, pronounce and remember. It also should be distinctive and free of any negative connotations (Turley, Moore, 1995).

Travis (2001) points out that every person in America could recite the mantras of the most popular brands as easily as their own name.

Today, this thought clearly resonates with people through recognizable slogans like: always Coca Cola, Nokia connecting people, Nestlé good food good life... Brand have names, and clearly define their business through them.

Branding in the lodging Industry

Branding and successfully managing strong brands is considered to be one of the key drivers of success in the hotel industry. Consumers often base their hotel-stay decisions on their perception of a specific hotel's brand name (Jiang, Dev, Rao 2002). As there are many brand names in the hospitality industry, hotel chains constitute a clas sic application of brand strategy. In fact brands are a quick way for hotels and hotel chains to identify and differentiate themselves in the minds of the customers (Prasad, Dev 2000).

In hotel marketing, branding may help to reinforce salient hotel attributes based on core or augmented aspects. It may also reduce consumer risks associated with the purchase of intangible hotel services (Onkvisit, Shaw 1989). Additionally, it may help hotels to achieve higher levels of repeat business especially for the regular user segments. More generally, branding can facilitate differentiation and positioning in a competitive marketplace (Connell 1992).

According to Dube and Renaghan study in 1999, frequent travelers indicated a strong reliance on a hotel's brand name and reputation when making a purchase decision. Since repeat guests are a hotel's richest source of revenues and profits, (Dube, Renaghan, 2000) it is common sense to develop and invest in the brand to generate incomes through the guest loyalty.

In saturated and highly competitive industries such as hospitality, the key to increasing and preserving market share is not just winning customers but keeping them. Brand loyalty is crucial in the hospitality industry because repeat business constitutes a large percentage of room and food sales. Brand loyal customers resist competitors' price cuts and help hospitality firms maintain high occupancy rates (Tepeci 1999). Although building and maintaining a brand loyal customer base is vital for competitiveness in the hospitality industry, it is hard to say that hospitality managers are successful in ensuring customers return to their properties (Lewis, Chambers, Chacko, 1995; Tepeci 1999). Previous studies showed that customers could easily switch among hotel brands (Warren and Ostergren, 1990; Tepeci 1999).

In their hotel study, Dubé and Renaghan (2000) figured out that the respondents mentioned brand name and reputation with the second-greatest frequency as a source of value driving their purchase, after the location.

Transformation to e-branding and its reasons

The Internet is now firmly established as a marketing tool. It serves as an integral part of the marketing mix, serving as a digital distribution channe l as well as an electronic storefront (Yelkur, DaCosta 2001). As the Web continues to be integrated into the global world of business, it is increasingly important for companies to differentiate themselves through brand strategies that exhibit clear messages and provide fulfilling experiences. Companies with well-established brand equity and brand power in the off- line world are well poised to extend their brand into cyberspace (Harvin, 2000; Nassar, Jones and Morgan 1998).

Brands are even more important in cyberspace than they are in most other channels or environments. With more and more choices from many providers that are relatively unknown, customers tend to choose a provider they know - one that represents a set of values or attributes that are meaningful, clear, and trusted, especially if they cannot see or confirm that the provider is «real» (Bergstrom 2000). Research shows that customers like to buy the brands they know and trust online (Harvin, 2000; Nassar, Jones and Morgan 1998).

Zook (2000) notices that registering a domain name has become increasingly easy and inexpensive, he also shows that Switzerland is the third country with the highest number of domains per 1000 people, after the US and Denmark.

Domain names

As the Internet became more stable and widespread, the distribution of names and network numbers quickly outgrew its initial clientele: military, government and governmentsponsored research organizations (O'Daniel; Wai 2000). The Domain Name System (DNS) helps users find their way around the Internet. Every computer connected to the Internet has one or several unique address (es) called «IP address» (Internet Protocol address); rather than typing «192.0.34.65,» people find the ICANN (The Internet Corp. for Assigned Names and Numbers) at www.icann.org (ICANN 2004). The Domain Name System (DNS) was then developed in the early 1980s to help combat this problem. Under the DNS system, one or

more names are assigned to represent a specific IP address. (Clark, Chou, Yen, 2001).

Securing an appropriate domain name is getting more and more difficult. That is because all types of commercial and non-commercial entities want to use the «.com» gTLD (generic Top Level Domain) for their domain names. Unfortunately, the current registry responsible for administering the «.com» gTLD, Network Solutions Inc. (now part of VeriSign Inc.), decided not to follow through with any type of enforcement mechanism and in the mid `90s removed any and all commercial restrictions for the «.com» gTLD. The ICANN recommended adding seven new gTLDs; info, biz, museum, name, coop, aero, pro (AFNIC 2005), to the existing list in order to provide more choices, more competition, and greater e- commerce opportunities to the public. These new domains are being established as a test to determine whether the introduction of new top-level domains would cause problems (Lovitz

2001).

Search engines bring less than one in ten visitors to a site (Statmarket, 2000; Raffa, Murphy, Mizerski 2003) and click through rates on advertising banners are less than one percent (Hanson, 2000; Raffa et al. 2003). With experience, Web users rely less on search engines and gravitate towards easy-to-remember domain names, such «as ford.com», to find

websites (Ries and Ries, 2000; Raffa et al. 2003).

In the past 40 years the domain name evolution on the web has been significant. In the 1970s the IP was a complicated number, in the 1980s domain names as the dot com appeared. In 2000 more and more extensions are created to face the increasing demand for a domain name that corresponds to the name of the brand in a dot com.

Murphy, Raffa and Mizerski (2003) arrived to the conclusion that 96%, except three of the world's top 75 brands, had a live «.com» site as an obvious name.

Adding new top- level domain names to the existing list would allow the domain name to be registered to a specific classification that would be similar to the way a trademark is registered. Adding new top- level domain names could create added confusion to the Internet (Clark, Chou, Yen, 2001). If domain names become more and more composite, perhaps internet users are going to use more and more search engines and finding information directly in the URL will become time consuming.

The study Methodology

A robot went through switch database (8/12/2002) and listed all the 2'448 «hotels» registered on switch in the «.ch». It provided also the registration date, the last modified date, the domain name, the name of the hotel, the city and the canton. From this database the restaurants, hotel service, others, and domain names were taken out to end up with 1 '671 Swiss hotels domain names.

If a hotel showed with several domain names, only the one with the best quality (see definition below) was taken into account.

Quality coding

An index representing the quality of the domain name was built, according to guidelines found in the literature.

Ingenieweb (2005) internet services has some advice on choosing a domain name; it has to be short (maximum 10 characters) so the visitor can memorize it and has less chances to mistype it, if possible avoiding numbers as the human brain retains more easily words, the domain name should not use hyphen as people never now if it is the «-» or the «_» sign. The «.com» is always the best choice if available.

Primary study: Q_algo

The base algorithm to determine quality of the domain name is defined as follows: 1 point if the word «hotel» is in the domain name

1 point if the domain name is less than 8 characters

1 point if the name of the hotel is in the domain name

2 points if the name of the town or city is in the domain name

The addition of those criteria gives the quality index.

This basic variable « q_algo » could not be used for statistical analysis (some levels were under represented). Thus a secondary variable has been built, merging some of the levels (0, 1=>0, 2=>1, 3, and 4 => 2).

Database coding explanation

To complete the database the following information have been determined and added to each domain name. Most of the codes are binaries to simplify the statistical calculation. (0 = no, 1 = yes).

Results and discussion

Bivariate analysis for dependency between quality and the other variables

Quality is significantly dependent of: The language

Take in Table 1.1 about here

The year of the domain name registration

Take in Table 1.2 about here

The fact of having many domain names for one hotel or not

Take in Table 1.3 about here

The region

Take in Table 1.4 + 1.5 about here

52.9% of the hotels which registered their domain name before 1998 are of high standing, and the proportion of high standing goes down with the years (or the proportion of middle and low standing increases). This effect is highly significant, even if the statistics is not precise because of the very small number of low quality names registered before 1998 (4).

Take in Table 1.6 about here

Quality increases obviously with the years (domain name registered more recently implies better quality). But it can be due to the correlation between the region and the year of registration.

Take in Table 1.7, 1.8, 1.9, 1.10 about here

There is a strong correlation between «having several domain names» and «does not answer any more» (either web site or domain name is not existing any more), but this effect might also result from our choosing only the best domain name when there are several.

Take in Table 1.11 about here

During the period of time from 1995 to 2003 that represents 8 years, 90.80% of the domain names were registered within 4 years; from 1997 to 2001.

Take in Table 1.12 about here

Repartition of the quality in 3 groups

Take in Table 1.13 about here

Dependency between the separate elements H, N, V

Proportion of domain names having the word «Hotel» (H), the «name» of the hotel (N) and the «town or city» (V) of the hotel in the domain name.

Prportion of the domain name having (H, N or V)

100.00

Significant summary of the dependencies between each couple of variables. The cells give the p-value, and are shadowed if the dependency is significant.

Interpretation example: cell Table2. 1: The fact that «the domain name contains the word hotel» is dependent on variable «DNS does not exist anymore» (p-value = 0.023).

Multivariate studies

The various bivariate studies (looking for dependencies between quality and the other variables at our disposal, looking for dependencies between the separate elements - H, N, V - and the other variables) showed dependencies with the variables «year of registration», «language» and «region». It is now important to check our findings with a multivariate study. We try to find a model predicting the quality of the domain name (or its composition) using several other variables, like the language, the region, the year of registration, at the same time. For example, a model that is able to take into account the mixed effects of being in a German speaking region and registered early.

Domains of a hotel located in a city, with domain name registered after 1998 are significantly more likely to contain the word hotel. The hotels being located in a German speaking region are less likely to have a domain name containing the word «ho tel», even if this effect is not significant.

Take in Table 3.1 about here

The fact that the hotel is in a German speaking region, implies that the domain name has 2.86 times more chance to contain the name of the hotel. What increases this probability is also the fact of belonging to the lake or mountain regions (especially mountain). When one goes up in standing, probability of having a domain name containing the name of the hotel decreases.

Take in Table 3.2 about here

The hotels in a German speaking region are nearly twice as likely to possess a domain name containing the name of the locality. The name having been registered after 1998 is also a strong predictor and the fact of having a high standing increases the probability for the domain name to contain the name of the locality. Being in town on the other hand decreases this probability. Hotels in town have 1/3 of the probability of having a domain name containing the name of the locality, compared to the hotels in the other regions (lake, mountain).

Take in Table 3.3 about here

Being in the German language speaking part, having registered after 1998, and being of a high standing improve quality of the domain name. Being in a city decreases it enormously.

Take in Table 3.4 about here

Limitations

The original database of 2'448 hotels was purified to 1 '672 hotels, which were of no use for the statistics (some were not hotels but chains, some domain names were pointing to all but hotels websites, and so on). All variables with more levels than three like the twelve Swiss touristy zones, the five types of stars, twenty-six cantons, had to be simplified. We used three levels of standing, the touristy zone was not examined, and the cantons were grouped in the three national spoken languages.

The World Wide Web, hotels classification, number of hotels and connection to website are changing every day. As this dissertation was prepared between 2004 and 2006 some of the information has been changing. New domain names were registered, replacing names we have in the database which sometimes do not exist any more.

Classifications with stars and number of rooms of the hotels taken are sometimes different from the hotel website and the www.swisshotels.com database. There is no fixed standard at this moment.

The human factor of the hotelier who buys his domain name was not searched and analyzed in this study. Results are based only on the domain names found on « switch. ch».

Conclusions

This study explored the Swiss hotels domain names and the determinant criteria that make a good quality domain name. The year of registration, the language spoken in the canton and the standing have an important correlation with the composition and quality of Swiss hotel domain names as showed in the results. There is no cause for purpose, but a correlation, an indication in the way in which people function when they register a domain name.

The results suggest that as the internet spread over the years, people learnt what a good domain name was. These results also emphasize the difficulty of obtaining a good domain name when located in a city where a lot of people register domain names of any kind.

According to Schegg, Steiner, Gherissi- Labben and Murphy study (2004), the top keywords used in search engines show that the most popular search terms are related to combinations of hotel, the city, the hotel's name, the region and activities/events.

In their study Angus and Oppenheim (2004) suggest to use existing words, such as noun or adjectives, as opposed to concocted ones. They also recommend short brand names that are easily memorized. The brand should enhance symbolic aspects that emphasize the unique qualities of the service.

The evolution of Internet and the numbers of domain names makes it very difficult to obtain the perfect domain name. Search engines are now the best support to collect the information and make a research for a hotel. The «old way» to search directly in the URL seems to be helpless for the brands not well known as one of the top 100 brands in the world.

Future research

Going through swisshotels.com database shows that some hoteliers have a domain name and a very different email contact address also some domain names are not online anymore, which is bad in terms of marketing. It would be of interest to edit a guide to the Swiss hoteliers who want to go «online» or improve his presence on the World Wide Web to avoid these mistakes.

For the hotelier who is trying to improve his presence on the Internet and the number of «click» on his website it is highly recommended to focuses on the website development rather than on the domain name itself as search engines goes through criteria found in the websites coding. A future research could be done to analyze the referencing words used in the website of Swiss hotels.

As this study did not take the hotelier's opinion in consideration to analyze the domain names a complementary study based on the hotelier choices to buy a domain name could help to understand the decision process of finding and buying a domain name in the Swiss lodging industry.

References

Aaker, D. A. (1991) «Managing Brand Equity», Free Press, New York, NY.

AFNIC (2005) «gTLD : domaines génériques» Association Française pour le Nommage Internet en Coopération online at : http://www.afnic.fr/doc/autres-nic/gtld [accessed 23 january 2005].

Angus, E., Oppenheim, C. (2004) «Studies of the characteristics of brand names used in the

marketing of information products and services. II: Internet related services» Aslib Proceedings: new information perspectives Vol. 56 No. 1, pp. 12-23

Assael, H. (1991) «Consumer Behaviour and Marketing Action» PWS-KENT, Boston, MA. Bennett, P.D. (1988) Dictionary of Marketing Terms, American Marketing Association, Chicago, IL.

Bergstrom, A. (2000) «Cyberbranding: leveraging your brand on the Internet» Strategy & Leadership. Chicago: Jul/Aug.Vol.28, Iss. 4; pg. 10

Clark, M. A., Chou, D. C., Yen, D. C. (2001) «Domain names management: a strategy for electronic commerce security» Information Management & Computer Security, Volume 9 No. 5.

Cleary, D.P. (1981) Great American Brands, Fairchild, New York, NY.

Connell, J. (1992) «Branding Hotel Portfolios» International Journal of Contemporary Hospitality Management, Bradford.

Dagnoli, J. (1990) «Beware line extensions», Advertising Age, Vol. 61, p. 62.

Dube, L., Renaghan, L., M. (2000) «Creating visible customer value» Cornell Hotel and Restaurant Administration Quarterly.

Dubé, L.; Leo, M. R. (2000) «Creating visible customer value» Cornell Hotel and Restaurant Administration Quarterly, Feb 41, 1, ABI/INFORM Global pg. 62.

Ginden, R. (1993) «The name game» Cheers, pp. 59-62.

Hambleton, R. (1987) The Branding of America, Yankee Books, New York, NY.

Hanson, W. (2000) Principles of Internet Marketing, Cincinnati: South-Western College Publishing.

Herbig, P. and Milewicz, J. (1993) «The relationship of reputation and credibility to brand success» Journal of Consumer Marketing, Vol. 10 No. 3, pp. 18-24.

ICANN (2002) `ICANN information', Internet Corporation for Assigned Names and

Numbers, online at: http://www.icann.org/general/ [accessed 23 november 2004].
Ingenieweb (2005) «Le bon choix», Les services web ingénieux online at:

http://www.ingeniweb.com/services/domain names/Choice [accessed 23 janvier 2005]. Jiang, W., Dev, C., S., Rao, V., R. (2002). Brand extension and customer loyalty: Evidence

from the lodging industry. Cornell Hotel and Restaurant Administration Quarterl.
Landler, M., Schiller, Z. and Therrien, L. (1991) «What's in a name?: less and less» Business

Week, July 8, pp. 66-7.

Lewis R.C., Chambers, R.E. and Chacko, H.E. (1995) «Marketing Leadership in Hospitality: Foundations and Practices» Van Nostrand Reinhold, New York, NY.

Lovitz, M. L. (2001) «Move over .com, make way for .biz» Marketing Management, Chicago, Jul/Aug.

McCarthy, E.J. and Perault, W.D. (1990), Basic Marketing, 10^th ed., Richard D. Irwin, Homewood, IL.

McNeal, J.U. and Zerren, L.M. (1981), «Brand name selection for consumer products», MSU Business Topics, Vol. 29 No. 2, pp. 35-9.

Morgan N. and Pritchard A. (2000), Advertising in tourism and leisure, Oxford: Butterworth Heinemann.

Muller C. C. (1998). «Endorsed branding», Cornell Hotel and Restaurant Administration Quarterly; Jun; 39, 3; ABI/INFORM Global pg. 90.

Nassar, M., Jones, E., Morgan, N. (1998) «Business Motivations: Hospitality brand-building websites - A pyramid of considerations.

O'Daniel, T., Wai, C. K. (2000), «Domain name and site hosting preferences: empirical evidence» Internet Research: Electronic Networking Applications and Policy, Volume 10, Number 4, pp. 308-3 16.

Onkvisit, S. and Shaw, J.J. (1989), «Service marketing: image, branding, and competition», Business Horizons, Vol. 32 No. 1, pp. 13-18.

Prasad, K., Dev, C., S. (2000). Managing hotel brand equity: A customer-centric framework

for assessing performance. Cornell Hotel and Restaurant Administration Quarterly. Raffa, L., Murphy, J. and Mizerski, D. (2003, «E-Branding: A Study of Domain Name Use

by the World's Top Brands» published in Electronic Markets 13(3), 30-40.

Ries, A. and Ries, L. (2000) «The 11 Immutable Laws of Internet Branding» New York:

HarperCollins.

Rooney, J. A. (1995) «Branding: a trend for today and tomorrow», Journal of Product and Brand Management; Volume 4 No. 4.

Schegg, R., Steiner, T., Gherissi-Labben, T., Murphy, J. (2004) «Using log file analysis and website assessment to improve hospitality websites».

Shimp, T.A. (1993), Promotion Management & Marketing Communications, The Dryden Press, Fort Worth, TX.

Skinner, S.J. (1990), Marketing, Houghton Mifflin Company, Boston, MA.

Statmarket (2000) `Search Engines Contribute Little to Site Traffic, online at: http://www.nua.ie/surveys/index.cgi?f=VS&art id=9053 56275&rel=true [accessed 20 January 2005].

Tepeci, M. (1999) «Increasing brand loyalty in the hospitality industry» International Journal of Contemporary Hospitality Management; Volume 11 No. 5.

The Economist (1988) «The year of the brand» The Economist, No. 309, pp. 95-100. Travis, D. (2001) «Branding in the digital age» The Journal of Business Strategy.

Turley, L.W., Moore, P. A. (1995) «Brand name strategies in the service sector», Journal of

Consumer Marketing, Volume 12 No. 4.

Warren, P. and Ostergren, N.W. (1990) «Marketing your hotel: challenges of the '90s" The Cornell Hotel and Restaurant Administration Quarterly, May, pp. 56-9.

Yelkur, R.; DaCosta, M., M., N. (2001) «Differential pricing and segmentation on the Internet: the case of hotels» Management Decision, Volume 39 No. 4.

Zook, M. (2000) «Internet metrics: using host and domain counts to map the internet», Current statistics Telecommunications Policy 24, pp.613-620.

Annexes

Table 1.1: The quality is not independent of the language

Quality |

Merged | a f i | Total

--+ +

0 | 435 51 28 | 514

| 84.63 9.92 5.45 | 100.00

| 30.04 36.96 32.94 | 30.76

1 | 576 67 38 | 681

| 84.58 9.84 5.58 | 100.00

| 39.78 48.55 44.71 | 40.75

2 | 437 20 19 | 476

| 91.81 4.20 3.99 | 100.00

| 30.18 14.49 22.35 | 28.49

Total | 1,448 138 85 | 1,671

| 86.65 8.26 5.09 | 100.00

| 100.00 100.00 100.00 | 100.00

Pearson chi2(4) = 16.8854 Pr = 0.002

Quality is not independent of the language. We particularly see this effect on the line corresponding to the highest quality: 30.18% of the domains in German are of high quality, against 14.49% for the French language and 22.35% for the Italian language. The German speaking sites are distributed about 1/3 by qua lity. For the French speaking sites, there is more quality 0 and 1, than 2, and very clearly. The effect is statistically significant, with a probability (p-value) of 0.002.

Table 1.2: The quality is dependent of the domain name registration year

Pearson chi2(6) = 46.4503 Pr = 0.000

Quality is not independent of the year when the domain names were registered. The effect is visible for example on the first column, among the 206 domain names registered before 1998; there are 42.72% of names of low quality (0). The proportion of low quality decreases when the years increase (39.33%, 28.86%, 24.97%). The proportion, in the <1998, high quality, is only 17.46%, and we see a linear evolution straightforwardly, of increase in quality. This effect is strongly significant.

Table 1.3: The quality is not independent of the fact of having or not many domain names

Quality is not independent of the fact of having or not several domain names. The effect is less apparent, and is not linear. The group of medium quality is the one which has the strongest proportion of "several domain names" (20.12% against 13.04% for quality 1 and 18.28% for quality 2). The effect is probably distorted by the fact that when a hotel had several domain names, only the highest in quality was taken. The effect is significant.

Table 1.4: The quality is not independent of the region

Pearson chi2(6) = 27.2733 Pr = 0.000

Quality is not independent of the region: typically for the cities, it is probably more difficult to obtain a good domain name because more domain names are used. Only 7.29% of the 96 domain names of city are of high quality, this proportion is of 29.6% for the mountain, for example. This effect could be still related to the year of registration, if there is a correlation between the year and the region:

Table 1.5

Pearson chi2(9) = 58.8200 Pr = 0.000

In the cities, the domain names were registered earlier (30% of the domain names of hotels located in a city were registered before 1998, against 10.45% of the domain names in the mountain, if we compare only these two). 2 1.5% of the names of city were registered after 1999, whereas 48.1% of the names in the mountain were registered after 1999. If the effect « year / quality » comes from the fact that internet is better used and understood and apprehended with the time being, then the region effect is perhaps especially a year effect.

Table 1.6

Pearson chi2(6) = 105.5103 Pr = 0.000

52.9% of the hotels which registered their domain name before 1998 are of high standing, and the proportion of high standing goes down with the years (or the proportion of middle and low standing increases). This effect is highly significant, even if the statistics is not precise because of the very small number of low quality names registered before 1998 (4).

Table 1.7: Quality related to the registration year, other point of view: ages of the domain names in relation with the quality

This confirms the effect detected above: the higher the quality, the lower the age of the domain name (in month). The mean age for the domain names goes from 82.9 1 months for quality 0, to 78.9 months for quality 1, and 75.5 for quality 2. The age decreases, and it acts as a tendency.

Same study but taking into account only the domain names that does not return a DNS error: Table 1.8

The effect is confirmed.

Considering quality as a continuous variable, we can look at the evolution of the mean quality as year of registration increases:

Table 1.9

yeargr | N mean

We clearly see that the quality increases with the year of registration.

Same study but taking into account only the domain names that does not return a DNS error: Table 1.10

yeargr | N mean

1998

2001

1996

1995

Year of registration

1997

Table 1.11: Strong correlation between « having many domain na mes » and « doesn't answer anymore »

More than |

One |

domain | DNS doesn't exist OR connection problem to website

name for |

this hotel | 0 1 | Total

+ +

0 | 1,230 150 | 1,380

| 89.13 10.87 | 100.00

| 84.83 67.87 | 82.59

+ +

1 | 220 71 | 291

| 75.60 24.40 | 100.00

| 15.17 32.13 | 17.41

+ +

Total | 1,450 221 | 1,671

| 86.77 13.23 | 100.00

| 100.00 100.00 | 100.00

Pearson chi2(1) = 38.3286 Pr = 0.000

Table 1.12: Frequency of year of registration

Year of | Registra |

tion | Freq. Percent Cum.

1995 | 4 0.24 0.24

1996 | 23 1.40 1.64

1997 | 179 10.90 12.55

1998 | 300 18.27 30.82

1999 | 395 24.06 54.87

2000 | 391 23.81 78.68

2001 | 226 13.76 92.45

2002 | 123 7.49 99.94

2003 | 1 0.06 100.00

Total | 1,642 100.00

Table 1.13: Repartition of the quality

28.5%

Quality repartition

2

40.8%

1

30.8%

0

Pearson chi2(2) = 3.5667 Pr = 0.168

0 | 293 38 208 396 | 935

| 31.34 4.06 22.25 42.35 | 100.00

| 56.78 39.58 52.39 59.82 | 55.95

1 | 223 58 189 266 | 736

| 30.30 7.88 25.68 36.14 | 100.00

| 43.22 60.42 47.61 40.18 | 44.05

Total | 516 96 397 662 | 1,671

| 30.88 5.75 23.76 39.62 | 100.00

| 100.00 100.00 100.00 100.00 | 100.00

TABLE 2.10

Pearson chi2(1) = 9.7358 Pr = 0.002 Pearson chi2(2) = 2.9878 Pr = 0.224

Pearson chi2(2) = 5.3672 Pr = 0.068 Pearson chi2(1) = 5.2572 Pr = 0.022

0 | 16 22 24 75 | 137

| 11.68 16.06 17.52 54.74 | 100.00

| 7.77 7.33 6.08 10.12 | 8.34

1 | 190 278 371 666 | 1,505

| 12.62 18.47 24.65 44.25 | 100.00

| 92.23 92.67 93.92 89.88 | 91.66

Total | 206 300 395 741 | 1,642

| 12.55 18.27 24.06 45.13 | 100.00

| 100.00 100.00 100.00 100.00 | 100.00

0 | 353 86 269 452 | 1,160

| 30.43 7.41 23.19 38.97 | 100.00

| 68.41 89.58 67.76 68.28 | 69.42

1 | 163 10 128 210 | 511

| 31.90 1.96 25.05 41.10 | 100.00

| 31.59 10.42 32.24 31.72 | 30.58

Total | 516 96 397 662 | 1,671

| 30.88 5.75 23.76 39.62 | 100.00

| 100.00 100.00 100.00 100.00 | 100.00

Composition of the domain name (h, n, y):

Table 3.1: Model: H according to «German» «after1998» «standing» and «region»

To be in the German speaking region decreases (0.71<1) the chance that the domain name has the word hotel, but in a way not completely significant (P-value=0.094). On the other hand, that the domain name was registered after 1998, and the fact of belonging to region 2 (city) increases the chance that the domain name has the word hotel.

Table 3.2: Model: N according to «German» «after1998» «standing» and «region»