Universtity of Amsterdam
Faculty of Economics & Business
.
Global e-commerce success: explaining of the country and
establishments differences at a major European airline
Master Thesis Research
Student
Vincent P. Lafeber – 5815568
Vincent.lafeber@zonnet.nl
Supervisor
Thomas Adelaar
t.adelaar@uva.nl
Library UVA version
Abstract
The internet is changing the world everywhere around the globe. Exploiting this channel gives airlines an opportunity to distinguish themselves from their competitors, due to increased customer satisfaction, lower costs and more choice of channels to customers. For this master thesis a major European airline, which started its website in 1995, is chosen. In 2009 the airline observed big differences over all countries where it was active with a website. In some countries less than 5% of the revenue is sold via the own internet site while in other countries this percentage is over 25%. The goal of this master thesis is to explain these differences. This paper starts with a thorough detailed explanation why it is so important for airlines to increase their direct sales through their own websites. Afterwards, literature on different perspectives will be briefly discussed, focusing on e-commerce, multi-channel distribution and strategy, but also making links to global variations in internet usage, wealth and cultural differences in countries and local airline industry pressure. A conceptual framework will be used, consisting of three clusters from an organisational, country and industry context. Measuring a myriad of variables on 72 countries will give some explanations on successful e-commerce execution by the airline. We find that effective execution of e-commerce and e-acquisition has a positive effect on the direct online sales, the size of the establishment in turnover and the market position do not increase the channel share of an airlines website. Also wealth distribution has an impact but cultural aspects seems not to matter. Market shares also do not seem to impact direct online sales whereas the product portfolio in terms of cheaper to more expensive products do influence the sales via the website of the manufacturer.
Preface
Writing a master thesis is quite a challenge. I completely agree with Saunders, who writes that it’s an iterative process, I also want to add that it’s often putting one step back and two steps ahead. I definitely overrated myself in the scope of this project but want to thank my supervisor Thomas Adelaar for steering where needed and sharing his confidence that I would bring the writing of this master thesis to a good end.
Table of c
o
ntents
1. Introduction ________________________________________________________________ 5
1.1 Research question ________________________________________________________ 6
1.2 Structure of the document __________________________________________________ 6
2. Literature review ____________________________________________________________ 7
2.1 Electronic commerce ______________________________________________________ 7
2.2 Distribution channels ______________________________________________________ 9
2.3 Airline Business _________________________________________________________ 10
2.4 Strategy ________________________________________________________________ 12
2.5 E-commerce and RBV _____________________________________________________ 13
2.6 Revenue size and resources allocation _______________________________________ 13
2.7 Webvertising ____________________________________________________________ 14
2.8 Execution _______________________________________________________________ 15
2.9 The model ______________________________________________________________ 18
2.9.1 Organisational context hypothesis ______________________________________ 18
2.9.2 Country context hypothesis ____________________________________________ 20
2.9.3 Industry context hypothesis ____________________________________________ 21
2.10 Control variables _______________________________________________________ 22
3.1 Sources ________________________________________________________________ 23
3.1.1 Organizational context sources: BI portal _________________________________ 23
3.1.2 Organizational context sources: Payment team ____________________________ 23
3.1.3 Organizational context sources: Country action plans (CAP) _________________ 24
3.1.4 Organizational context sources: Central e-acquisition team __________________ 24
3.1.5 Organizational context sources: MEDEAM ________________________________ 25
3.1.6 Country context sources: Euromonitor ___________________________________ 25
3.1.7 Country context sources: Internet World Statistics (IWS) ___________________ 26
3.1.8 Country context sources: HTTP Watch ___________________________________ 26
3.1.9 Country context sources: CIA fact sheet __________________________________ 27
3.1.10 Country context sources: ITIM international _____________________________ 27
3.1.11 Industry context sources: MIDT ________________________________________ 27
3.2 Consolidating the data ____________________________________________________ 27
3.3 Tests planned for execution _______________________________________________ 28
3.3.1 Variables explained ___________________________________________________ 29
4. Results of analysis & Discussion _______________________________________________ 30
4.1 Descriptive statistics _____________________________________________________ 30
4.1.1 Distribution mix – Channel shares _______________________________________ 30
4.1.2 Establishment size – Revenues _________________________________________ 31
4.1.3 Resources commitment _______________________________________________ 31
4.1.4 Effectiveness of e-acquisition ___________________________________________ 31
4.1.5 Email _______________________________________________________________ 32
4.1.6 Website reach _______________________________________________________ 33
4.1.7 Search & Metasearch ROI ______________________________________________ 34
4.1.8 Loading speed _______________________________________________________ 35
4.1.9 Gini coefficient _______________________________________________________ 35
4.1.10 Cultural Dimensions _________________________________________________ 36
4.1.11 Payment methods and Conversion ratio _________________________________ 36
4.1.12 Conversion ratio ____________________________________________________ 36
4.1.12 Market shares ______________________________________________________ 37
4.1.13 Coupon value _______________________________________________________ 37
4.2 Correlations _____________________________________________________________ 38
4.3 Unattained variables and cases _____________________________________________ 38
4.4 Regression analysis ______________________________________________________ 40
4.4.1 Direct online share ___________________________________________________ 40
4.4.2 Indirect off-line share _________________________________________________ 43
4.4.3 Direct off-line share ___________________________________________________ 44
4.5 Accepting and Rejecting the hypothesis ______________________________________ 46
5.1 Discussion ______________________________________________________________ 48
5.2 Limitations _____________________________________________________________ 50
5.3 Further research _________________________________________________________ 51
6. References _________________________________________________________________ 52
7.2 Regression 1 (n=49) _____________________________________________________ 56
7.2.1 Direct Online regression _______________________________________________ 57
7.2.2 Indirect Offline regression _____________________________________________ 58
7.2.3 Direct Offline regression _______________________________________________ 59
7.2.4 Indirect Online regression _____________________________________________ 61
7.3 Regression B (n=39) _____________________________________________________ 62
7.3
Revenues and Losses of all Airlines in the World ____________________________ 63
1. Introduction
Airlines are facing though times as the margin on tickets is under pressure. In the good years of 2006 and 2007 the airline in our research published net profits varying from 9,36 to 12,40 million euro on a total number of passengers carried between 70 and 73,5 million. This leads to a net turnover per ticket of between 12 and 13 euro. The following years 2008 and 2009 a loss was booked of respectively 186 and 1285 million euro. Over 4 years the net profit was almost € 9 million. If this profit is divided over the yearly average number of passengers carried this leads to a net profit per passenger of 30 cents. This low margin endangers the long term survival of the company whereas long term survival is very important to its stakeholders. One group of stakeholders, shareholders, want their shares to grow in value and receive dividends. Shareholders want a company to maximize their value and profitability (Brealy, 2006).
To be more profitable a company can increase its margin either via increasing the revenue or decreasing its costs. Distribution costs are a substantial cost driver for airline companies and the power of the distributors is substantial (Porter, 1979). A part of this airlines strategy is focused on lowering the distribution cost by focusing on the creation of the direct online distribution channel, their own sales website. This way the company gives direction to its actions, harvesting on its resources and competencies to gain advantages over the competition in the dynamic world of airline business (Johnson et al, 2006). The airline studied here started their website already in 1995 as an information and brand website. Since 1997 the website was also commercially exploited by selling tickets via the website. By the year 2010 76 websites in 76 countries where serviced direct online. Some countries sell over 25% of their revenues via the own website whereas others do not sell more than 5% of their tickets via the direct online channel. How can it be that in times where airlines really needs this shift towards e-commerce it is not happening in quite some of the countries? The purpose of this research is to try to identify factors that create these differences. With the results the management can make plans on how to increase the direct online share in the countries who are lagging behind.
1.1 Research question
Scientists in business studies find the internet very interesting, as it is changing the rules of the game. A lot of research is being done already in America and Europe as well as in Asia. However not a lot of research is looking at a global scope. Rob Law and Rita Leung (Law, 2000) study the Hong Kong consumers, Deborah Colton (Colton, 2010) studies only English speaking countries namely the US, Canada, Australia, New Zealand, UK, Ireland, India and Singapore and Buhalis and Li (Li, 2006) study the Chinese online consumer. Secondly not a lot of research is done within one company. This research is done within a global operating European airline company using all countries where there is a website that is owned ad branded by the airline. The main question we would like to answer by doing this research is:
“What factors can be identified to explain the differences in distribution mix for a major European airline” The plan is to investigate a wide set of possible factors and to conclude per factor if there is an
influence on the direct online share. Also we want to know if the factors influence the distribution mix in a positive or negative way. In order to decompose the research question some explanation will follow. Factors are variables that can influence the distribution mix. These can be resources of the an establishment in a country, contextual factors of a country or contextual factors of the airline industry in that country. A country will be one where the airline has a website where it sells tickets. When tickets are sold via the website this is one distribution channel, the so-called direct online channel. Apart from the website, travel agents are also a distribution channel. There are online travel agents (Indirect online distribution channel) or physical outlets in cities (Indirect off-line distribution channel). The last channel which is part of the whole distribution mix is direct sales via offline channels like the various call-centres around the world and the city and airport ticket offices where tickets can be bought by customers. The four channels together add up to 100%.
1.2 Structure of the document
First of all, existing literature will be reviewed on e-commerce, distribution and the airline industry. Than Jay Barney’s resource based view will be examined as well as well as Michael Porters external view on
strategy and competition. Afterwards it will be describes where all variables where collected from. These variables are put into a conceptual model that will be tested; the method used for testing will also be described. Before testing can be done the data will be analyzed for each variable that was collected. After explaining the results of the regression analysis, conclusions will be drawn. Interpreting the data within the context of some limitations will be discussed and at the end advice for further research on the subject will be given. In the appendix more detailed analysis output can be found as well as a list of scholars who have done research on these subjects. What these scholars have written can be read in the following paragraph.
2. Literature review
In order to study the subject of direct online sales within a global airline company several fields of research are reviewed. First of all, the growth of the internet as a commercial channel will be described. Why are both consumers and companies so focused on this distribution channel? This channel cannot be seen apart from the other distribution channel which will be discussed afterwards. As written in the introduction, the profit margins of airlines are very small; the reason for this will be discussed.
E-commerce is changing the rules of the game, but in the end it is the company that has the power to use this new technique wisely or unwisely. Many scholars have studied the competitive power of the company from the resource based view perspective and the external power perspective. The same applies to how to use e-commerce to the advantage of the company, e.g. selling directly to the customer. Companies have influence on e-commerce by using different methods of getting the consumer to their website, adding features to their website and allocating resources to this channel. Also the company can have advantages due to history or location, therefore country differences effecting e-commerce will also be discussed.
2.1 Electronic commerce
E-commerce can be defined as business activities conducted over the internet (Zhu 2004). E-commerce brings value for both the customer and the company. Amit and Zott cluster these value creation
Efficiency means lower efforts for customers to buy an airline ticket when the internet is used as it enables customers to search prices and benefits of many different suppliers from behind their home computer in a simple and easy way (Buhalis, 2008). Therefore the customer puts less effort in searching, has almost full transparency of the market and has more knowledge available on product differences and prices (Buhalis and Law 2008). On the other hand airline tickets can be complex especially due to the complex revenue management strategies which means that customers might lack the expertise to search, find and book the right product fitting their needs themselves (Öörni and Klein 2008). Another advantages for the customers is that they can do business 24/7 from anyplace anywhere as long as they have an internet connection and a computer at their disposal (Law and Leung 2000). Leyland and Pitt (Pitt, Berthon et al. 1999) say that the internet kills distance, homogenize time and make location irrelevant. The latter is also easier because the number of internet enabled devices is increasing and we can go online wireless via our Smartphone, laptop and computer from home and at our workplace (Horrigan 2009). Another advantage for customers
mentioned by Buhalis & Law (Buhalis and Law 2008) is that customers receive higher quality levels for services offered via the web.
The advantages for the airline company are numerous as well. First of all costs will be lower because the customer participates in the work that the buying and delivering of the service incorporates and staff costs are one of the biggest costs of an airline company. Why do costs go down because of the internet is also because of its global nature; it enables selling on a global scale very easily, leading to economies of scale (Steinfield, Mahler et al. 1999). Secondly, it enabled airline companies to change the power in the industry. Only if the internet as an enabling technology is used wisely, airlines could sell directly to customers, lowering the power of travel agents (Porter 2001). Companies often value e-commerce incorrectly because they or their government subsidize e-commerce investments, stock-buyers value e-commerce higher, and creative accounting results look better. Buy.com, a big retailer in the US, sold many of their products below cost price hoping to gain money via other business models. Direct customer contact leads again to
customer retention and input from customers enabling improvements of the quality of the services delivered (Buhalis and Law 2008). So, to conclude, we could say that the costs of doing business go down for both consumers and suppliers because of the internet (Wigand 1997). However, both Wigand and Porter say that the internet is only an enabler that can add value when used wisely. Porter says that logical thinking and existing strategy thinking still hold in the new economy. Many people thought that the internet era was changing the world. Debra Howcroft and Grover and Ramanlal argue that the changes are not that fundamental. They mention 7 myths, ideas about how e-commerce would change the world. First of all the new economy would change political, social and cultural world structures, secondly the definition of success would no longer be profit making (revenue-cost) and thirdly gaining wealth would happen in a short time span. The fourth myth stated that it would be easier competing by any company with the bigger competitors changing any industry to a level playing field. Also it was expected that e-commerce would make everything virtual and no material was needed anymore and the last, seventh myth, stated that we all go shopping online and traditional shopping outlets would disappear (Howcroft, 2001, Grover, 1999).
2.2 Distribution channels
The direct distribution via the internet is only one distribution channel that an airlines can use to distribute their tickets. Most airlines also make use of online travel agents (OTA’s) like Expedia, Interpark (Korea) or Cheaptickets (Netherlands). This channel is called the indirect online channel. Also in the off-line
distribution there are two channels. The first one is the oldest and biggest (roughly 2/3th distribution channel share)(Bilotkach, 2009) channel is indirect off-line channel which are the brick and mortar travel agent found quite often found in shopping streets. The last channel is the direct off-line sales channel, consisting of the call-centers and ticket offices of the airlines. The advantages of travel agents is that the airline can reach a more extensive customer base but the downside is the high costs of commissions and fees to Global Distribution Systems like Amadeus and Sabre (Koo, Mantin et al. 2009). Some argue that distributing via multiple channels like most airlines do creates complexity and thus additional overhead costs (O'Connor
and Frew 2004). Most budget airlines like Ryanair, Jetblue and Tiger Airways have chosen a single distribution strategy where they create a base of loyal customers they can control by price. Based on the success of these airlines Koo et al. assumes this to be a successful strategy. Steinfeld, Adelaar and Neslin argue that using a multiple channel strategy can create synergy effects via shared infrastructures,
marketing and customer bases leading to more loyal customers who spend more money at your company (Steinfield, Adelaar et al. 2002, Neslin et al. 2006). Lack of trust is one of the major reasons why customers do not buy online. If a company has a well known brand and physical retail locations where customers can come to complain or return products this channel has a positive effect on trust and thus on the acceptance of the direct online channel (Steinfield, Mahler et al. 1999; Goersch 2002).
2.3 Airline Business
In the past each respectable country had an airline which was owned by the government. Nations pride flew everywhere, there were agreements on landing permissions between countries and agreements on prices between airlines. Due to technological innovations in aviations technology airplanes became bigger and more fuel efficient and unit cost dropped. The creation of global Computerized Reservation Systems (CRS) in the 1970s and the internet (1994) have transformed the operational and strategic practices in the airline industry dramatically (Buhalis and Law 2008). A CRS like Sabre allowed real-time inventory
management from any remote location (Belobaba et al. 2009). Initially the CRS system was available through airline employees only, but later on it became available to travel agents. With the merging of CRS systems the Global Distribution System (GDS) was born. Worldwide, via any travel agent or airline system, real-time availability was shown and bookings could be made (Belobaba et al. 2009). In spite of all these cost reducing IT advancements, over the last 50 years only two brief periods of reasonable profitability could be noted (1965-1968 & 1975-1978). In general the airline business has been a very marginal business (Doganis 1991). Economic recessions and roaring fuel prices explain this. More than 30 years ago there came an end to bilateral country agreements and IATA cartel agreements. In 1978 the US government
stopped regulating the airline industry (Bowen 2009). This resulted in the emergence of low-cost carriers like Southwest airlines and Ryannair. The deregulation lead to the bankruptcy of at least 120 airlines like Pan Am and TWA (Freiberg, 1998). Competition grew, margins became smaller and airlines merged into to megacarriers like Oneworld, Skyteam and the Star Alliance. The three alliances together earn 80% of the worldwide passenger revenues (Star: 40%, Sky & OW 40%)(ATW, 2010). Southwest airlines was the only US
airline to make money each year from 1973 to 1998 (Freiberg, 1998). Recent developments like the banking crisis accellarated this negative trend. Sales of business and first class seats have plunged since fall 2008 (see figure 1). In 2008 all airlines around the world reported a net loss of 16 billion US dollar, in 2009
this was 10 billion whereas for 2007 a profit was reported of almost 13 billion US dollar (Flint, 2010).
Figure 1.-Airline Passenger traffic worldwide by cabin class. Premium traffic monitor & Monthly Traffic Analysis (IATA OCT09) In 1991 with the Gulf war another recession made airlines look for more cost cuttings. A lot was to gain via reducing the distribution costs which represented traditionally up to 25% of the revenues. In 1993
commission fees contributed to 13% of the costs and GDS fees were around 3 to 4 USD per segment. Low cost airlines focused only on keeping their costs down, especially distribution costs. This triggered legacy carriers (nation flag carriers) to focus even more on lowering costs (Bowen, 2009). By 2007 around 20% of all tickets were sold directly by the website of airlines.
2.4 Strategy
European carriers are affected by the laws of the European Union amongst others the laws on fair competition. The EU loves a perfect contestable market where above normal profits are eliminated. According the contestability theory (CT) this is possible if there are no costs involved by entering or exiting the market (Lockett, 2010). Competition comes from other airlines but also the power of consumers, suppliers and substitutes influence the margin that an airline can make (Porter, 2001). This perfect competitive market is worst for long run profitability (Porter, 1979). Porter suggests to influence the balance or to move to an industry where the forces are lower. Barney searches the long term profitability within the company. Either way searching for opportunities in a dynamic industry exists when there are competitive imperfections in a factor or market or when a market of perfect conditions like the CT theory dreams about does not exist (Alvarez, 2010). Porter stated already that the internet is an enabler that can be used wisely or unwisely meaning that the opportunities are out there but only if combined with the strengths and weaknesses of the internal company they can be harvested (Porter 2001). Wigand says that we have to add value and that that is possible if relationships between IT, strategy and processes are optimized and mistakes corrected and the organization is adapted to the changed environment (Wigand 1997). They both say that in order to stay ahead of the competition the internal resources and
competencies have to be sustainably different. Barney says that resources and competencies remain competitive if they are rare, unique, immobile and imperfectly imitable (Barney 2000). The resources and competencies also have to be valuable to the company or to the customer. They should lead to more effective and efficient execution of work. These competencies could be derived from events in history that if they do not occur again make copying by competitors impossible. If the competition cannot figure out why a company is successful and the employees of that company are also unable to do this this again leads to a sustainable advantage as the competition cannot perfectly understand how or what to copy. Barney calls this causal ambiguity and social complexity. Finally it is important that competitors have no substitute that could fulfill the same needs of customers. All kinds of resources in relation to e-commerce and
multi-channel strategy are already studied from the RBV view by scholars like financial resources, firm size, physical outlet infrastructure and web experience (Adelaar 2009). A set of variables will be reviewed from the RBV perspective. In this research both external factors will be reviewed as internal factors to see where the opportunities lie to gain a sustained competitive advantage via e-commerce.
2.5 E-commerce and RBV
A big part of e-commerce is based on information technology (IT). Although a lot of hard- and software can be bought off the shelf, the linking to legacy systems and existing processes can create synergies that can lead to performance advantages (Zhu 2004). In order to gain sustained competitive advantage via IT the focus should be on the IS capabilities as the assets are easier to be copied by the competition. An example of an SCA (sustained competitive advantage) in this case would be the ability to react quickly to market chances and build IT solutions to act on them (Wade, 2004).
Online booking penetration differs around the globe. In North America over 60% of tickets are sold online, whereas for South America this percentage is approximately 30% and for Asia around 14%. Reasons are the low internet and computer penetration and the lower levels of trust in the internet security and payment methods online (Belobaba, Swelbar et al. 2009).
2.6 Revenue size and resources allocation
The commitment of the company’s top management and the management of the foreign establishment towards e-commerce is important as this strategical apex sets out the direction for the organization and thus for all employees (Mintzberg, 1983). This is especially so in strongly hierarchical and formal
organizations: these organizations usually are so formal and hierarchical because of their age. Most European airlines which used to be state owned have such a hierarchical structure. If the resources like marketing budget and manpower are allocated more to e-commerce, the success of e-business might be ahead (Zhu and Kraemer 2005). The size of the firm also matters; the bigger it is the more resources can be allocated if the management has this focus. Although Zhu and Kraemer say there is a risk of inertia for
bigger firms, this probably doesn’t apply for the subsidiaries, as none of them is so big that inertia might happen. Although no definition is given of what is a small or large firm, most of the airline offices abroad count less than 50 heads per establishments and are therefore assumed to be small enough not to risk inertia. On the contrary, the head office might work somewhat inefficient but the strategy of the airline studied has chosen for a more decentralized managed field organization that the impact of the head office is expected to be minimal.
2.7 Webvertising
Getting customers to your website via advertising is called webvertising (Lovelock, 2007). There is an enormous number of websites that reduce the likelihood that people will find your web store therefore it is important to draw the attention of the consumer (Steinfield, Mahler et al. 1999). There are several channels that can be used to drive traffic to your website. First of all there is search engine advertising, then there is banner advertising, also known as display, another relatively low cost acquisition channel is email marketing and two more recent ways to get customer to the website are via affiliate marketing and metasearch (Lovelock, 2007). In march 2010 over 15,4 billion core search requests where inserted in one of the search engines like Google (Berman 2010). Most search engines have an organic and a paid part on their result page. The organic part can be influenced two ways, either via increasing the relevance on a website or by smartly using the knowledge on how the algorithm of search engines works (for as much as is known on these algorithms). Research shows that click through rates differ based on how smart the keywords to bid on are chosen: the length is important as well as including retailer specific information or the brand name (Ghose, 2009). Any company like an airline can pay for certain search words of customers, there’s no fixed price, it is just a bidding system (Berman 2010).
Another way to search in the enormous pile of data that the internet offers customers is to make use of metasearch websites. Metasearch engines increase coverage of a search and return large numbers of results focused on relevance and present alternative views of information needs (Jansen et al. 2007). A
Metasearch engines like Kayak.com do not index themselves like Google and Yahoo do but they let other search websites do the search simultaneously, so the customer only asks one website a question and gets answers from multiple websites.
2.8 Execution
Once customers are on the website it is important to keep them there: the website has to be sticky (Lovelock, 2007). Customers like a website that is easy to use and one that is useful (Pavlou 2003). One of the most logical things that make a website easy to use is one that loads quickly. Another important topic linked to ease of use is the trust. Lack of trust is probably the main reason why customers do not buy online (Steinfield, Mahler et al. 1999, Pavlou 2003). Consumers will not easily choose to buy online because the open and global nature of the world wide web creates uncertainty on losing their money when buying online (Pavlou 2003). Various digital online methods of payment like real time online banking, banktransfers or debit cards help consumers transact with convenience and confidence via the internet (Chen 2008). A study in Hong Kong found that the most important factor for a successful website was to have secure payment methods (Law 2003). Thus a company should add multiple payment options to the website making sure it fits the customers local needs, offer the payment options they trust and put secure logo’s on the website (Pasman, 2009). An option like Real time online banking (RTOB) is branded
differently per country but grouped as one method in the analysis. RTOB in the Netherlands is Ideal, in Singapore it’s eNets, India Netbanking and in Australia it is Bpay. Trust, according to Pavlou, has an underlying variable which is called perceived ease of use. Making a booking online should be without any effort. A website that is easy to use will lead to visitors staying longer on the website which in itself
increases the chance to convert a visitor into a consumer by persuading him to make a sale. A website that is quickly displayed on the screen increases this ease of use (Lovelock and Wirtz 1996). Before a website can convert a visitor into a sale (conversion) the customer has to be attracted to the website, this is done via webvertising
2.9 Internet penetration and speed around the globe
Even though the number of internet users is growing at an enormous speed there are still a lot of countries where a very limited number of people are connected to the internet (Bradshaw, 2001). All countries in Africa have at least 90% of their inhabitants without access to the internet. On the other hand in European countries like Norway, the Netherlands and Sweden around 50% of the people were
connected to the internet in 2001. In Asia there are big differences; Korea, Singapore and Japan have high percentages of internet users (+- 40%) whereas Afghanistan and Iraq have negligible numbers of internet users. The same picture we see for America where the USA & Canada show high number of internet
connections and in most Latin America low number of people are connected (Akamai, 2010, Gehrke, 1999). With such great disparities within, between and across countries we cannot expect internet sales
possibilities to be the same around the world. Another external factor that influences internet sales success is the speed of the internet in countries. Users of websites are not very tolerant to slow websites (Galetta, 2003). The slower the website the more negative the attitude of consumers towards the website and their behaviour (buying). Servers can be slow, computers of users can be slow, also telephone lines can be the bottle neck or the lack of fiberglass or other broadband alternatives (Gehrke, 1999). Broadband connections have a speed of at least 5 megabytes per second (MBPs) (Akamai, 2010). South Korea has the highest level of broadband connectivity around the globe with an average of 15Mbps. Over 75% of the Koreans have broadband internet access. Other countries that score high are Japan (60% > 5MBPs), Hong Kong (46%), Romania (42%) and Sweden (39%). The global penetration is very low, with only 1 person out of every 100 people having broadband connections. The same applies for the number of internet
connections. One more factor we can look at is how much of the money spent in a country is spend online by those consumers. Some scholars state that around 75% of the worldwide expenditure is spend only in the USA (Oxley, 2001). To conclude we can say that worldwide there are big differences in the number of internet connections, the speed of the connection and the online expenditure.
2.10 Level of economic development
Another factor that is studied in the realm of e-commerce is the level of economic development (Oxley, 2001). If consumers have limited amounts of money to spend on computers and internet connections the likelihood that they will buy online is also limited.
2.11 Culture
A third country specific factor is culture. One size doesn’t fit all, concepts are created from the perspective of a culture and thus work best within that culture (Hofstede, 1993). Hofstede categorizes 5 important dimensions in cultures. The first is about equal distances between people, the second describes wether people rather act individualistic or prefer to work in groups. Masculinity is the third dimension and is the opposite of femininity. The fourth dimension describes how much risk people are wiling to take or to avoid. And the last dimension characterizes the long term thinking of people within a culture versus opportunistic short term views.
2.11 Competitive pressure
For this research competition will not be limited to other airline companies. Distribution channels also have power and put pressure on the airline company. Furthermore we define substitutes and suppliers like the GDS systems (e.g. Amadeus) as competition (Porter, 2008). According to Porter it is the business structure that drives the profitability. With over 60 airlines competing around the globe there is quite some internal rivalry (ATW, 2010). Market share is an important and accurate measurement to measure brand equity (Aaker, 1996). Brand awareness is an important component of brand equity. It reflects the salience of the brand in the customers mind and is about recognition and recall. If an airline has a high market share, the brand awareness will also be higher and the chance that a customer books directly with the well known airline’s website is more likely. On the other hand, if travel agents have a good brand name, the customer will probably book via the travel agent, who has more influence on the airline choice.
2.9 The model
Based on the existing literature on the resource based view, outside market and forces view and
e-commerce a model was created that should explain the effects on the channel distribution. The model consists of three categories, namely the internal organizational/establishment context, the specifics of a country and thirdly competitive factors from the airline industry. In order to verify if correlations should not be explained by other variables
some control variables where added to the research.
2.9.1 Organisational context hypothesis
The first 5 hypotheses look at the organizational or establishments context. How big is the establishment in terms of power in the market and how establishments organize themselves regarding the e-commerce channel. Size matters, as it will lead to a higher spending power opening more possibilities if a company or establishment wants to boost the direct online channel. Therefor the first hypothesis to be tested is: H1: The bigger the establishment is the more developed and organized the establishment will be on e-commerce leading to a higher direct online share.
If two establishments both spend 40% of their budget to boost e-commerce but the budget of one is twice the one of the other, the mathematics of economies of scale apply (Zhu and Kraemer, 2005). So we expect bigger establishments to profit from economies of scale and have relatively higher web sales.
Having a certain size, presence and spending power doesn’t mean that the management also wisely subscribes their resources towards the e-commerce channel. Does the management really assigns more of their employees to work on the e-commerce channel and spend a big part of their marketing budget on e-commerce, then this will actually lead to more sales via the airline website. The second hypothesis can be drawn as follows;
H2: The financial and FTE commitment of the establishment management towards e-commerce has a positive influence on the direct online share.
Supposedly this is also a matter of mathematics as we look at relative channel share. If one establishment is double the revenue size as the other but the latter one spends 75% of their budget on e-commerce and the first only 25% they spend absolutely the same but the first has double the absolute sales to reach to have in the end the same percentage level. Another factor that is expected to have influence is the cultural factor in the organization: commitment (Mintzberg 1983, Kraemer 2005). If a manager believes strongly in e-commerce there’s a bigger chance in reaching more sales via e-commerce. The top management of the studied airline believes that e-commerce is the future. If however establishments do not believe this, their commitment and (marketing) spending towards e-commerce will be lower.
The following three hypotheses measure the quality and effectiveness of these resources. Having allocated resources is one thing but deploying them effectively is just as important. Hypothesis number 3 looks at the effectiveness of the e-acquisition channels. Get as many customers as possible to the website via email, search, display or even social media and try to convert them to making a sale on the airlines website. Online acquisition has a positive effect on website traffic (Ilfeld, 2001). Ilfeld also says that brand building via online advertising will not work neither will off-line advertising lead to site visits that much. Therefore site-visits is an important goal on itself to reach.
H3: More efficient spending of marketing funds will lead to a higher share of the direct online channel. Get more consumers to your website with less money spend will increase the efficiency leading to more value for money. This means choosing smarter the search keywords to bid on (Ghose, 2009) and choosing the right subject line for your email, the right length of text and number of imgages (Chittenden, 2002). The same applies for all webvertising channels like Metasearch; knowledge and experience will lead to more efficient use of marketing spend and thus to more sales via the direct online channel.
Not all customers who arrive on the website will eventually also make the sale. Although the booking tool is the same for each country there are differences made per country. One is the number of payment methods which is one of the main reasons not to buy online due lack of trust (Steinfeld, 1999, Pavlou
2003). As Credit Cards are known often to be misused the trust will be higher if alternative trustworthy payment options are offered (Law, 2008).
H4: Better execution of e-commerce will lead to more trust to finish the booking via the website leading to a higher share of direct online sales.
2.9.2 Country context hypothesis
The following two hypotheses state that the environmental context of a country has influence on the online buying behavior of airline tickets via the airline website. Even if an establishment is highly committed and deploys effectively their resources to increase the e-commerce channel it still might not work because the environment is just not ready for e-commerce. A company could loose customers during their website visit if the experience of the visit is not good, the speed of going through the website is one of the most important ones (Gherke, 1999 ). A quick loading speed time will lead to a higher number of people finishing what they came to do on the website (Galletta, 2003), for the ones who where ready to make the buy this will lead to a higher share of the website channel. Advanced e-commerce we can measure via the speed of the internet in countries based on the number of broadband connections (Bradshaw, 2001). In many African countries the telephone lines are so bad that internet access is hardly possible especially at a decent speed. So the combination of the number of internet connections and the speed of these connections define the fith hypothesis;
H5: In a country where people are more advanced in e-commerce the share of the direct online channel for airline company will also be higher.
It might however still be that consumers are e-savvy but mainly interact online or buy smaller less expensive items online like CD’s and books but not higher risk products and services like airline tickets. In fact it is even the question if people are wealthy enough to buy airline tickets to abroad destinations in the first place.
The 6th hypothesis is all about the ability and willingness of consumers to buy airline tickets. If a country is
problems of buying this online. If countries are poor, the customers of KLM are the rich in that country who are than relatively more rich as average wealthy people in Europe and have much better computer equipment. This hypothesis is described as follows;
H6a: Where consumers are wealthy and used to buy more expensive products online: the direct online share of the airline will be higher.
H6b: Where countries are poor and the rich have more to spend: these customers will more easily buy online.
Culture is the last country dimension that is expected to play an important role. From the five dimensions of Hofstede, two are expected to influence the channel choice. The first is individualism; as many
passengers travel in groups that cannot be booked via the website the part of the market that can be sold tickets via the website becomes smaller. In other words, the potential segments in the market that can make use of the website are different if a country’s culture is very collectivistic like China, Korea or Japan. The second dimension is risk avoidance. As the main reason for consumers not to buy online is the risk they expect when choosing this channel it is expected that countries with a high level of uncertainty avoidance will choose the direct online medium less easily. This leads to the following two hypotheses; H7a: Countries with a collectivistic culture are expected to book less via the internet site of the airline H7b: Countries with high uncertainty avoidance will book less via the internet site of the airline
2.9.3 Industry context hypothesis
The last hypothesis, number seven, focuses on the attention and power that the airline has in a market. The idea is that the higher the market share the higher the brand awareness and thus the trust that customers have to buy directly from the companies website (Aaker, 1996). Where the indirect channels often have a high share the customers trusts them and buys with a travel agent who advises to buy a ticket from a specific airline. Therefore the last hypothesis is formulated like this;
H8: The higher the market share of the airline the more tickets are sold through their own website
It might very well be that customers buy more often via the website of an airline because the ticket value is lower and the risk is also lower. Thus, for countries that sell a lot of tickets with a low ticket value (intra European traffic or tail-end traffic) the chance is likely that customers have no problem buying these tickets online. We therefore conclude with the eighth hypothesis as follows:
H9: A low average coupon value leads to relative high sales via the website
Above hypothesis visualized in a picture leads to the conceptual framework shown in picture 2;
Figure 2 Conceptual Framework
2.10 Control variables
One control variables is planned to be tested as well. Levying a booking fee might discourage customers to book online. If customers have to pay an additional fee to book online this might be a barrier leading the consumer directly in the arms of the travel agent.
3. Method
The research is mainly built around quantitative data. A set of 161 variables on 195 subjects were gathered in order to scope down later, but keeping the possibility to easily add a variable again. There are airline offices in 96 countries, 69 of these are destinations where direct flights of the airline operate to, 72 countries have a website. In Lithuania there is a website, however flights are sold of Lithuanian airlines. Ticket revenues were registered in 185 countries out of the 195 that were in the dataset.
3.1 Sources
Both internal airline and external resources where used to collect data. Two external sources where used, the Euromonitor and the Internet World Statistics (IWS).
3.1.1 Organizational context sources: BI portal
The airlines has an business intelligence portal (BI-portal) which is a data warehouse that contains the most important Key Performance Indicators (KPI’s) for the airline like revenues and coupons. These figures were collected per distribution channel for the year 2009 and from this data the average coupon value and the share per channel were also calculated. Out of the 196 countries that were selected via the
Euromonitor only 2 where not matched in the BI-portal.
This portal was also used to calculate the average coupon value. This is an internal source; nevertheless it reflects the possibility to sell more or less short haul flights compared to long haul flights. The more short haul possibilities the lower the average yield will be.
3.1.2 Organizational context sources: Payment team
A special department within the airline company is working on the online payment processes and options. This department keeps an overview of all payment options per website and the fees leveraged per country. Next to the payment method of Credit Card we also accept bank transfer, pay at the office/cash, pin, debit card, realtime online banking and Western Union. The master table shows per payment option (MPO) if it’s available and sums the total per country as a separate variable.
3.1.3 Organizational context sources: Country marketing plans (CMP)
The country marketing plans have a predefined format that each country has to fill in. It contains amongst others information on the marketing budget, e-commerce budget and division of this e-commerce budget over the different e-acquisition channels also elucidating the return on investment (ROI) of each channel. The channels are, from most used to least used; search, email, display, affiliation, met search and others. Because social media channels where not yet widely used within the airline company in 2009, these channels were not specifically inventoried. The number of staff working for a country is also available in the CMP, however it was difficult to tell exactly per employee how much time per week (% FTE) should be allocated to which country. The country action plans are requested by the e-sales managers of the
e-commerce department.
3.1.4 Organizational context sources: Central e-acquisition team
The advantage of online advertising above off-line advertising is that all clicks of the customer can be tracked. By adding tiny URL codes to website addresses all kinds of information can be gathered. Each channel is tagged as well as each page in the website including each step in the booking tool. If a customer reads an email and clicks on the email to go to the airlines website this is tracked and stored in a
webanalytics tool. If the customer finishes a booking either at the same time or in the future this can be linked back to the email of the airline read. Via a technique called floodlight tagging, all clicks are
measured via DART which allows to link a click or a booking to all acquisition channels a customer has seen or used. The last channel used by a customer will get the revenue assigned once a booking is made. Only for the channel Search and Metasearch worldwide reports were available. So only for these channels the ROI per country could be taken into account for the analysis. For some countries like Japan, Korea and the Philippines among them search was not handled centrally and thus the ROI figures were supplied to the local establishment only. After a few phone calls most of these countries ROI figures could be obtained afterwards, except for Japan where the media agency could not work with DART and has therefore not tagged the search advertisements in a correct way. For email there were two problems. The tagging in the
booking tool was not working for the smaller countries and the recalculation to the euro did not work for all countries. For the display channel reports were only available locally and quite some countries did not fill in their ROI in their country action plan.
Next to the ROI per channel the budgets spent per channel were needed. The information supplied by the establishment on their local spend per acquisition channel was added to with the money spent by the headquarters budget. The costs for sending out emails are paid centrally and can be calculated by multiplying each 1000 emails sent with a fixed cost of a few Euros.
3.1.5 Organizational context sources: Information Centre
Quite some report creation is outsourced to another company, the information centre which reports the main KPI’s and FPI’s per establishment (country) each month. From these reports the number of website visitors, booking tool visitors, booking tool conversion rates and email database subscribers were copied. Because countries differ in size and internet penetration it would be unfair to compare site visits and email database sizes with each other as such. Therefore two new variables were created that adjust for these differences. The email database size and site visits were divided by the number of people in a country that had an internet connection. So it says something about the coverage of the reachable group. The website reach variable (visitors/connections) should be read with care because one internet connection can be used by multiple consumers and these consumers can visit the website multiple times a year.
3.1.6 Country context sources: Euromonitor
The Euromonitor was accessed through the HES institute in Amsterdam. From this database 200 countries were selected to gather information on ICT, Economics and country descriptive. There are 15 variables on ICT information containing data about telephone and internet infrastructure, investment money spend and usage by the consumers in the specific countries for the years 1998 till 2009. The Economic variables total 36 and cover the wealth of the country and their consumers and focus on spending on travel and leisure for the years 2005 to 2009. The number of inhabitants and rates of exchange are part of the
country descriptive. The latter two are important in order to compare variables in a later stage between countries. At first glance the database shows that the Gross National Income of Thailand is 2.8 times bigger than the one of the USA. This seems strange, as the USA has 4.7 times the number of inhabitants as
Thailand and is 3.8 times bigger. The explanation is that the GNI for Thailand is expressed in Thai Bath that rated 34.3 to the US dollar in the year 2009. Part of the variables were also expressed in a per capita variable that corrects for the difference in number of inhabitants per country. Euromonitor did not contain data for each subject (country) for every variable that was selected, in that case it was left blank. For four countries the available information was so low and the relevance to the airline company was so small that these where omitted. These countries were American Samoa, North Korea, Moldova and Montenegro. Therefore the total number of subjects is 196. In total 45 queries were run on the Euromonitor database.
3.1.7 Country context sources: Internet World Statistics (IWS)
The website of the Internet World Statistics (IWS) is a very useful online database which keeps track on the usage of internet in most countries worldwide. Basically the database contains the number of internet connections and the number of inhabitants in most countries in the world. All 196 countries that where selected from the Euromonitor could be matched with the IWS database.
3.1.8 Country context sources: HTTP Watch
Via a free to download program called HTTP Watch each e-commerce manager in each establishment could test and report on the speed that the local airlines website needed to load in their country. Instructions were put in a word document and distributed. From 34 countries results came back. One problem was that some e-commerce managers handled a wider area as where they lived, they could not easily travel abroad to do this test. Secondly the instructions did not explicitly tell the testers to complete the whole booking process including payment where the testers could choose for a payment method that doesn’t require direct payment, like bank transfer. With the current results it is unknown which countries did finish all steps and which countries finished only the steps upto the payment step.
3.1.9 Country context sources: CIA fact sheet
The intelligence agency of the United States makes some very interesting facts and figures available through the internet. For 266 countries economic, geographical and political data is available. This data is updated every few weeks however for the Gini index that was used for this research. The recency of the information differed between 10 to 1 year. Also information for Gulf countries was unavailable.
3.1.10 Country context sources: ITIM international
Itim intenational is an international consultancy organization that uses the concepts of Geert Hofstede. Although the database of Hofstede and all replications focus on employees of companies and not
necessarily on general individuals we can assume that the working population of a country well represents the potential customers of an airline company. Although it is not very clear how recent the data is of this database this doesn’t harm the use of it because cultures do not change quickly over time.
3.1.11 Industry context sources: MIDT
Each time when a travel agent creates a booking through its central reservation system (CRS) this is recorded by the CRS. These records are copied each month on a tape and sold to for example airlines. The combined data of all CRS systems gives an enormously valuable view on the relative market position of each airline on different routes per country. For this research 3 variables where created. These 3 variables represent the market shares of our studied airline from a country to Europe, to the North Atlantic (USA + Canada) and to the whole World. For each country separately three queries had to be run this was a very labour intensive exercise. Because of the former reason 110 out of the 196 countries where queried.
3.2 Consolidating the data
All different data tables were consolidated into one excel sheet. This was only possible in Excel 2007 because the older versions of Microsoft Excel did not support more than 256 columns (2^8) and the total number of columns including country headings in front of most variables totaled 478. A check was done if each variable was matched on the right row height matching the country in the first column. Afterwards all
country heading columns were deleted and a pivot table was created to review the data. Some errors were found and corrected. In order to export all data to SPSS, first all formats like separators and currency indicators had to be deleted. Also in SPSS all variables had to be coded in order to make them fit. At the same time all variable names were added manually in SPSS.
3.3 Tests planned for execution
Based on the existing literature and the variables available for testing, we came up with a set of variables where we want to know the total effect of the set on the direct online share. The most logical test to use in this case is multiple regression. The less data available the lower the number of subjects and the less reliable the outcome of the research will be. For this same reason two tests will be run for each of the four channels. One for a set of variables with a number of subjects as high as possible, and one for a bigger set of variables which has at least a number of subjects above the threshold of 31 which is needed to do any statistical interference in the first place according the central limit theorem (Field, 2009 (p42). The hypothesis testing will finally measure the model as shown in figure 3.
The variables in red italics were planned to be obtained but could either not be gathered for a sufficient number of establishments, or they caused problems of e.g. multicollinarity during the regression tests. Reasons per variable will be described further in this document.
3.3.1 Variables explained
This paragraph describes how to read and interpret each variable that is used in the analysis.
The first variable in the organizational context is establishment size. This variable is expressed in turnover in Euros. This is the net turnover, meaning the money collected from the passenger for the ticket he or she bought. Airport taxes, Fuel surcharges and travel agent commissions explain the difference between gross and net. The following variables show the return on investments for search and metasearch and are calculated as follows. The costs involved in a specific acquisition channel are deducted from the revenues that are generated by that channel and the outcome is divided by these costs again. The last used
acquisition channel is the one used to determine which acquisition channel generated the sale.
For the email acquisition channel another measurement can be calculated; the email database penetration. This variable is a calculation where the non frequent flyer email subscribers to the general e-newsletter are divided by the number of website visitors. Comparing the sizes of the email database would not be fair as the number of people living in a country, as well as the level of internet penetration, differ.
The website reach is a calculation where the number of website visitors is divided by the number of people having an internet connection in a specific country.
Effectiveness measurements of the acquisition channels follow the execution measurements. The variable number of multiple payment options is a sum of the seven available options to pay online on the website: Credit Card, Debit Card, Banktransfer, realtime online banking, Western Union cash payments, Cash and Pin transactions. If one of these options is available this is counted as one and the sum is totaled. The second variable measuring the execution is the conversion ratio. Different from the conversion ratio this percentual variable measures how many people who started the booking process also finished it.
The next construct is the country context. For the internet penetration the number of people having an internet connection is divided by the number of people living in a country, the same calculation is followed for the number of broadband connections leading to the broadband penetration.
Wealth is calculated via the Gross Domestic Product. The definition of the GDP is the total of all produced goods and services in a country against market prices. The purchasing power parity variable recalculates the GDP for a country per capita to a value that gives you the same ability to buy a same set of products or services in a specific country. The third and last variable of the travel readiness construct is the Gini coefficient that says something about the distribution of wealth amongst the citizens in a country. Values go from zero to hundred where zero means no inequality between wealth amongst the citizens.
The market shares are calculated by dividing the total airline ticket sales sold via CRS/GDS systems divided by all ticket sales in that country via these systems.
The last variable is the coupon value which is expressed in euro’s. A person who flies from Warsaw via Frankfurt to New York and back has 4 coupons.
4. Results of analysis & Discussion
In this chapter we will analyze the data which was gathered. First the descriptives for each construct will be analyzed followed by an overview of the variables that could not be collected or did not yield yet any value to participate in the regression analysis. The outcome of the regression analysis will be describes and based on this some hypotheses can be accepted or rejected.
4.1 Descriptive statistics
In this paragraph we will describe the basic statistics of the variables. How is the data distributed amongst the different subjects (countries), what is the minimum, maximum, average and mode. If there are special outliers or special cases these will be mentioned as well.
4.1.1 Distribution mix – Channel shares
interested in the direct online share or what is sold via the airlines website. Secondly we would like to know if the direct online share is low what this means for the other channels. Are the shares of the call centers high, or that of the OTA’s or the travel agents. The Scandinavian countries and the United Kingdom all have direct online share percentages of above 20% whereas India and China have direct online
percentages of around 5%. Eastern African countries like Ethiopia have a share of around 10%.
4.1.2 Establishment size – Revenues
The establishment size is measured by the total revenue irrespective of channel. 70% of
the revenue is generated by the top 10 countries. The home market alone contributed up to 23% of the total turnover in one year. Most revenue is generated in Europe (67%), the USA, Canada, Africa and Asia contribute around 10% and the Caribbean and South America with no more than 3%. The revenue is obviously not normally distributed; it has a positive skewness of more than 8.
4.1.3 Resources commitment
Budgets are based on the revenues earned in each country. It is then upto the establishments how they divide their budgets. If they have a high commitment towards e-commerce they will probably allocate a big share of their marketing budget towards e-commerce. In general we could say that the commitment towards the e-channel is high because the average share spend on e-commerce of the total marketing budget is 35%, whereas the average share of the direct online channel is 15%. Establishments are definitely spending more money on this channel than it is currently worth.
4.1.4 Effectiveness of e-acquisition
The second set of constructs we wanted to measure is the effectiveness of the deployed resources. 45% of the traffic driven to the website comes from online advertising channels.
Figure 7: Site traffic divided per origin/acquisition channel
Email is the most affordable acquisition channel followed by Metasearch, Display, Search and Affiliation. The acquisition costs via the indirect off-line channel are much higher. This is especially due commissions even though these have come down from 12% to around 4% on an average booking value of 820 euro (=33 euro which is almost half the cost). The next paragraph focuses on the email channel.
4.1.5 Email
The email database and reach of internet users via email is described followed by the website visits and website reach. Then the ROI of e-acquisition channels and the ROI of the booking tool itself are covered.
Table 1 :Ranking of establishments based on email, revenue and website penetration
The average per country is around 85.000 email addresses. An interesting finding is that Spain and
Top 15 email size
Non FF Top 15 Revenue Penetration
1 UK USA Norway
2 Spain UK Switzerland
3 Italy Germany Spain
4 Germany France Denmark
5 Switzerland Norway Sweden
6 Canada Italy UK
7 Norway Japan Finland
8 Sweden Canada Greece
9 India Sweden Portugal
10 France Spain Belgium
11 USA Denmark Italy
12 Denmark Switzerland Ireland 13 South Korea Belgium UAE
14 Belgium China Singapore
Origin of bookings 38,197 39,992 48,160 97,225 200,658 520,768 38,197 39,992 48,160 97,225 200,658 520,768 Meta Search Affiliation E-mail Display Search
Switzerland have a relatively big email database compared to their relative position in terms of revenue contribution to the company. The opposite is true for China and Japan. The USA and Canada are relatively small compared to their revenue position. Looking at the email penetration we see that by average 10% of the website visitors of a country are also reading the newsletters. In South Korea and Spain around 30% of the visitors get the newsletter. For Brazil, the USA, China and Japan the opposite is true, less than 3.5% of the website visitors are subscribed to the local newsletter.
4.1.6 Website reach
The website reach can be measured for around 70 countries. It is interesting to see that the people in the some of the smaller islands visit the website frequently and that in total more visits are generated than there are people living in those countries. If we exclude the home markets and their colonies, which have very high reach percentages of 130 to 300+ percent, the average is around 10%. The Nordic countries, the UK and Switzerland score high and China, South Korea, India and Sudan score low. The single correlating effect on the direct online share is quite strong with Pearson correlation result of 0,523 (P<0,000).
4.1.7 Search & Metasearch ROI
We can measure return on investment (ROI) for the search acquisition channel and the metasearch acquisition channel and we can measure how many customers who started the booking process have actually finished, the latter is called the conversion rate of the electronic booking tool (EBT). The EBT conversion rate was available for 65 countries for. Nigeria and Ghana were excluded as outliers. These two countries make an enormous amount of fraudulent bookings which actually do not materialize and are no actual passengers. The conversion rate for search is high for African countries like Ethiopia, Libya, Uganda and Tanzania. The rate is low for Australia, New Zealand, Lithuania and Estonia. The range is quite small between 2% and 10%.
For 45 countries search was measured. Countries missing were among others Japan because the tagging is not in place. For many African countries search could not be measured as no online acquisition is done at all, most of these countries are just not mature enough for commercial internet marketing. The UAE has an ROI over 200, which is extremely high, South Africa and China are around 100. India and Turkey have an ROI of below 30, but Germany with exactly an ROI of 30 is also surprising. The low ROI for Germany might be explained by the size of the investment, the more you invest the lower the ROI; the investment of Germany is several hundreds of thousand euros, where the investment of the UAE and South Africa are just a few thousand euros. The correlation between Search and direct online share is positive but
negligible showing from the scatterplot in figure 5 to find out that the UAE, the UK and Norway are maybe biasing the data but the regression line is not very steep.
Metasearch ROI results were available for 68 countries. As metasearch has a very open structure, the chance is high that there is always some sales in a country. The risk for our analysis could be that really small investments of a few hundred euros lead to high ROI figures. However it does not have to mean that investing in those countries would result in the same ROI. To check if this assumption is correct we ran a correlation test.
