1
TELECOMMUNICATION INDUSTRY:
SELF-SERVICE TECHNOLOGY AS THE
SINGLE POINT OF CONTACT AND THE
CUSTOMER - FIRM RELATIONSHIP
Supervisor:
Dr. Flier
Date:
9th January 2017
MSc Executive Programme in Management
University of Amsterdam
Student:
Filip Lukić
ID:
10899650
2
Statement of Originality
This document is written by Student Filip Lukić who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
List of Content
Abstract ... 5 1. Literature Review ... 6 1.1 Introduction ... 6 1.2 Self-service Technology ... 6 1.3 SST-Customer Channels ... 71.4 SST and Customer Lifecycle in Telecommunication industry ... 9
1.5 SST and Customer Satisfaction ... 10
1.6 Discount and loyalty ... 11
2. Research question ... 13
3. Theoretical framework and Hypothesis ... 13
3.1 SST-only customer support and loyalty ... 14
3.2 The effect of discount on SST – loyalty ... 15
3.3 Theoretical Model ... 16
4. Research Design ... 18
4.1 Procedure introduction ... 18
4.2 Study type ... 18
4.3 Sample selection and data collection ... 20
4.4 Measurement of variables ... 23
5. Analytical Strategy ... 26
5.1 Recoding of variables ... 26
5.2 Reliability... 26
5.3 Computing Scale Means ... 27
5.4 Correlations ... 28 6. Results ... 29 6.1 MANCOVA ... 29 6.2 Regression ... 34 7. Discussion ... 37 8. Conclusion ... 44 List of Literature ... 46 Appendix 1 - Vignette 1 ... 50 Appendix 2 - Vignette 2 ... 51
4
Appendix 3 - Vignette 3 ... 52
Appendix 4 - Vignette 4 ... 53
Appendix 5 - Vignette 5 ... 54
Appendix 6 - Survey: English version ... 55
Appendix 7 - Survey: Dutch version ... 58
5
Abstract
Previous studies have found that by forcing self-service technology (SST) towards customers would have a negative impact on loyalty. This paper aims to extend this by studying elements that could positively affect the relationship between SST in customer support (CS) and the customer loyalty. The conceptual models aims to test whether offering stand alone Self-service customer support by mobile operators, without personal support, would have influence on customer satisfaction and retention. Furthermore this study aims to test whether providing discounts as compensation for self-service only, would have a positive influence on satisfaction and retention. This model contains two independent variables which are SST and Discount while the dependent variables are Satisfaction and Retention, those two variables present loyalty. Additionally a set of control variables are adopted, age, gender, educational level, income level and technological readiness. A vignette based survey study has been performed for to collect the data. In total 224 participants have participated in the survey, the sample population consisted out of consumers of telecommunication services predominantly from the Netherlands. The data has been analysed by performing MANCOVA, Post Hoc and Regression tests. The results showed that by providing SST as a single point of contact during the after sales phase, customer support, would have a negative impact on satisfaction and retention. On top results showed that by offering compensation discount levels of 10% and 20% would not be enough for to positively moderate this relationship. However results did show that by providing discounts at a level of at least 30% would positively moderate this relationship. Nonetheless by providing 30% discount, consumers would not become more satisfied and loyal compared to the status quo (SST + Personal CS). Yet, consumers would maintain the same level of satisfaction and loyalty, nothing more.
6
1. Literature Review
1.1 Introduction
Precisely one century ago the first roots of the self-service concept were introduced. Clarence Saunders brought in 1916 a new retail concept that would transform the way services were provided to consumers. He launched the first self-service grocery shop (Piggly Wiggly, Memphis). This was totally different from other groceries, as at that time the shop personnel would gather the goods for the shoppers. Saunders noticed that customers were an unused resource, which made him introduce the shop policy of customers serving themselves. Soon after, he patented this new concept and issued franchises to hundreds of groceries in the U.S. Many other firms and industries followed this way of driving B2C business. The self-service B2C concept could generally be defined as customers performing tasks by themselves, which previously were done for them by company personnel (Meuter et al., 2000; Salomann et al., 2006).
1.2 Self-service Technology
Customer Self-service advanced over time with the involvement of technological developments. This trend to transform the service economy in a mass-production-like way revolutionized the manufacturing industry, with services provided in large volumes for lower costs (Economist, 2004). Increasing technological innovations, with in parallel still high labour costs, make many firms choose to invest in integration of automated self-service technologies into their service delivery processes. The services provided to end customers are becoming more technologically driven with an increasing choice of customer self-service cannels. Self-service technologies (SST) are technological interfaces that enable customers to
7 take advantage of a service without any interference of real employees (Meuter, Ostrom, Roundtree, and Bitner, 2000). The introduction of SST’s enforces the potential of improving productivity and service quality while saving on costs. This drives firms to increasingly move to SSTs for offering their services (Weijters, Rangarajan, Falk, and Schillewaert, 2007).
1.3 SST-Customer Channels
Previous literature has defined a set of commonly used SST interfaces and their purposes. The three main interface channels defined are telephone, online and interactive kiosks. Telephone interface includes Interactive Voice Response (IVR) systems whereby users choose the information they need during a phone call and the system consequently answers with an automated voice (Bitner et al., 2002; Meuter et al., 2000). Online SST helps users to self-serve through online websites. Interactive kiosks are seen as a commonly used interface. These are stand-alone machines operated by the consumers (Bitner et al., 2002; Meuter et al., 2000). With the emergence of applications through smartphone and tablet devices we can add “Apps” as a fourth interface. The separation of Apps from Telephone and Online is due to the variety of extra features an App can provide compared to Telephone and Online (e.g. location based, interaction with 3rd party services, notifications, calls etc).
The literature defines three main purposes of SST interfaces. Customer support purposes contain mainly supporting customers on questions about all possible matters that arise in the after sales phase when consumers are actually using the product/service. Another purpose is the transactions which mainly involve payment of the product/service.
8 Self-Help is defined as a third purpose, which entails customers can inform themselves, arrange, adjust and design their product/service by themselves - within the boundaries of possibilities offered by the company (Bitner et al., 2002; Meuter et al., 2000).
In Figure 1 an overview is presented of currently used SST channels and their purposes, based on a review of the academic literature and my work experience with companies.
Figure 1: SST Customer Channels, Purposes and Examples
SST Interfaces/Channels
Telephone/IVR Online Interactive Kiosks Mobile Apps Customer Service - Order status - Flight info - General info - Account info - Package tracking - Troubleshoot - Bank ATM’s - Hotel/flight check-in/out - Account info - Troubleshoot - Invoice info Transact. - Tele- banking
- Prepaid top-up - Billing status - Online banking - Online shopping - Parking machine - Pay at pump - Banking - Invoice payment - Money transfer Self-Help - Information telephone lines - Change address - Distance learning - Change account info/settings - Tourist info - Blood pressure machine - Change account info/settings - Change product
(source: Meuter et al., 2000)
In my research I will research the relationship of Discount and SST only within the Customer service purposes of SST interfaces. This segment of the SST is not yet researched independently. In addition, the four different SST interface channels could be taken into consideration in comparing consumer attitudes toward different channels and the relationship to retention. P u r p o s e s
9
1.4 SST and Customer Lifecycle in Telecommunication industry
SST Innovations enable different industries to use SST in different ways. In some industries SST is part of the firm’s core service offered to customers, while in other industries SST is mainly part of complimentary services. In the telecommunications industry SST is increasingly taking an important role in the operator - customer relationship as operators increase their focus on self-service (Gupta, 2009; Van Riel et al., 2004; Wallin et al., 2009). The below graph shows customer touch points categorized into personal and SST contact during the customer lifecycle of Telco customers.
Figure 2: SST and Customer lifecycle
(source: T-Mobile; Vodafone; Telefonica; 2016)
The customer service unit takes over the customer relationship in after-sales when the sale is made and the consumer starts using the services. This phase in the customer relationship is marked as important in regards to satisfaction, loyalty and retention (Ernst, 2010; Lele, 1997, Saccani, 2007). The after-sales area of Telco’s is increasingly pushed towards SST customer touch point by Telephone IVR systems, user self-management via website, app or
10 self-service kiosks in Telco’s physical shops. Currently, customers still have the possibility to engage in personal contact with the company, call centers, telco shops as well as via email, chat and company forums. Customers currently still have the choice of using SST as well as personal customer service (T-Mobile; Vodafone; Telefonica; 2016).
Previous research has neglected to specifically study the telco industry regarding the relationship of offering SST only within the after-sales customer support phase and its effects on consumer retention. In my study I will study this relationship within the telco industry.
1.5 SST and Customer Satisfaction
As SSTs are considered beneficial from a firm’s perspective, more and more service providers actively “push” their customers toward SST channels (Langer et al. 2012; White et al. 2012). One of the methods for stimulating consumer usage of SSTs is by making the traditional full-service relatively unattractive through applying additional pricing for traditional full-service usage. Another increasingly used method is the complete replacement of traditional services with SSTs (Reinders, Dabholkar, and Framback, 2008). This forces customers to use SST. Customers however may be willing to rather avoid SSTs if they are not comfortable with using the technology (Meuter, Ostrom, Bitner, and Roundtree,2003).
Lin and Hsieh (2006) indicate that forcing customers to use SSTs could lead to a decrease in positive perception and evaluation of the firm’s services. Reinders (2008) shows that forcing SST’s usage can lead to negative attitudes in using SSTs and could additionally have behavioural consequences. Moreover, recent research shows that the value customers can receive from SST channels differs from personal service channels in a way that it does not
11 allow a total substitution of these channels (Kumar and Telang 2012). Customers that use SST are not satisfied by definition, yet are stuck with it (Buell et al. 2010). Selnes and Hansen (2001) state that forced usage of SST, by substituting it for personal service, can harm customer loyalty. Scherer at all (2015) however argue that customers are most likely to exit a service relationship when only one service channel, either SST or personal, is offered.
The main obstacle for SST is getting customers on board to try and use new SSTs for the first time. This mostly involves an important behavioural change by which old habits need to be overcome. It also requires the customer to change its behaviour in self-service situation. Customers become co-producers of the service, with responsibility for the delivery services and for their own satisfaction (Bendapudi and Leone 2003; Meuter and Bitner 1997).
Earlier research identified that by providing SST-only services would have a negative effect on customer loyalty. I will test this assumption during my study and expand this study by investigating the effects of discount.
1.6 Discount and loyalty
Wieseke et al. (2014) found that loyal customers on average obtain greater discounts, which in turn drives customer loyalty. This loyalty–discount cycle is particularly achieved if customer loyalty is based on price as opposed to quality and if customers and sales people have longer relationship. Previous research found that loyal customers could expect a price discount or better service as a reward for their loyalty (Dowling and Uncles 1997).
Other studies however focused on the price-related consequences of customer loyalty, showing that customer loyalty leads to lower price sensitivity (Guadagni and Little 2008; Srinivasan, Anderson, and Ponnavolu 2002), which results in higher price levels for products
12 or brands (Chaudhuri and Holbrook 2001). Many customer retention approaches are driven by loyalty programmes and customer discounts, yet some studies indicate that the driver of repurchase is the high-quality customer service and smartly managed formal and informal communications with customers (Vavra and Pruden, 1998).
Loyalty programs in general include discount features giving price discounts on selected products and services for its members. Such discount elements supply customers with immediate rewards for their purchases (Yi & Jeon, 2003). Discount features do drive customers to purchase products or use services which they usually buy in other stores. The discount features stimulate consumers to show loyal purchase behaviour to firms (Lewis, 2004). Hence, money saving (collecting) features can create switching costs to consumer. This can drive loyal behaviour as it would be less beneficial for customers to change the retailer or service provider (Zhang et al., 2000). Zhang et al (2000) suggest that direct discounts do not create switching costs, while delayed rewards do create it, implying that customers need to take efforts to earn their discount over a time period.
Earlier research neglected to study the effect of beneficiary elements, such as discount, to neutralize the negative effect of SST on loyalty. To add value to earlier literature I will study if and to what degree providing SST discount during the subscription live time has a positive effect on customer loyalty.
13
2. Research question
This paper aims to extend existing literature by studying the element(s) that could positively affect the relationship between self-service only customer support and customer loyalty. This will be evaluated by answering the following questions:
- What is the effect of SST only customer support on customer loyalty?
- What is the effect of discount on the relation between SST only customer support and customer loyalty?
- What is the effect of different discount levels on this relationship?
3. Theoretical framework
and Hypothesis
There are many studies trying to identify how customer loyalty can be influenced. Researchers define customer loyalty by retention, meaning that existing customers stay with the company over time. A key question for organisations is how to maintain the relationship with customers. In order to do so, companies are making ongoing trade-offs between cost efficiency strategies and customer retention initiatives (Weijters, Rangarajan, Falk, & Schillewaert, 2007).
Organizations are increasingly driving SST into their internal and external processes as these would bring operational and cost effective benefits (Langer et al. 2012; White et al. 2012; Weijters, Rangarajan, Falk, & Schillewaert 2007). Studies show that by adding SST channels to traditional customer service channels adds value to customer satisfaction (Kumar and Telang 2012; Scherer, Wünderlich and Wangenheim 2015). Yet organizations are increasingly replacing traditional channels with SST channels. This full replacement is argued to have
14 negative effects on satisfaction and loyalty (Lin and Hsieh 2006; Reinders 2008; Scherer at all 2015; Scherer, Wünderlich and Wangenheim 2015; Selnes and Hansen 2001).
In general, companies can maintain customer satisfaction by providing incentives like discounts (Dowling and Uncles 1997, Wieseke et al. 2014). The effect of discount on the relationship between SST and loyalty is neglected by current studies. This study aims to investigate if discount can be used by firms to drive customer acceptance of SST only CS to maintain loyalty. It will contribute to existing literature by adding discount as a possible positive influencer of the SST-Loyalty relationship. The study will also contribute to corporate management as it could open the black box of loyalty and discount to provide possible strategic options to bridge the relationship between SST only CS and retention.
3.1 SST-only customer support and loyalty
Previous studies found that introducing SST has a negative effect on loyalty. Lin and Hsieh (2006) state that forcing consumers to adopt SST only could lead to a reduced positive perception of the firm’s services. Reinders (2008) shows that forcing SST’s usage can lead to negative attitudes in using the SSTs and could additionally affect behavioural consequences. Selnes and Hansen (2001) state that forced usage of SST can harm customer loyalty. Hence, Scherer at all (2015) argue that customers are most likely to exit a service relationship when only SST is offered. In contrast to this, Liljander et al. (2006) state that a high degree of technological readiness of the consumers will result in SST to be perceived positively.
H1: SST only customer support has a negative effect on customer loyalty in terms of satisfaction and retention, compared to traditional SST + Personal support.
15
3.2 The effect of discount on SST – loyalty
Earlier research has shown that discount in general can have positive effect on customer loyalty towards firms. Loyal customers do expect to receive a price discount or a better service as a reward for their long term loyalty towards the firm (Dowling and Uncles 1997). Switching costs for consumers could increase when continued discounts are received from a particular firm (e.g. loyalty programs or subscriptions). This would make the choice for choosing another retailer or service provider unfavourable and costly for consumers (Zhang et al., 2000). It however seems that earlier research neglected to embed discount as a possible moderator in the relationship between self-service technology only, and the customer loyalty (Wünderlich and Wangenheim 2015).
This opens the gap for further research about how the negative effect of SST (provided only) on loyalty can be neutralized by elements such as discount. Based on findings from previous research (Dowling and Uncles 1997; Vavra and Pruden 1998; Wieseke et al. 2014) regarding the general positive relationship between consumer discount and firm loyalty, I aim to test if this positive effect would also apply to the relationship between SST and customer loyalty in the form of a moderating variable.
H2: Discount has a positive moderating effect on the relation between, SST only customer support, and customer loyalty in terms of satisfaction and retention.
16
3.3 Theoretical Model
This paper aims to extend existing literature by studying certain element(s) that could positively affect the relationship between self-service (SST) only customer support and customer satisfaction and retention within the mobile operator market.
Figure 3: The Conceptual Model
The conceptual models aims to test whether offering stand alone Self-service customer support by mobile operators, without personal support, would have influence on customer satisfaction and retention. Furthermore the aim is to test whether providing discount as compensation for self-service only, would have a positive influence on satisfaction and retention. This model contains two independent variables which are SST-Only and Discount while the dependent variables are Satisfaction and Retention, those two variables present loyalty. Additionally a set of control variables are adopted, age, gender, educational level,
17 The following hypothesis are constructed to test this model:
Hypothesis 1: SST only customer support has a negative effect on customer loyalty,
compared to traditional SST + Personal support.
Hypothesis 2: Discount has a positive moderating effect on the relation between, SST only CS,
and customer loyalty in terms of satisfaction and retention.
To test other effects on the main relationships as defined in the model, the following sub-hypothesis will be tested, each containing different discount levels. It will be tested whether higher discount levels would result in higher loyalty.
H2A: Discount level of 10% has a positive moderating effect on the relation between,
SST only CS, and customer loyalty.
H2B: Discount level of 20% has a positive moderating effect on the relation between,
SST only CS, and customer loyalty.
H2C: Discount level of 30% has a positive moderating effect on the relation between,
18
4. Research Design
4.1 Procedure introduction
This research was conducted as part of a MSc study at the University of Amsterdam. The research aims to measure the relationship between self-service only customer support and customer loyalty in terms of satisfaction and retention. On top it aims to measure the effect of discount on this relationship. The data for to test this relationship has been collected through a vignette survey study which was distributed to a selected sample population through online and print format during November-December 2016.
4.2 Study type
The type of study executed is of a quantitative nature. An experimental vignette survey has been performed. A quantitative vignette study contains two components: a vignette experiment as the core element, and a traditional survey for the parallel and supplementary measurement of additional respondent-specific characteristics, which are used as covariates in the analysis of vignette data (Atzmüller and Steiner, 2010). Nowadays in marketing strategy studies consumer and business decision-making is studied by various methods. For instance conjoint analysis can be used to understand respondents´ valuation of product attributes and their future purchase behaviour. Such a study gets close to understanding how people make decisions and what people actually value in products and services. However when people lack experience with a certain product or service, or when moderating variables are important, vignettes are mostly used to ease product and service evaluation by the respondents (Wason et al., 2002).
19 The vignette is actually a short, carefully constructed description of a person, object, or situation, representing a systematic combination of characteristics, which are presented to the respondent (Atzmüller and Steiner, 2010). It has been defined as short descriptions of a person or social situation that contain precise references to what are thought to be the most important factors in the decision-making or judgement-making processes of respondents (Alexander and Becker 1978, p.94). This method has mostly been used to study attitudes, perceptions and beliefs of people (Finch, 1987).
In social sciences the vignette technique is preferred over traditional survey methods because the latter of traditional studies is argued to provide less reliable and biased self-reports as a result of the direct and technical questions. However the vignette technique provides respondents with hypothetical situations which closely resemble real-life situations (Frey and Neckermann, 2013). Above all, by presenting a set of factors makes it possible for respondents to evaluate the whole situation rather than indicating their preference for isolated factors (Frey and Neckermann, 2013).
In this study, the vignettes contain hypothetically illustrated customer support and mobile subscription situations, with a set of real-life common issues that mobile operator customers face. Respondents will respond to the different vignettes, which should reveal their preferences for the different situations with regards to satisfaction and retention. Because in real life, the situation whereby self-service is offered as the only tool for customer support of mobile operator customers, doesn't exist yet in Benelux, Germany and Europe in general. Therefore the vignette survey is used as respondents would give more reliable judgements to such hypothetical situations. The used vignettes can be found in appendices 1-5. There are basically three types of experimental designs that are used: within-subjects design,
20 mixed design and between-subjects design. In this study a between-subject vignette design is used, which holds that every respondent will be exposed to a single randomly assigned vignette (Atzmüller and Steiner, 2010). This is with the reason to not bias the judgement of the respondents. As the vignettes hold certain discount levels, it would influence the decision if a respondent would be exposed to multiple vignettes with multiple discount levels. Therefore each respondent has been exposed to one single vignette rather than to all possible vignettes.
4.3 Sample selection and data collection
The survey was distributed to a selected sample population predominantly in online (Qualtrics) and secondly in print format. The online survey format consisted out of an English and Dutch version and the participants could switch between the languages. The main version was created in English and was afterwards translated into Dutch by the back-translation technique. Before the survey was launched, a test survey has been provided among a test group of five participants. According to their feedback the wording in some parts of the survey and vignettes has been adapted accordingly to have it better understandable for participants. Then, the survey has been conducted in a cross-sectional timeframe of three weeks during November and December 2016. The survey contained five different vignettes. Each participant would be exposed to one randomly assigned vignette. The further survey questions were the same for all vignettes and participants.
The selected sample group contained consumers of mobile operator services, predominantly from the Netherlands. Next to that the selected sample contained persons belonging to the "Millenials-Gen Y" and "Baby boomers" age groups. This was chosen for to hold consistency by building on earlier research conducted by Pagani (2004). This study used the same age
21 groups to research the difference in perceived quality of services and customer support of mobile network operators. Two sampling techniques were used. Firstly the convenience sampling technique was used, formal and informal network was approached through personal invitation. Next to that, snow-ball technique was used, the subjects that have received the personal invitation were kindly asked to further spread the survey among their own network. This worked out very well as the initial survey was spread to 94 persons by personal invitation, while the total number of recorded respondents counted 224.
Data collected
The actual collected data sample contained 224 cases, all customers of mobile operator services (controlled by survey question on operator). 61% of respondents were male, while 39% were female. Respondents were predominantly from the "Millenials-Gen Y" (66%.) age group followed by the "Baby boomers" (24%), "Generation X" (8%) and "Centennials- iGen-Gen Z" (2%.) age groups. The respondents' country of residence was mostly the Netherlands (88%), followed by persons from Belgium (10%) and Germany (2%). Regarding educational level, we can see that most participants hold a higher professional degree-HBO (39%), followed by University education (29%), vocational education-MBO (22%) and high-school education (10%). Regarding income level, 24% earns a monthly net income between €1.501 and €2.000, followed by 18%, earning €2.001-2.500. When it comes to the frequency of the usage of personal customer support in real life for whatever product or service, we can see that the largest group, 67%, uses personal customer support very rarely (no more than once or twice a year) followed by 29% that uses it occasionally (several times a year). The following figures shows a detailed view on the collected data.
22
Figure 4: Data collection in terms of number of respondents per element
Gender Age
Male 138 (61,6%) Centennials- iGen (<21 years) 4 (2%) Female 86 (38,4%) Millenials- Gen Y (21-39 years) 147 (66%)
Generation X (40-51 years) 20 (9%) Baby boomers (52-70 years) 53 (24%)
Total 224 (100%) 224 (100%)
Country of residence Vignettes
Belgium 21 (9,4%) SST Only 44 (19.6%)
Germany 5 (2,2%) SST Only + 10% Discount 45 (20.1%)
Netherlands 198 (87,9%) SST Only + 20% Discount 45 (20.1%)
SST Only + 30% Discount 46 (20.5%) SST Only + Personal contact 44 (19.6%)
Total 224 (100%) 224 (100%)
Educational level Income level
High school 22 (9,8%) €0-499 11 (4,9%)
Vocational education (MBO) 50 (22,3%) €500-1.000 18 (8,0%) Higher professional (HBO) 87 (38,8%) €1.001-1.500 16 (7,1%)
University education 65 (29,0%) €1.501-2.000 53 (23,7%) €2.001 – 2.500 41 (18,3%) €2.501-3.000 37 (16,5%) €3.001-4.000 28 (12,5% €4.001- 5.000 7 (3,1%) €5.001-6.000 2 (0,9%) €6.001-or more 11 (4,9%) Total 224 (100%) 224 (100%)
Personal CS usage Current mobile operator
Rarely (no more than once or twice a year)
153 (68,3%) KPN Vodafone
57 (25,4%) 67 (29,9%) Occasionally (several times
a year)
66 (29,5%) T-Mobile Other
43 (19,2%) 42 (18,8%) Frequently (several times
a month)
2 (0,9%) Regularly (once a month) 3 (1,3%)
23 4.4 Measurement of variables
The questionnaire answering format consisted out of closed ended questions which were scaled according to Likert scale. Making use of a Likert scale reduces the time a respondent need to fill out the survey (Jamieson, 2004). The independent variables consist out of SST and discount. These were computed into a variable consisting five different vignettes (groups). The dependent variables were retention and satisfaction, while the control variables include age, gender, income level, educational level and technological readiness.
Customer support type and Discount
The variables SST and discount were integrated into five different vignettes. The vignettes contained the factor of customer support type and discount level with different values across the vignettes. Following figure illustrates the different vignettes, appendices 1-5 hold the actual vignettes used in the survey.
Figure 5: The vignette groups
Factor Vignettes Description
Customer support type Self-service only SST only without discount and discount
Self-service only + 10% disc.
SST only, yet customers receive a discount for this
Self-service only + 20% disc. Self-service only + 30% disc. Self-service + Personal
Traditional CS, offering SST and personal contact, however no discount related to this
The customer support type "self-service + personal CS" has been added as a vignette, the aim with this is to be able to statistically compare "self-service only" with current real-life
24 customer support "SST + Personal" for to measure if there is a difference in positive or negative effect on loyalty.
Different discount levels are integrated for to make the hypothetical vignettes more feasible for the participants for making more reliable judgement. Next to that it gives the possibility to show if there are different effects on loyalty according to the different discount levels. These three particular discount levels were chosen to keep feasibility with real-life mobile operator discount levels, as mobile operator give discounts for subscription plans as their core product. However for non-core products and services, in general discount is not given (T-Mobile, Vodafone, 2016). Additionally percentage levels are used as it gives a general feeling of more benefit according to previous studies (Hardesty and Bearden, 2003).
Satisfaction
For the satisfaction variable a seven point likert-scale (ranging from 1 = ‘completely disagree' to 5 = ‘completely agree') was used. Measurement questions were adapted from Liljander et al. (2006). This measure contained four statements related to consumer satisfaction according to the shown vignettes. In general the higher the number on the scale, the higher the level is of consumer satisfaction.
Retention
For the retention variable a seven point likert-scale (ranging from 1 = ‘completely disagree' to 5 = ‘completely agree') was used. Measurement questions were adapted from Coelho et al. (2012), Gerpott et al. (2000). This measure contained three statements related to consumer retention according to the shown vignettes. In general the higher the number on the scale, the higher the level is of consumer retention.
25 Technological readiness
For to measure if the level of affinity with technology could affect the result of the study, the technological readiness control variable has been used with a seven point likert-scale (ranging from 1 = ‘completely disagree' to 5 = ‘completely agree') measure. Measurement questions were adapted from Liljander et al. (2006) and Lin (2006). This measure contained four statements related to technological readiness of the participants.
Other variables
Next to technological readiness control variable, the variables of gender (1='male' 2='female') and age were used as control variable. The age measure had an open ended outcome, when data was collected the answers were scaled according to the general age/generation groups (iGen, Gen Z or Centennials: Born 1996 and later, Millennials or Gen Y: Born 1977 to 1995, Generation X: Born 1965 to 1976, Baby Boomers: Born 1946 to 196) (Kumar and Lim, 2008), these scales were made for to hold consistency by building on earlier research on mobile operator services conducted by Pagani (2004) and to hold generalisability. The income level control variable was measured by a 10-point scale (ranging from '1=€0-499' to '€6.001 or more'). The educational level was measured with a 5-point scale (1 = ‘High School’, 2 = ‘vocational education (MBO), 3 = ‘Higher professional education’ (HBO), 4 = ‘University education'). Nevertheless two more variables were taken into the questionnaire, these were frequency of personal customer support usage in real-life (1='Rarely (no more than once or twice a year)' to 4='Frequently (several times a month') and country of residence (1='Netherlands', 2='Belgium', 3='Germany'). These were taken in for in case it could explain further context.
26
5. Analytical Strategy
Analysis were done by using IBM SPSS Statistics 22 software. Response omissions were minimum (11), and were replaced using Hotdeck (Myers, 2011). The main analysis technique used is the analysis technique of multivariate analysis of covariance (MANCOVA). This technique is an extension of analysis of covariance (ANCOVA). With MANCOVA the statistical differences on multiple dependent variables by an independent grouping variable, while controlling for a third variable, the covariate (multiple covariates can be used) can be tested. Covariates are added so that it can reduce error terms and so that the analysis eliminates the covariates’ effect on the relationship between the independent grouping variable and the dependent variables. Additionally a regression analysis has been performed for to test the model and relationships of variables. Furthermore, a cut off point of p = .05 has been used throughout the analysis.
5.1 Recoding of variables
The coding of counter-indicative items has been done in the Qualtrics, therefore it was not needed to again recode the variables in SPSS. Moreover this applied to one question of the variable retention ("I would terminate my contract with this mobile operator") and with one question from technological readiness ("Whenever something gets automated, I need to check carefully that the machine or computer is not making mistakes"). These were recoded in Qualtrics from 1= completely disagree <–> 7= completely agree to 7= completely disagree <–> 1= completely agree.
5.2 Reliability
The reliability of variables analysis has been performed for the variables technological readiness, satisfaction and retention. Cronbach’s alpha is used for to indicate the internal
27 consistency of all the items within the scale, and to evaluate whether certain questions should be deleted to improve the consistency. After running the reliability check, one question needed to be removed from the variable technological readiness ("Whenever something gets automated, I need to check carefully that the machine or computer is not making mistakes"), this is for to achieve a favourable Cronbach’s alpha (from 0.43 --> to 0.81). All the three variables now achieved a Cronbach’s alpha greater than 0.7, meaning a high level of internal consistency.
Figure 6: Reliability by Cronbach's Alpha
Variable Cronbach’s Alpha
Technological readiness 0.81
Satisfaction 0.89
Retention 0.86
5.3 Computing Scale Means
Certain scale items needed to be computed into new variables. This has been done for technological readiness (3 items), satisfaction (4 items) and retention (3 items). The means were calculated into a new variable using SPSS. Figure 7 indicates the means and standard deviations of the variables. Furthermore the five vignettes were computed into a single variable in SPSS called “CS Plan”, holding 5 item (groups) (view items in chapter “5.4 Measurement of variables”). The open ended answers on age were scaled into 4 age groups (iGen, Gen Z or Centennials: Born 1996 and later, Millennials or Gen Y: Born 1977 to 1995, Generation X: Born 1965 to 1976, Baby Boomers: Born 1946 to 196) (Kumar and Lim, 2008).
28 5.4 Correlations
A bivariate correlation test has been performed in SPSS. There is a high positive relation between satisfaction and retention (r = 0.780), meaning that when satisfaction increases by 1 on the scale, retention would increase by 0.78 on the scale. Moreover we see a tendency to a positive relation between CS plan - and satisfaction (r = 0.366) and retention (r = 0.305), the higher the item is on the CS plan scale, containing discount, the higher satisfaction and retention. Nevertheless we see a small relation between educational level and income level (r = 0.287) as well as with technological readiness (r = 0.243). Also there is a small relation between technological readiness and satisfaction (r = 0.226) as well as with retention (r = 0.152). Could be illustrated as the higher the level of technological readiness, the higher the level of satisfaction and retention.
Figure 7: Means, Standard Deviation, Correlations Means, Standard Deviations, Correlations
Variable M SD 1 2 3 4 5 6 7 8 1.Technological read. 4.94 1.26 (0.81) 2. Satisfaction 4.15 1.42 0.226** (0.89) 3. Retention 4.07 1.42 0.152* 0.780** (0.86) 4. CS Plan 3.01 1.41 -0.002 0.366** 0.305** - 5. Educational level 3.89 0.96 0.243** -0.074 -0.092 0.053 - 6. Income level 4.95 2.09 0.028 -0.115 -0.128 0.022 0.287** - 7. Age 2.36 0.75 -0.342** -0.154* -0.131 -0.05 -0.045 0.44** - 8. Gender 1.37 0.48 -0.168* -0.042 -0.071 -0.05 0.002 -0.032 0.77 -
** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed)
29
6. Results
6.1 MANCOVA
MANCOVA is an extension of analysis of covariance (ANCOVA). With MANCOVA the statistical differences on multiple dependent variables by an independent grouping variable, while controlling for a third variable called the covariate (multiple covariates can be used) can be tested. Covariates (control variables) are added so that it can reduce error terms and so that the analysis eliminates the covariates’ effect on the relationship between the independent grouping variable and the dependent variables.
In this study the independent variable is the Customer support plan (CS Plan), consisting of the groups "SST Only", "SST Only + 10%", "SST Only + 20%", "SST Only + 30%" and "SST + Personal". The two dependent variables are Satisfaction and Retention. The control variables included are Technological readiness, Age, Gender, Income level and Educational level.
By performing MANCOVA analysis with the mentioned variables, this study aims to determine whether there is a significant difference between Customer support plan (CS Plan) groups in relation to the level of satisfaction and retention, especially whether the SST Only group shows a significantly negative difference compared to SST + personal group. I also tested whether the SST + discount groups show significant positive difference related to satisfaction and retention, compared to the SST Only group.
Descriptive statistics
Adjusted means of CS Plan groups show that the satisfaction mean was lower in the SST Only group (M = 3.51, SE = 0.21) compared to the SST + Personal group (M = 5.02, SE = 0.20). The groups that contain SST + discount level show a higher satisfaction level when the discount is
30 higher. This counts also for the retention levels. Here it is observed that the retention level was less in SST Only group (M = 3.61, SE = 0.22) compared to the SST + Personal group (M = 4.75, SE = 0.21). Table below gives an overview.
Figure 7: Adjusted means
Adjusted means SST Only SST + 10% SST + 20% SST + 30% SST + Personal
Variable M SE M SE M SE M SE M SE
Satisfaction 3.509 0.211 3.770 0.205 4.043 0.214 4.402 0.209 5.021 0.207 Retention 3.605 0.220 3.732 0.214 3.799 0.223 4.442 0.218 4.752 0.216
MANCOVA Model
After adjustment of the control variables, the multivariate test showed the overall model for the CS plan variable to be statistically significant. The Wilks' Lambda test gave F(8, 366) = 4.248, p < .05; Wilks' Λ = 0.837; partial η2 = 0.085, meaning that there is a significant differences in the CS Plan groups. The model also shows that the control variable Technological readiness is statistically significant, the Wilks' Lambda test gave F(2, 183) = 6.044, p < .05; Wilks' Λ = 0.938; partial η2 = 0.062. The other control variables were not significant (p > .05). Therefore there is no significant impact of age, gender, educational level and income level.
Observing the Tests of Between-Subjects Effects gave a statistically significant difference in Satisfaction between the CS plan groups, F(4, 184) = 8.039 p < .05, partial η2 = .149. The partial η2 indicates that there is a high effect size. It gave as well a statistically significant difference in Retention between the CS plan groups, F(4, 184) = 5.358 p < .05, partial η2 = .104. However the partial η2 indicates that there is a moderate effect size. This concludes
31 that for both variables Satisfaction and Retention the CS plan groups have a statistically significant main effect.
For Technological readiness there is a significant main effect for Satisfaction variable, F(1, 184) = 11.404 p < .05, partial η2 = .058. Yet the partial η2 indicates that there is a low effect size. However there is no significant main effect for Retention (p > .05). Regarding the other control variables, there was no significant main effect on both Satisfaction and Retention variables (p > .05).
Figure 8: Results ANOVA test for Satisfaction DV
Variable SS DF MS F η2 Sig. Age 0.719 1 0.719 0.434 0.002 0.511 Gender 0.121 1 0.121 0.730 0.000 0.788 Educational level 5.323 1 5.323 3.210 0.017 0.075 Technological readiness 18.907 1 18.907 11.404 0.058 0.001 Income level 0.696 1 0.696 0.420 0.002 0.518
Customer Support Plan 53.314 4 13.329 8.039 0.149 0.000
Error 334.847 189 1.772
Figure 9: Results ANOVA test for Retention DV
Variable SS DF MS F η2 Sig. Age 1.366 1 1.366 0.758 0.004 0.385 Gender 0.265 1 0.265 0.147 0.001 0.702 Educational level 4.168 1 4.168 2.313 0.012 0.130 Technological readiness 6.678 1 6.678 3.706 0.020 0.056 Income level 0.978 1 0.978 0.543 0.003 0.462
Customer Support Plan 38.621 4 9.655 5.358 0.104 0.000
32
Post-Hoc Test
To further test the differences between the groups and to reveal whether SST Only has a negative effect on the dependent variables, and whether discount could moderate that relationship, a multiple comparison test is performed for the variable CS Plans through Tukey post-hoc tests. This shows the differences between the CS Plan groups related to satisfaction and retention. Again, the multivariate test gave that CS plan variable holds statistically significant for the main effect regarding the whole model. The Wilks' Lambda test gave F(8, 246) = 3.342, p < .05; Wilks' Λ = 0.814; partial η2 = 0.098.
Satisfaction DV
The Tukey post-hoc tests revealed a statistically significant difference between" SST only" and "SST + Personal" groups (p = .000) in satisfaction. It shows that "SST Only" produces a significantly lower satisfaction level compared to "SST + Personal". As the satisfaction level was 1.49, 95% CI [-2.32, -0.65] lower for "SST Only" compared to the "SST + Personal" customer support, implying that "SST Only" has a negative impact on satisfaction, compared with traditional customer support offering "SST + Personal" support.
Furthermore it shows that there are no significant difference between SST Only and "SST only + 10% discount" (p = .848) and "SST only +20% discount" (p =.572) groups. This shows that when adding discounts of 10% and 20% to "SST Only", it would not change the impact on satisfaction, neither positively nor negatively compared to "SST Only" without discount.
However, results show a statistically significant difference between "SST only + 30% discount" and " SST only" (p = .012) in satisfaction. "SST only + 30% discount" produces a significantly higher satisfaction level than "SST Only" without discount. As the satisfaction
33 level was 0.98, 95% CI [0.15, 1.82] higher for "SST Only + 30%" compared to the "SST Only" without discount.
On top of that there is no significant difference in satisfaction between "SST Only + 30% discount" and "SST + Personal" (p = .449) customer support. This implies that when adding discount of 30% to "SST Only", it would change the impact on satisfaction positively compared to "SST Only" without discount, while there will be no difference in satisfaction between "SST Only + 30% discount" and traditional customer support "SST + Personal". This could be interpreted that by offering 30% discount for SST Only, this would close and compensate the negative satisfaction gap between SST only and current CS offering, SST + Personal.
Retention DV
Regarding Retention, the Tukey post-hoc tests revealed a statistically significant difference in between "SST only" and "SST + Personal" groups (p = .003). It shows that "SST Only" produces a significantly lower retention level compared to "SST + Personal". As the retention level was 1.15, 95% CI [-2.00, -0.29] lower for "SST Only" compared to the "SST + Personal". This means that "SST Only" has a negative impact on retention, compared with traditional customer support offering "SST + Personal" support.
Furthermore it shows that there are no significant differences between "SST Only" and "SST only + 10% discount" (p = .989) and "SST Only +20% discount" (p = .988). This implies that when adding discounts of 10% and 20% to "SST Only", it would not change the impact on Retention, neither positively nor negatively compared to "SST Only" without discount.
However, results show a statistically significant difference in Retention between "SST only + 30% discount" and " SST only" groups (p = .030). "SST only + 30% discount" produces a
34 significantly higher retention level than "SST Only CS". As the retention level was 0.91, 95% CI [0.05, 1.76] higher for "SST Only + 30%" compared to the "SST Only" without discount. The higher retention mean indicates that people tend to be more loyal.
On top of that there is no significant difference in retention between "SST Only + 30% discount" and "SST + Personal" (p = 0.937) customer support. This shows that when adding discount of 30% to "SST Only", it would change the impact on retention positively compared to "SST Only" without discount. No difference is shown between "SST Only + 30% discount" and traditional customer support of "SST + Personal" support. This could be interpreted that by offering 30% discount for SST Only, would close and compensate the negative retention gap between SST only and current CS offering, SST + Personal.
6.2 Regression
Linear regression analysis is performed for a general understanding of the model and relationship between the variables.
Satisfaction DV
Hierarchical multiple regression was performed to investigate the ability of Customer support plan to predict levels of consumer satisfaction, after controlling for gender, age, technological readiness, income level and educational level. In the first step of hierarchical multiple regression, five predictors were entered: gender, age, technological readiness, income level and educational level. This model was statistically significant F (5, 188) = 3.24; p < .05 and explained 7.9 % of variance in Satisfaction.
After entry of Customer support plan at Step 2 the total variance explained by the model as a whole was 22% F (6, 187) = 8.35; p < .05. The introduction of Customer support plan
35 explained additional 13% variance in Satisfaction, after controlling for the mentioned control variables (R2 Change = .13; F (1, 187) = 31.26; p < .001). The final model showed that two out of six predictor variables were statistically significant, with CS plan having a higher Beta value (β = .37, p < .05) than Technological readiness (β = .25, p < .05). This means that when CS plan increases by one, thus more discount, the satisfaction will increase by 0.37. Additionally when the Technological readiness level of people increase by one, the satisfaction will increase by 0.25.
Figure 10: Hierarchical Regression Model of Satisfaction
Variable R R² R²-change B SE β t Step 1 .29 .08 Age -.12 .15 -.06 -.78 Gender -.00 .21 .00 -.01 Technological readiness .27 .09 -.24** 3.02 Income level -.05 .11 -.11 -1.49 Educational level -.17 .11 -.11 -1.48 Step 2 .47 .22 .14 Age -.08 .14 -.04 -.62 Gender .06 .19 .02 .33 Technological readiness .28 .08 .25** 3.44 Income level -.03 .05 .05 -.71 Educational level -.19 .10 -.13 -1.86 CS Plan .37 .07 .37*** 5.59
Note. Statistical significance: *p <.05; **p <.01; ***p <.001
Retention DV
Hierarchical multiple regression was performed to investigate the ability of Customer support plan to predict levels of consumer retention, after controlling for gender, age, technological readiness, income level and educational level. In the first step of hierarchical multiple regression, five predictors were entered: gender, age, technological readiness, income level and educational level. This model was not statistically significant (p > .05) and explained 5.5 % of variance in Retention.
36 After entry of Customer support plan at Step 2 the total variance explained by the model as a whole was 15% F (6, 193) = 5.25; p < .05. The introduction of Customer support plan explained additional 9% variance in Retention, after controlling for the mentioned control variables (R2 Change = .09; F (1, 187) = 19.44; p < .001). The final model showed that two out of six predictor variables were statistically significant, with CS plan having a higher Beta value (β = .30, p < .05) than Technological readiness (β = .16, p < .05). Meaning that when CS plan increases by one, thus more discount, the retention will increase by 0.30. Additionally when the Technological readiness level of people increase by one, the Retention level will increase by 0.16.
Figure 11. Hierarchical Regression Model of Retention
Variable R R² R²-change B SE β t Step 1 .06 .03 Age -.11 .15 -.06 -.75 Gender -.13 .21 -.04 -.61 Technological readiness .17 .09 .15 1.93 Income level -.06 .05 -.09 -1.16 Educational level -.16 .11 -.11 -1.38 Step 2 .14 .12 .09 Age -.09 .14 -.05 -.61 Gender -.08 .20 -.03 -.37 Technological readiness .19 .09 .16* 2.17 Income level -.05 .05 -.07 -.94 Educational level -.18 .11 -.12 -1.65 CS Plan .30 .07 .30*** 4.41
37
7. Discussion
The results from the statistical analysis show that there is support for the hypotheses 1: SST
only customer support has a negative effect on customer loyalty, compared to traditional SST
+ Personal support. It concludes that in case mobile operators would impose self-service only
in the customer support, without giving an option to reach out to personal contact, customers would become less satisfied and would become less loyal to the company. It seems that customers are significantly more satisfied with the current customer support format whereby they could reach out to both self-service and personal contact. This result actually confirms findings from previous studies. Even more, findings from this study gave that self-service technology as a single point of contact without any compensation would dissatisfy all segments of the population and trigger them to become less loyal to their mobile provider. This study also shows that the age, gender, educational, income level and technological readiness differences among people didn't change their attitude towards self-service only. Furthermore, these findings show that in real-life most of the people (67%) use the personal customer support only once or twice a year, followed by 29% of people that uses it only a couple of time a year. This doesn't even count for only mobile operators, yet for all the products and services people use in total, beyond telecommunication services. A general conclusion is that overall, people are not reaching out often to whatever personal support of whatever branch. It seems that people, even if they don't use the personal support very often, want to keep this as a save harbour in case something happens. This could actually be further researched to see what are the reasons behind this. Are people insecure with mobile services? Did they have bad experiences and therefore want to keep personal support? Or are people perhaps still not sufficiently technologically advanced to
38 entirely trust technology over personal contact, while the society is rapidly pushed towards automatization and robotization?
Therefore it is a good next step to measure if these feelings are entirely true, or whether it could be influenced by compensation such as discount incentives.
The second hypotheses stated that a combination of SST only support and discount would unconditionally positively influence loyalty. Hypothesis 2: Discount has a positive moderating
effect on the relation between, SST only CS, and customer loyalty in terms of satisfaction and
retention. Findings of this study rejected this hypothesis as discount is not always by default
positively influencing the relationship. The findings also show that in case mobile operators would provide 10% or 20% discount to customers for using only self-service technologies in customer support, this would not positively change customers' satisfaction and retention. It seems that people perceive 10% and 20% as insufficient for leaving their right for personal customer support, even independently of personal characteristics such as age, gender, education, income and technological readiness. These findings therefore consequently reject hypothesis 2A (10% discount) and 2B (20% Discount).
Nevertheless findings point out that in case the mobile operator would provide 30% discount for SST Only, customers would stay satisfied and loyal. Even more, by providing 30% discount it would improve the relationship between SST Only and satisfaction. Customers would not change their attitude compared to traditional support, while it would improve compared to SST only without discount. This makes that hypothesis 2C: Discount level of 30% has a
positive moderating effect on the relation between, SST only CS, and customer loyalty is
supported. Customers would accept SST only support in case they get 30% discount. Discount of 30% would trigger people to drop their right for personal contact and accept
39 self-service technology as their single point of contact. This is a significant contribution to existing literature as this could open the black box of self-service technology as single point of customer service and the customer satisfaction and loyalty levels. At a rate of 30%, customers tend to be satisfied and loyal even if they are not able to interact personally with their mobile service provider. (While people nowadays tend to use their Smartphone as an integral part of their daily lives.)
The model of satisfaction and loyalty could be further explained (R2) with additional variables beyond this study. An extension could be made by the variable "Optimism" of people. More optimistic people tend to be more optimistic about future and tend to be happier and more satisfied. Such people would be more capable to believe that issues could work out positively eventually (Alarcon, Gene, Bowling, Nathan, Khazon, 2013). The variable "Social support" could additionally be used, this makes people be able to cope more effectively, manage problems and feel better. This consists out of the perception that a person is cared for and is assisted by the society. This could be further distinguished between "Emotional" and "Informational" support which are closer to personal and SST only customer support (Langford, Bowsher, Maloney, Lillis, 1997). These variables are more part of social and psychological studies, however they may be able to further explain the model (increase R2). Such variables make a deep dive into the different customer types/personal characteristics of people. Yet also it goes deeper into the overall life attitude of people, this could show different influences on satisfaction and loyalty to particular situations such as the ones researched within this study.
40 This study reveals that people are somehow stuck between rapid technological advancement of services and their basic human need for personal interaction. Even the Netherlands, which is one of Europe's leading countries in adaption of new technologies (World economic forum, 2016), where people are early adopters of new technology. it seems that people are not yet ready for a customer-company relationship interface solely based on technology. People feel comfortable to have the possibility to make use of both human contact and technological self-service. Having a human to interact with about issues gives a save feeling to people that it will be solved. However in reality telecom firms are cutting costs in their customer support. Next to continuous improvements of self-service technology channels, operators are hiring students and temporarily workers in customer support. In reality this means that the focus of operators is not 100% on offering the highest quality of personal support. This results in cases whereby customers are not sufficiently supported during one support interaction, yet several interactions are needed during a time span of days or weeks. This is neither cost efficient for the firm nor satisfactory for the customer. On the other hand current customer technology channels are still in rapid development and are not yet without mistakes. This makes people still distrust support services that are solely technologically driven. Nowadays with whatever (branch) services people use, there are still mistakes, errors or days when systems are completely out of order (e.g. online banking) (NU, 2015). On top companies are not sharing much information with customers about the source of system failure, and even tend to believe customers are too demanding regarding technological capabilities (De Telegraaf, 2014). This makes customers more uncertain about what is happening, whether hackers have attacked their data/services or whether the firm is not capable enough and is making mistakes. Such factors certainly stresses the customer relationship with technology (De Telegraaf, 2014). People have the
41 aim to solve their issues as efficient as possible, without much hassle. Personal services and technological services could be more effectively integrated by introducing technology in a controlled manner by managed phases. In first phase as a complimentary service, when offering personal contact, agents should educate and enrich customers on the capabilities of solving issues with the firm's technology channels. This would make customers at least try such services in next phase. In such a method the SST channel would be complimentary at first contact, yet it could become the main method of service in next phase. However this can only work in case the company as well as its support agents have the willingness to take more time and patience for to let customers adopt SST. Currently telecom firms impose time limit targets to their support agents when speaking to customers (T-Mobile: 6 minutes), and agents are reviewed according to this. The relationship could work if technological solutions are fully capable to operate without mistakes, for to not cause distrust in the customer-technology relationship. On top it is important that issues are solved within the first point of contact, this will build trust and satisfaction.
This study provides important implications to managers of mobile operators, and as well as for managers of firms that are heavily investing into self-service technologies. By introducing SST as standalone point of contact this study suggests it could have serious negative influence on customers. It is a good question whether the cost and process efficiency benefits of self-service technologies could stand against unsatisfied and declining customers. This study shows that customers nowadays are not easily accepting every type of discount. For low margin markets, discount levels such as 10% and 20% have much impact, while it would not be enough for customers to stay satisfied and loyal. Managers should take this as a learning as it could damage their customer base by introducing self-service technologies
42 unconditionally, especially in case of single point of contact. However this study shows that at a level of 30% discount, customers would stay satisfied and loyal. However an important fact is that with 30% discount, consumers will not become more satisfied and loyal compared to the status quo (SST + Personal CS). Yet, consumers will maintain the same level of satisfaction and loyalty, nothing more. Such findings are important management implication as it offers solid ground for the development of customer support strategy, customer communication strategies, as well as general strategy of firm level approach. Managers are now able to make tradeoffs and calculate how sinful it would be to implement SST only against discount levels of at least 30%, from a financial perspective. However it provides additionally a good tool to influence attitude of existing customers, in case self-service technologies are implemented. The retention of existing customers is nowadays high on the agenda of telecommunication operators, as well as maintaining revenue as revenues of the telecommunication market are declining, and as well in coming period (NRC, 2016). However this study shows that customers perceive customer support as a very important aspect of using the mobile services, even if customers almost never reach out to personal customer support they seem still to be attached to it. The customer support relationship seems to still be a sensitive topic for consumers. Managers should take this in consideration when making strategy, to prevent on customer strategy mistakes and financial losses.
Next to implications for businesses, this study offers insights to existing literature on the topic of self-service technology. The outcome of this study extends existing literature by offering a solution for bridging the negative relation between person's satisfaction and self-service technologies when offered as single point of contact. This study shows that with discount levels as high as 30% would neutralizes customers' attitude towards SST from a satisfaction and retention perspective. Therefore discount can be used as an influencing
43 variable, however not unconditionally. This study reveals that discounts below 30% would not affect consumers' satisfaction and retention. Furthermore, studies can build on this study by further researching consumers' attitudes towards personal as well as self-service technologies from a social science perspective. In times when the society is imposed towards automatization and robotization, it is important to understand the relationship between humans and machines.
