Abstract

Service quality and customer satisfaction are very important concepts that water enterprises must understand and measure from the customers' perspective to satisfy their needs. The main objectives of this research were to assess the level of customer satisfaction on urban water supply services of Southern Region, Ethiopia, and identify major determinants. Quantitative data were collected from 8,413 customers in seventeen towns, using a questionnaire based on the SERVQUAL model. Qualitative data were collected from customers via focus group discussion, and interviews were used with utility employees and officials. The results showed 47% of customers were satisfied with the water supply enterprise services, while 43% were dissatisfied for various reasons. The customer satisfaction score was below the acceptable level for all service quality dimensions, and understanding of customers, communication, and responsiveness were far below the benchmark. The correlation analysis revealed the existence of a monotonic, positive relationship between customers' total satisfaction and service quality dimensions. The proportional odds model indicated that total customer satisfaction was highly dependent on the nine service quality dimensions used in this research.

INTRODUCTION

Access to safe water and sanitation is a human right, as recognized in 2010 by the United Nations General Assembly, resolution 64/292 (United Nations General Assembly 2010). However, in 2015, only 91% of the global population used an improved drinking water source (96% of the global urban population). In other words, more than 660 million people worldwide used unimproved drinking water sources, nearly half of them in sub-Saharan Africa, because of low coverage and rapid population growth (UNICEF and World Health Organization 2015).

In Ethiopia, as in most developing countries, urban water supply infrastructure services are provided by state-owned water supply and sanitation bodies. The national policies for water supply are set by the Ministry of Water, Irrigation and Energy (Ministry of Water and Energy 2011). In 2006 the government adopted a Universal Access Plan (UAP), (Ministry of Water and Energy 2011), to achieve 98% access for rural water supply, and 100% for urban water supply and sanitation by 2012. The responsibility is given to the respective municipalities of the town, with the public involved as board members. According to DHV consultants BV; T&A consultant's plc (2003), public organizations and nongovernmental organizations (NGOs), often with the assistance of external support agencies, have constructed numerous water supplies in the country. At that time, the regions were operating over 600 urban and peri-urban water supplies. Many were not giving the service intended, however, and water supply services often remain unsatisfactory to users.

The water supply in Southern Nation Nationalities and Peoples' Region (SNNPR) – one of the 9 regional states in Ethiopia – urban areas is equally lacking water utility recovery and is not progressing towards its goal. Customers are not served at the standards required and their expectations are not met, due to problems in the sector's governance and drinking water coverage, and discontinuity of service because of limited maintenance. These and other factors are affecting efforts to provide better service levels (Southern Nation Nationalities and Peoples; Regional State, Ethiopia 2014).

Service quality and customer satisfaction are very important concepts that water enterprises must understand, to remain competitive in business and hence grow. Companies must know how to measure service quality and customer satisfaction from the consumers' perspective, so that they understand their needs and satisfy them.

Customer satisfaction or dissatisfaction (CS/D) is a function of the disconfirmation arising from discrepancies between prior expectations and actual performance (Bolton & Drew 1991). Customer satisfaction is associated with ‘value’ and is based on the amalgamation of service quality attributes (Athanassopoulos 2002).

Parasuraman et al. (1985) suggested that, when perceived service quality is high, that will lead to an increase in customer satisfaction. They support the idea that service quality leads to customer satisfaction, in line with papers by Saravanan & Rao (2007) and Lee et al. (2000), who acknowledged that customer satisfaction is based on the level of service quality provided.

Various models (like SERVQUAL, SERVPERF) have been developed to measure service quality. For this research the SERVQUAL model was preferred because it is the most widely used method of measuring service quality and customer satisfaction. Its main advantage is that it can be successfully applied in a variety of service industries, due to its ease of modification and adaptation to specific company requirements.

SERVQUAL can be used for complex analysis, as it provides a basis for control of the so-called non-financial variables such as customer satisfaction, and enables monitoring of service quality over time, identification of service components that are particularly good or bad, benchmarking of results with competitors, and the measurement of total customer satisfaction with a particular service (Đukić & Kijevčanin 2012).

The SERVQUAL model used ten dimensions to evaluate service quality, i.e., reliability, responsiveness, competence, access, courtesy, communication, credibility, security, understanding the customer, and tangibles. Most of these require consumers to have had some experience, to be able to evaluate levels ranging from ideal down to completely unacceptable. Parasuraman et al. (1985) further linked service quality to satisfaction by pointing out that when the expected level exceeds that perceived, the perceived quality is less than satisfactory and will tend towards totally unacceptable. When the expected level equals that perceived, perceived quality is satisfactory, however, and when expected service is below the perceived level, perceived quality is more than satisfactory and will tend towards ideal.

Franceschini et al. (2010) adopted Parasuraman et als (1985) SERVQUAL model for research used to assess water and sewerage quality, and customer dissatisfaction, and found it the most appealing model for such work. According to them, the ten variables affecting service quality and customer satisfaction are – courtesy, reliability, assurance, responsiveness, understanding/knowing the customer, access, competence, communication, tangibles, and credibility.

In this study the version of SERVQUAL used was customized to exclude credibility, because of the nature of the industry. In fact, the framework used was that of (Franceschini et al. 2010), adjusted to reflect the country and regional context. The main objective of the research was to assess the level of customer satisfaction and its determinants on urban water supply services in SNNPR, Ethiopia.

MATERIALS AND METHODS

Study area

The study was conducted in 17 towns in SNNPR that offer clean drinking water to customers. SNNPR is in southern and south-western Ethiopia, and accounts for more than 10% of its area. The population is estimated at nearly 18 million, around 20% of the country's population (Terefe & Welle 2008). It is overwhelmingly rural, with only 8% living in urban areas.

Figure 1 shows the 17 towns chosen for the study: Aleta Wondo, Arba Minch, Bonga, Dilla, Durame, Jinka, Tepi, Worabie, Butajira, Wolyta Soddo, Hawassa, Halaba, Hossana, Wolkite, Shonie, Mizan Aman and Yirgalem.

Figure 1

Map of the study area showing the towns in SNNPR, Ethiopia.

Figure 1

Map of the study area showing the towns in SNNPR, Ethiopia.

Research design

Survey research is concerned with assessing existing conditions or relationships, prevailing practices, beliefs, points of view or attitudes held, ongoing processes, influences felt, and developing trends (Best 1970).

Sampling and sample size

The target population was the customers (heads of households) of the water supply enterprises in the selected towns. In order to ensure that the sample was representative, towns were selected on the basis of population size and agro-ecological climatic conditions. A sample comprising 8,965 customers was invited to participate in the survey from the total of 134,727 in these towns, based on formula 1 below.
formula
(1)

where,

  • n = Sample size

  • N = Total number of households with tap water in the region

  • = Confidence level at α level of significance or (1-α) confidence level (Z0.025 = 1.96)

  • P = estimated satisfied population proportion (assumed to be 0.5 which is the optimal value for determining sample size)

  • d = relative margin (1%) of error between estimated and population actual proportion

Data collection

In order to collect valid and reliable data, questionnaires, interviews, and focus group discussions (FGDs) were used. The questionnaire was based on the service quality measurement dimensions courtesy, reliability, assurance, responsiveness, understanding/knowledge of the customer, access, competence, communication, and tangibles, as noted above. Instrument reliability was checked by calculating Cronbach's alpha, which was found to be 92.8% and thus suitable for the purpose. According to Bland & Altman (1997), Cronbach alpha values exceeding 70% are a good indicator of the internal consistency (reliability) of a data collection instrument. FGDs were conducted in each town with 6 to 10 people representing different community members, to elicit quality control data. Semi-structured interviews were also held with water supply office representatives in each town.

Data analysis

Descriptive statistics were used to estimate customer satisfaction levels, and Spearman rank correlation was employed to assess the association between total satisfaction and the nine service quality dimensions. A proportional odds model using SAS® (Statistical Analysis system) version 9.1 was used to identify the determinant service quality dimension explaining total customer satisfaction.

The proportional odds model formula (2) (Agresti 2002) is:
formula
(2)
where j is a total satisfaction category (strongly disagree (SD), disagree (D), neutral (N), agree (A), strongly agree (SA)), X is an independent variable vector (nine service quality dimensions), α is a cut point, and β is a logit coefficient vector.

Conceptual framework and operational definition of attributes

The conceptual framework of the variables included is presented in Figure 2.

  • Responsiveness- is the willingness of enterprises to help customers and provide quick service.

  • Reliability- refers to the extent to which the promised service is actually delivered.

  • Assurance- refers to the safety of the tap water

  • Courtesy- measures whether the enterprise cares for customers and the attentiveness with which this is done.

  • Tangibility- refers to the enterprises' facilities, machinery and personnel.

  • Access- refers to ways to contact the service provider

  • Competence- refers to the competence of the service provider's staff in technical and customer service matters

  • Communication- the use of appropriate communication tools when there is a need to furnish information to customers

  • Understanding/knowing the Customer- refers to the service provider's effort to understand customer needs

Figure 2

Conceptual framework of service quality dimensions for customer satisfaction.

Figure 2

Conceptual framework of service quality dimensions for customer satisfaction.

RESULTS AND DISCUSSION

Eight thousand nine hundred sixty five (8,965) questionnaires were distributed of which 8,413 were returned, a 90.6% response rate. According to Dillman (1978), a response rate of 70% and more is sufficient for a survey research, these results are thus sufficient for analysis.

Customer satisfaction with water supply services in SNNPR

Customers were asked to rate their total satisfaction with the water supply services provided and their responses are presented in Figure 3.

Figure 3

Customers satisfaction with water supply enterprise in SNNPR.

Figure 3

Customers satisfaction with water supply enterprise in SNNPR.

As can be seen, 46.90% of customers were satisfied with the drinking water supply services, 43.27% were dissatisfied, and the remaining 9.83% were indifferent. This implies that the level of satisfaction is below average, i.e., the proportion of satisfied customers is below 50%.

Some problems corroborating the low customer satisfaction result indicated in Figure 3 were identified through FGDs and interviews. The main problems obtained from the FGDs include: shortage of water, lack of proper and timely maintenance, poor fixture quality, technical (maintenance workers) skill gaps, and failure to give customers information regarding water supply interruptions, water quality problems, and bill-related issues.

The interviews with water supply office workers revealed that the water supply facilities in some towns were working beyond their design life. In some towns, system upgrades have started but were not completed on schedule, so that significant numbers of residents did not have water connections or sufficient pressure in their taps. Furthermore, the high population growth and fast expansion of the towns challenged the capacity of the water supply enterprises to meet customer demand.

The findings are also supported by Addis Zemen (2017). The main problems and complaints of Ethiopian urban water supply customers are: increased water demand because of expansion of towns, late completion of water project construction, incorrect calculation of population growth in towns, poor operation and maintenance of facilities after construction, and assignment of inexperienced/less qualified workforce for construction.

The FGDs, the water supply personnel interviews, and the literature all substantiate strongly the low customer satisfaction level result obtained from the primary data.

Customers satisfaction on the nine service quality dimensions

Table 1 is a summary of the customer satisfaction scores for the nine service quality dimensions. It is intended to indicate explicitly which dimension accounted for the lower satisfaction scores. According to several authors (Nielsen & Levy 1994; Market-Directions 2009; Sauro 2011), a reasonable comparison bench mark for comparing the means of a five point rating scale is at least 80% of the maximum score. On that basis, the bench mark used to interpret the mean results in Table 1 is 4 points.

Table 1

Average customer satisfaction scores on the service quality dimensions

Service Quality DimensionMeanStd. DeviationStandard scorea
Competence 3.28 1.0174 −0.71 
Assurance 3.26 0.9489 −0.78 
Courtesy 3.24 0.9680 −0.79 
Tangibility 3.21 0.9160 −0.86 
Reliability 3.05 0.9421 −1.01 
Access 3.04 0.9394 −1.02 
Communication 2.91 0.7320 −1.45 
Responsiveness 2.79 0.9542 −1.27 
Understanding Customers 2.60 1.0727 −1.31 
Service Quality DimensionMeanStd. DeviationStandard scorea
Competence 3.28 1.0174 −0.71 
Assurance 3.26 0.9489 −0.78 
Courtesy 3.24 0.9680 −0.79 
Tangibility 3.21 0.9160 −0.86 
Reliability 3.05 0.9421 −1.01 
Access 3.04 0.9394 −1.02 
Communication 2.91 0.7320 −1.45 
Responsiveness 2.79 0.9542 −1.27 
Understanding Customers 2.60 1.0727 −1.31 

aThe standard score is computed from the benchmark mean (4) and respective standard deviation.

The mean score for all service quality dimensions is below the benchmark. This means customers are dissatisfied on all service quality dimensions, which is consistent with the total satisfaction result shown on Figure 2. While all dimensions are below the benchmark level, the standard scores for understanding customers, responsiveness, and communication are far below, indicating that customers are more dissatisfied with these dimensions.

Correlation between total satisfaction and service quality dimensions

Correlation analysis was used to explore the association between total satisfaction and the nine service quality dimensions, so a non-parametric Spearman rank-order correlation measure was used – see Table 2.

Table 2

Correlation between total customer satisfaction and service quality dimensions

CorrelationReliabilityResponsivenessCompetenceAccessCourtesyCommunicationAssuranceUnderstanding customerTangibility
Spearman 0.61* 0.63* 0.55* 0.62* 0.62* 0.54* 0.67* 0.49* 0.65* 
CorrelationReliabilityResponsivenessCompetenceAccessCourtesyCommunicationAssuranceUnderstanding customerTangibility
Spearman 0.61* 0.63* 0.55* 0.62* 0.62* 0.54* 0.67* 0.49* 0.65* 

*Correlation is significant at the 0.01 level (2-tailed).

Table 2 indicates that, based on Spearman correlation cutoff point (Campbell & Swinscow 1996), total customer satisfaction has a strong positive correlation with reliability, responsiveness, access, courtesy, assurance, and tangibility. On the other hand, it has only a moderate positive relationship with competence, communication and understanding customer, which was statistically significant at the 99% confidence level. Generally the analysis indicated that there is a positive, monotonic correlation between total satisfaction and all service quality dimensions in this study.

Proportional odds model

A proportional odds model was used to assess the relationship between the outcome (total satisfaction) and explanatory variables (service quality dimensions). The score test for the proportional odds assumption (χ2 = 39.7, DF = 27, p-value 0.0546) is insignificant at the 5% level, indicating that the data satisfy the parallel lines assumption, although the test's p-value is very small (0.0546) and closer to the level of significance. However, the SAS/STAT® User Guide warns that the test may reject more often than it should, especially if the sample size is large or there are many predictor variables in the model stated by Elkin (2017). In this study the sample was large, so the p-value is very small. The model estimated how likely the response (total satisfaction) is to be in category j (response category) or above, as against a category below j (SA, A, N, D, SD), given the nine service quality dimensions.

Table 3 shows four cumulative logits (SA vs A, N, D, SD; SA, A vs N, D, SD; SA, A, N, vs D, SD; and SA, A, N, D vs SD). Odds ratios (ORs) are interpreted as the association between a unit score increase for that service quality dimension and being in a satisfied direction level of total satisfaction versus an unsatisfied direction. The coefficients are parameterized so that positive coefficients translate into higher chances of being satisfied in total satisfaction for every unit increase in the score of a given service quality dimension, compared to an unsatisfied level. In this model, intercepts represent the log odds of being in category j (SA, A, N, D), or a higher versus a lower satisfaction category.

Table 3

Proportional odds model parameter estimates with standard error and ORs

ParameterEstimateStandard ErrorP-ValueOR
Intercept (SA) −12.8281 0.1186 <0.0001 − 
Intercept (A) −9.0496 0.1422 <0.0001 − 
Intercept (N) −8.3276 0.1466 <0.0001 − 
Intercept (D) −5.3992 0.1773 <0.0001 − 
Reliability 0.4865 0.0425 <0.0001 1.627 
Responsiveness 0.4371 0.0402 <0.0001 1.548 
Competence 0.2814 0.0246 <0.0001 1.325 
Access 0.2466 0.0424 <0.0001 1.820 
Courtesy 0.1804 0.0425 <0.0001 1.198 
Communication 0.1217 0.0442 0.0059 1.129 
Assurance 0.3015 0.0464 <0.0001 1.352 
Understanding 0.1208 0.0221 <0.0001 1.128 
Tangibility 1.1188 0.0403 <0.0001 3.061 
ParameterEstimateStandard ErrorP-ValueOR
Intercept (SA) −12.8281 0.1186 <0.0001 − 
Intercept (A) −9.0496 0.1422 <0.0001 − 
Intercept (N) −8.3276 0.1466 <0.0001 − 
Intercept (D) −5.3992 0.1773 <0.0001 − 
Reliability 0.4865 0.0425 <0.0001 1.627 
Responsiveness 0.4371 0.0402 <0.0001 1.548 
Competence 0.2814 0.0246 <0.0001 1.325 
Access 0.2466 0.0424 <0.0001 1.820 
Courtesy 0.1804 0.0425 <0.0001 1.198 
Communication 0.1217 0.0442 0.0059 1.129 
Assurance 0.3015 0.0464 <0.0001 1.352 
Understanding 0.1208 0.0221 <0.0001 1.128 
Tangibility 1.1188 0.0403 <0.0001 3.061 

The respective intercept estimates are the log-odds of falling into or above a given response category, without the use of any service quality dimension. It can be seen in Table 3 that the intercepts are increasing, indicating the odds of being satisfied versus unsatisfied.

A single parameter, slope, describes the effect of respective dimensions on customer's total satisfaction level, and it is the increase in log-odds of falling into or above any category associated, compared with a unit increase for a given service quality dimension, holding all other dimensions at same level. In Table 3 all slopes are positive, indicating a tendency for the total satisfaction level to be highly likely to move in the satisfied direction as a service quality dimension increases by unit score.

When the OR exceeds 1, the total satisfaction is highly likely to be satisfied as all service dimension scores increase. For instance for the reliability dimension the OR is 1.627 (Table 3), for a unit increase in score of reliability the customers total satisfaction is 1.627 times the unsatisfied direction. Alternatively, a unit increase in reliability score the total satisfaction in the satisfaction direction is 62.7% higher than the dissatisfied direction. Ojo (2010) found that reliability contributes the most variation to customer satisfaction, but, in this study, tangibility, access, reliability and responsiveness had significant influence on total satisfaction, from the highest to lowest respectively. A study by Jayaramu et al. (2014) also indicated that reliability and responsiveness lead to higher consumer satisfaction.

Generally, customer satisfaction is significantly affected by reliability, responsiveness, competence, access, courtesy, communication, assurance, tangibility and customer understanding. The effects of tangibility, reliability, responsiveness, and assurance account for most effects, relatively, with respect to total customer satisfaction with water supply services.

CONCLUSION

Water supply utilities in Ethiopia, as in other developing countries, are publicly owned and customers do not have a choice of service provider. This study of the SNNPR water utilities has shown, through data analysis of nine service quality dimensions, that customer service quality satisfaction was 47%, which is below average. The major reasons for this were failure to treat and inform customers properly, expiry of design period, poor construction, use of low or poor quality materials, and poor utility employee capabilities. The utilities should, therefore, improve employees' capabilities and competence in all attributes used to measure customer satisfaction and upgrades the water supply systems. Water supply employees at this level need, particularly, to be trained in technical matters and customer handling techniques.

ACKNOWLEDGEMENTS

The authors would like to express their heartfelt gratitude to Arba Minch University for the facilities made available for the project, and Southern Nations and nationalities civil service bureau for providing financial support. Experts from the civil service bureau, each town's water enterprises, regional water bureau and Arba Minch University Institutional Transformation Directorate are also acknowledged for their contributions.

REFERENCES

Addis Zemen
2017
Why Towns Thirsty of Water
.
Addis Zemen News paper
.
Addis Ababa
,
Ethiopia
.
Agresti
,
A.
2002
Categorical Data Analysis
, 2nd edn,
Wiley & Sons
.
New Jersey
.
Athanassopoulos
,
A. D.
2002
Customer satisfaction cues to support market segmentation and explain switching behavior
.
Journal of Business Research
47
(
3
),
191
207
.
Best
,
J. W.
1970
Research in Education
, 2nd edn,
Prentice-Hall. Englewood Cliffs
,
New Jersey
.
Bland
,
J. M.
&
Altman
,
D. G.
1997
Statistics notes: Cronbach's alpha
.
BMJ
314
,
512
.
Bolton
,
R. N.
&
Drew
,
J. H.
1991
A longitudinal analysis of the impact of service changes on customer attitudes
.
Journal of Marketing
55
(
1
),
1
10
.
Campbell
,
M. J.
&
Swinscow
,
T. D. V.
1996
Statistics at Square One
, 9th edn.
BMJ Publ. Group
,
London
.
DHV consultants BV; T&A consultant's plc
2003
National Water Supply and Sanitation Master Plan
.
Dillman
,
D. A.
1978
Mail and Telephone Surveys, The Total Design Method
.
Wiley & Sons
.
New Jersey
.
Đukić
,
S.
&
Kijevčanin
,
V.
2012
Service quality as determinant of customer satisfaction
.
Journal of Economics and Organization
9
(
3
),
311
325
.
Elkin
,
E.
2017
Beyond Binary Outcomes: PROC LOGISTIC to Model Ordinal and Nominal Dependent Variables
. ).
Franceschini
,
F.
,
Galetto
,
M.
&
Turina
,
E.
2010
Water and sewage service quality: a proposal of a new multi-questionnaire monitoring tool
.
Water Resource Management
24
(
12
),
3033
3050
.
Jayaramu
,
K. P.
,
Kumar
,
B. M.
&
Prasanna
,
R. K. K.
2014
Customer satisfaction with domestic water supply in India – a study in Hubli city
.
Journal of Environment and Earth Science
4
(
9
),
105
116
.
Lee
,
H.
,
Lee
,
Y.
&
Yoo
,
D.
2000
The determinants of perceived service quality and its relationship with satisfaction
.
Journal of Service Marketing
14
(
3
),
217
231
.
Market-Directions
2009
Discussion Paper on Scales for Measuring Customer Satisfaction
. ).
Ministry of Water and Energy
2011
Urban Water Supply Universal Access Plan 2011–2015
.
Addis Ababa
,
Ethiopia
.
Nielsen
,
J.
&
Levy
,
J.
1994
Measuring usability: preference versus performance
.
Communications of the ACM.
37
(
4
),
66
75
.
Ojo
,
V. O.
2010
Customer Satisfaction: A Framework for Assessing the Service Quality of Urban Water Service Providers in Abuja, Nigeria. PhD Thesis
,
Loughborough University
,
Loughborough
,
UK
.
Parasuraman
,
A.
,
Zeithaml
,
V. A.
&
Berry
,
L. A.
1985
Conceptual model of service quality and its implications for future research
.
Journal of Marketing
49
(
1
),
41
50
.
Saravanan
,
R.
&
Rao
,
K. S.
2007
Measurement of service quality from the customer's perspective: an empirical study
.
Total Quality Management and Business Excellence
18
(
3
),
435
449
.
Sauro
,
J.
2011
How to Interpret Survey Responses: 5 Techniques
. ).
Southern Nation Nationalities and Peoples’ Regional State, Ethiopia
2014
Annual Report
.
Civil Service Bureau (unpublished)
,
Hawassa
.
Terefe
,
B.
&
Welle
,
K.
2008
Research-inspired Policy and Practice Learning in Ethiopia and the Nile Region (RiPPLE)
:
Working Paper 1: Policy and institutional factors affecting formulation & implementation of Sanitation and Hygiene strategy
,
Addis Ababa
,
Ethiopia
.
UNICEF and World Health Organization
2015
Progress on Drinking Water and Sanitation - 2015 Update and MDG Assessment
.
World Health Organization
.
United Nations General Assembly
2010
The Human Right to Water and Sanitation Resolution 64/292
.
United Nations
.