Water Resources Management (2022) 36:3481–3503

https://doi.org/10.1007/s11269-022-03200-6

Customer‑Driven Water Supply Systems: Synergizing System

Reliability and Customer Satisfaction with Bowtie Analysis

Husnain Haider1 · Majed Alinizzi1 · Md. Shafiquzzaman1 · Saleem S. AlSaleem1 ·

Mohammad Alresheedi1 · Rehan Sadiq2

Received: 24 February 2022 / Accepted: 23 May 2022 / Published online: 1 June 2022
© The Author(s), under exclusive licence to Springer Nature B.V. 2022, corrected publication 2022

Abstract
Conventional interview surveys for assessing customer satisfaction require supplementary
resources and yield a low response rate. Covering fewer customers who generally face ser-
vice interruptions is another limitation that hinders the identification of the root causes of
failure. The survey findings might mislead the utilities by showing unrealistic satisfaction
levels. Customer dissatisfaction risk subjects to the time between the complaint registration
and the utility response pending full resolution of the issue. In the present research, bow-
tie analysis (BTA) reveals an innovative synergetic correlation between system reliability
and customer satisfaction with water utilities. Using complaints reported due to service
failures, the fault tree analysis (FTA) assesses the reliability of system components (pres-
sure, water quality, and structure). On the opposite side of the bowtie, event tree analysis
(ETA) evaluates the efficiency of the utility response and translates it into customer dissat-
isfaction risk. Exploring over 1500 registered work orders in a utility operating in British
Columbia (Canada) reveals that 80% of complaints originated from structural (primarily
service connections) issues. Simulating the BTA for five improvements showed that eco-
nomical proactive maintenance and an efficient complaint response mechanism improve
the system reliability from 58 to 86% and reduce the cumulative risk by 62%. The remain-
ing risk reduction requires expensive improvements, e.g., source change and conventional
treatment installation, which require customer approval based on willingness to pay. The
study reveals that a complete bowtie analysis effectively integrates both the FTA and ETA
and provides a novel customer-driven decision-support tool for water supply systems and
other elsewhere applicable.

Keywords Sustainable water supply system · System reliability · Customer satisfaction ·

Bowtie analysis · Fault tree analysis · Event tree analysis

* Husnain Haider
[email protected]; [email protected]
1
Department of Civil Engineering, College of Engineering, Qassim University, Qassim 51452,
Buraydah, Saudi Arabia
2
Professor and Provost, School of Engineering, University of British Columbia, Okanagan Campus,
3333 University Way, Kelowna, BC V1V 1V7, Canada

13
Vol.:(0123456789)
3482 H. Haider et al.

1 Introduction

Unlike with other products, residential customers blame their water supplier for all sorts of
service failures (OFWAT 2020). Although suppliers strive to deliver reliable, responsive,
and affordable services to attain customer satisfaction (USEPA 2017a), water supply sys-
tems (WSSs) undergo structural, hydraulic, and water quality failures during routine opera-
tions (Mesalie et al. 2021; Beygi et al. 2019; Yannopoulos and Spiliotis 2013). Customer
acceptance (tolerance) at a specified water rate is correlated with customer satisfaction
(CS), while conflicting customer preferences (needs) in utility services lead to customer
dissatisfaction (Sayles et al. 2021; Amano et al. 2016). Considering that customers accept
bottled water for drinking and intermittent supply for other uses at low water rates, fur-
ther quality decline will increase dissatisfaction and generate complaints. A reliable system
operating with fewer failures and complaints achieves high CSL, and reported complaints
should be promptly responded to and resolved, as delay increases dissatisfaction (USEPA
2017b). Customer satisfaction assessment is an important and regular activity for improv-
ing system reliability (SR) and ensuring customer confidence (Vloerbergh et al. 2007).
Complaint resolution not only improves CS but also eliminates the root cause of a
defect. This process establishes a stimulating synergistic relationship between SR and CS
(Haider et al. 2016a). Complaints can have multiple causes, i) system failure, e.g., service
connection (SC) issues (Maziotis et al. 2020); ii) unplanned interruptions, such as main
breaks (Tian et al. 2022); and iii) administrative or documentation errors, such as overbill-
ing (Haider et al. 2016b). The utility provider needs to promptly answer the email or phone
call, schedule a customer visit, identify the cause of the failure, and quickly resolve the
complaint. The utility competence throughout the complaint resolution process is impor-
tant. Since a complex issue may take longer to identify and resolve, an early visit followed
by sharing the findings with the customer reduces dissatisfaction (Haider et al. 2016a). The
literature reports two types of customer satisfaction assessment methods for water suppli-
ers: i) questionnaire surveys (Stitzel and Rogers 2022; Ohwo and Agusomu 2018) and ii)
indices that evaluate the complaint response mechanism (OFWAT 2012). Both methods
mainly evaluate the effectiveness of the utility’s response. Instead of criticizing the avail-
able methods, some limitations are outlined hereafter regarding effectively synthesizing the
information from the complaint record.
In the first method type, customers are interviewed about their experiences and opinions
on the utility’s practices and policies, e.g., satisfaction questionnaires, strategic improve-
ment methods, subjective social indicators, and SERVQUAL (Vloerbergh et al. 2007).
The questionnaire assesses the CS level (CSL) on key aspects, such as the utility’s per-
sonnel competency and response efficiency (Han et al. 2015; Hanson and Murrill 2013;
Ohwo and Agusomu 2018). SERVQUAL uses provider service quality dimensions, such
as responsiveness to customers, reliability in delivering promised services, source to tap
safety, empathy for customers, tangibility of equipment and personnel, and access to sup-
plier (Vloerbergh et al. 2007). The method involves collecting qualitative data on explana-
tory variables and comparing them with the desired service level using descriptive statis-
tics (Kassa et al. 2017). The strategic improvement method (SIM) evaluates the relative
importance of certain factors, e.g., the call-center response and the billing mechanism
(Vloerbergh et al. 2007), and assesses CS based on the utility performance of these factors.
Several studies have collected qualitative data through questionnaire surveys and employed
statistical methods, such as the ordered logit model (Anh et al. 2022), exploratory factor
analysis (Li et al. 2020) and conjoint analysis (Amano et al. 2016), to ascertain important

13
Customer‑Driven Water Supply Systems: Synergizing System… 3483

factors (or indicators) and use them to assess customer preferences or satisfaction. Despite
low response rates (Sivo et al. 2006) and low survey frequencies (Franceschini et al. 2010),
surveys without incentives or legal binding produce a small number of responses (Smith
et al. 2019). Given the possibility of overlooking customers who experienced service fail-
ures in reality, the findings might not reflect actual CS and may be misleading. The long
interval between the complaint reporting and the interview can also generate ambiguous
findings, as the interviewees may have forgotten their actual satisfaction level.
In the second method type, the indices primarily evaluate the number of telephoned and
written complaints and the response time (WUG 2014; OFWAT 2012). Privately owned
utilities in the UK and Europe use water rates and a service incentive mechanism index
(SIMI), which aggregates quantitative information on complaint responses and qualita-
tive information from surveys to generate a score between 1 and 100. In 2018–2019, the
SIMI ranged from 75 to 90 for water suppliers in the UK and Wales, with an average
score of 84.4 (OFWAT 2022). Although the index includes abandoned calls and resolved
complaints, it does not consider the complaint type (minor or major), allied severity (low
pressure vs. no water), or personnel findings at the customer’s location, which essentially
defines the complaint’s complexity and the subsequent resolution time.
An analysis of complaint records reveals the true state of CS, identifies the root causes
of complaints, and estimates failure probabilities. Subsequently, prioritized proactive
maintenance can reduce the complaint frequency and help field personnel by lessening
the severity of the issue (Dey et al. 2015). This probability and consequence relationship
turn the complaint origination response system into a risk analysis problem. Haider et al.
(2016a) were the first to define customer dissatisfaction risk (CDR) as the ‘risk of not meet-
ing the desired service level’. They estimated CDR using failure mode and effect analysis
(FMEA), which combined the failure probability and the consequence in one step, while
the analysis contains two discrete components: i) SR, which reduces the probability of
complaints, and ii) CS, which is based on the utility response. To the best of our knowl-
edge, no customer-driven risk analysis approach has separately evaluated SR and CS to
facilitate the decision-making of water suppliers and achieve long-term sustainability.
The present study develops a novel customer-driven approach for managing WSSs.
Bowtie analysis (BTA) synergistically structures fault tree analysis (FTA) to evaluate SR
and event tree analysis (ETA) to assess CS based on the utility response. The proposed
approach is not a substitute for existing interview-based methods; rather, it facilitates util-
ity management to effectively implicate complaint data in the decision-making process. An
application of the framework to a smaller water utility in Okanagan Basin (BC, Canada)
evaluates its pragmatism.

2 Methodology

2.1 Customer‑driven Framework

Figure 1 illustrates the customer-driven framework for improving SR and CS in WSSs.

The complaints record reveals the root causes of water quality (WQ), pressure, and struc-
tural failures. The analysis identifies the failed components in a system, i.e., source, treat-
ment, distribution network, or the in-house plumbing system. The present research devel-
ops three fault trees and event trees for each complaint type, i.e., structural, pressure, and

13
3484 H. Haider et al.

Step 1: Baseline data collection and analysis of customer

complaints

Step 2: Root cause analysis and identification of failure modes

with high RPN using FMEA (see Haider et al. 2016)

Step 3: Detailed risk analysis using Bowtie analysis (BTA)

Step 3.1: Development of Step 4: Development of event

fault tree for improving
tree for improving customer
system reliability and
satisfaction and estimation of
estimation of exceedance
probabilities (FTA) frequencies (ETA)

System reliability Customer dissatisfaction risk

Yes Yes
System is Is the system Is the risk Customers are
reliable! No
reliable? acceptable? satisfied!

No No

Step 4: Prioritization of high-risk system failures and functional

events after the complaint registration

Step 5: Application of improvement actions

(Risk Mitigation)

Stop

Fig. 1  Customer-driven approach for improving system reliability and customer satisfaction in water supply
systems

WQ. Finally, the prioritization of high-risk failure modes and events can help with plan-
ning improvement actions.

2.2 Case Study Area and Complaint Analysis

Smaller utilities in British Columbia encounter various challenges such as low treatment
levels, economies of scale, and financial resources (Haider et al. 2014, 2016b) and experi-
ence the highest count of boil-water advisories across Canada (WATERTODAY. CA 2020).
The study area here is a medium-sized utility consisting of 16,000 consumers served by
6,400 connections in the Okanagan Basin with mixed (residential, commercial, public, and
agricultural) land use and rolling terrain. Without a conventional filtration system, the util-
ity supplies chlorinated Okanagan Lake water. Most of the 125 km long distribution mains
are less than 25 years old.

13
Customer‑Driven Water Supply Systems: Synergizing System… 3485

A complaint comprises useful information, including i) failure location; ii) reporting

time; iii) response time (until the visit of utility personnel); iv) root cause of failure;
vi) nature of the problem, i.e., utility side (US) or homeowner side (HS); vi) extent of
the problem, i.e., minor issue vs. major issue; vii) time until the complaint was fully
resolved; and viii) the CSL, e.g., ultimately happy, somewhat dissatisfied, and highly
dissatisfied (e.g., shouting at utility staff) (Haider et al. 2016a). To implement the frame-
work presented in Fig. 1, this study uses a three-year record of 1,500 work orders (com-
plaints), estimated failure probabilities (­ Pi) and consequence events. Figure 2 shows that
structural failures originated most (81%) of the complaints, while inadequate source
WQ caused over 80% of WQ complaints.

Fig. 2  Analysis of customer complaints in the study area: (a) overall percentage distribution of 1500 com-
plaints, (b) percentage distribution of structural complaints, and (c) percentage distribution of structural
(ST), water quality (WQ), and pressure (PR) complaints along with the system components

13
3486 H. Haider et al.

2.3 Bowtie Analysis

BTA is a risk assessment method that integrates the causes and consequences of failure. On
the left side of Fig. 3, the fault tree analysis (FTA) evaluates all possible failures to assess
SR, while on the right side, the event tree analysis (ETA) evaluates the utility’s response
and culminates at a given level of CDR as a consequence of BTA. The overall outcomes
inform about the present state of SR and CS, prioritize high-risk events, and estimate CDR
for different scenarios.

2.3.1 Fault Tree Analysis

A fault tree is a tree-shape diagram that establishes the logical interrelationships between
critical events (complaint type) and their root causes (Rausand and Hoyland 2003). The
complaints analysis in Fig. 2 provides the basis for FTA development. The FTA for pressure
failures in Fig. 4a shows the complaints as the intermediate events and their root causes
(basic events) as the inputs. The utility received two types of pressure-related complaints.
The ’no water’ complaints originated from unplanned interruptions without informing cus-
tomers prior to the event, while ‘low-pressure’ complaints followed from multiple issues
occurring on both the utility and homeowner sides. Complaints about having no water
can lead to high CDR if not resolved immediately. Table 1 lists all root causes of pressure
complaints. The root causes on the US resulting low pressure complaints include delays in
switching-on booster station ­(X1,4), pressure relief valve failure ­(X1,5), SC failure ­(X1,6),
and meter issues (­ X1,7 to ­X1,10) (Fig. 4a). Plumbing failure ­(X1,11), clogged filters/sprinkler
system ­(X1,12), and frozen pipes ­(X1,13), identified as HS issues by the utility field staff, led
to complaints of low pressure.
The FTA for water quality complaints in Fig. 4b was initiated from socioeconomic con-
cerns in addition to traditional aesthetic WQ problems. Some customers complained about

Fault tree analysis (FTA) Event tree analysis (ETA)

Initiating Utility-side Resolved Outcomes Frequency

Responded
event issue same day (CSL) ()
x1

Yes High 4.71 x 10-3

Root causes

Yes 0.73
Yes 0.16 No - -
X1
x…
2

0.87 0.27
No 15 - -
Intermediate events
xm

0.84
Top event

Yes - -
X2

0.463 Yes 0.0000

0.25 No - -
…

No 0
0.13
No - -
Xn

0.75

System reliability assessment Customer satisfaction assessment

Fig. 3  Vignette of bowtie to assess system reliability using FTA (left side) and customer satisfaction using
ETA (right side)

13
Customer‑Driven Water Supply Systems: Synergizing System… 3487

Pressure
complaint

(No water)
Emergency Low pressure

Unplanned X1,3 Homeowner side

Utility side issue
interruption issue

X1,1 X1,2 X1,4 X1,5 X1,6 Meter issue X1,11 X1,12 X1,13

X1,7 X1,8 X1,9 X1,10

(a)

Water quality
complaint

Aesthetic WQ
Economic loss Social concern
complaint

X2,1 X2,2 Taste in water Color in water Odour in water X2,12 X2,13

X2,3 X2,4 Homeowner side X2,9 Homeowner side

Utility side issue
issue issue

(b) X2,5 X2,6 X2,7 X2,8 X2,10 X2,11

Fig. 4  Fault tree analysis (FTA) for, (a) pressure complaints, (b) water quality complaints, and (c) structural
complaints

stained clothes and bathroom fixtures (­X2,1) and the rapid clogging of the in-house filtra-
tion system ­(X2,2), while others voiced concerns about the health of elderly, juvenile, and
ill persons ­(X2,12) in their families. Table 2 shows that most of the aesthetic WQ complaints
originated due to the inadequate quality of supplied ­(X2,3- to ­X2,5) or source ­(X2,9) water.
Figure 4c illustrates the FTA for three main types of frequent structural complaints,
including SC issues, minor water leaks, and major water leaks (see Figs. 2c and 4c).
While responding to ON/OFF requests ­(X3,1), the utility staff found minor ­(X3,2) or
major ­(X3,3) issues culminating in the need for unplanned proactive maintenance. Both

13
3488 H. Haider et al.

Structural
complaint

Service connection
related issue Major water leak
A

Service X3,22 X3,23

Meter issue Utility side issue
connection issue

Inspection/ Curb box

ON/OFF issue X3,9 X3,10 X3,11 X3,12 X3,19 X3,20 X3,21
maintenance issue adjustment

X3,1 X3,2 X3,3 X3,4 X3,5 X3,6 X3,7 X3,8

Minor water leak

Utility side issue X3,17 X3,18

(c)
X3,13 X3,14 X3,15 X3,16

Fig. 4  (continued)

the minor ­(X3,5) and major (­X3,6) repairs were also identified in response to “SC inspec-
tion/maintenance” requests. Such complaints essentially add to proactive maintenance,
as the issues were resolved before the failure occurred. Other structural complaints were
initiated due to curb box adjustment ­(X3,7 and ­X3,8) and various types of meter issues
­(X3,9 to X
­ 3,12) (see Table 1). Minor water leaks on the US shown in Fig. 2c were caused
by stainless steel rod ­(X3,13), gate valve ­(X3,14), galvanized nipple ­(X3,15), and service
line ­(X3,16) failures. Major water leaks occurred when the service line was damaged
­(X3,19), the pressure gauge failed ­(X3,20), and the hydrant broke (­X3,21). Sometimes, ice
melt and surface runoff drainage ­(X3,23) resulted in customer complaints.
The FTA adopted a probabilistic approach to deal with independent basic events
(root causes), where any basic event can cause the occurrence of the intermediate or top
event. The following equation was used for OR-gate:
n
∏
Px = 1 − (1 − Pf ), i = 1, 2, 3, … , n (1)
i=1

where Px is the probability of failure of an intermediate or top event using the OR-gate, Pf
represents the probability of the ith basic event (or intermediate event), and n is the number
of basic events that lead to the occurrence of the top event. Individual Pf is essentially the

13
 Table 1  Basic events (root causes) behind pressure, water quality, and structural complaints
X Symbol Basic event (root cause) Probability of
basic event (­ Pf)

Pressure complaints ­(X1) X1,1 Unplanned flushing of mains 2.74 × ­10–3

X1,2 Unplanned flushing of water hydrants 3.65 × ­10–3
X1,3 Unplanned interruption due to water main break 1.80 × ­10–3
X1,4 Booster station (BS) turned OFF 1.82 × ­10–3
X1,5 Pressure releasing valve (PRV) clogged 8.21 × ­10–3
X1,6 Service connection failure 1.82 × ­10–3
X1,7 Dirt inside meters 6.39 × ­10–3
X1,8 Mechanical parts failure 1.96 × ­10–2
X1,9 Dole valve failure 1.46 × ­10–2
X1,10 Strainer clogged 1.46 × ­10–2
X1,11 In-house plumbing failure 1.64 × ­10–2
X1,12 Filter/sprinkler system failure 1.82 × ­10–3
Customer‑Driven Water Supply Systems: Synergizing System…

X1,13 Frozen pipes 2.74 × ­10–3

Water quality complaints ­(X2) X2,1 Stained clothes, electric appliances, or washroom fixtures 4.56 × ­10–3
X2,2 Quick clogging of in-house filters and membranes needs early maintenance 9.12 × ­10–3
X2,3 Taste in water due to source water quality 3.65 × ­10–3
X2,4 Taste in water due to high chlorine in supplied water 3.65 × ­10–3
X2,5 Colored water due to presence of algae in source water 2.19 × ­10–2
X2,6 Colored water due to flushing of water main 1.28 × ­10–2
X2,7 Colored water due to clogging of in-house plumbing pipes 1.82 × ­10–3
X2,8 Colored water due to clogging of in-house filtration systems 2.74 × ­10–3
X2,9 Odor in water due to high chlorine in supplied water 1.19 × ­10–2
X2,10 Odor in water due to clogging of in-house filtration system 9.12 × ­10–4
X2,11 Odor in water due to uncleaned in-house water storage reservoir 9.12 × ­10–4
X2,12 Health issues among elderly, juvenile, or ill persons 6.39 × ­10–3
3489

13
 Table 1  (continued)
3490

X Symbol Basic event (root cause) Probability of

basic event (­ Pf)

13
X2,13 Customer concerned about WQ and asked for WQ testing results 8.21 × ­10–3
Structural complaints (­ X3) X3,1 Service connection turned ON/OFF without repair 3.14 × ­10–1
X3,2 Service connection turned ON/OFF with minor repair 3.65 × ­10–2
X3,3 Service connection turned ON/OFF with major repair 1.37 × ­10–2
X3,4 Inspection of service connection required and completed without any issues 8.85 × ­10–2
X3,5 Inspection of service connection with minor repair 4.29 × ­10–2
X3,6 Inspection of service connection with major repair 6.40 × ­10–3
X3,7 Curb box adjustment without excavation 4.93 × ­10–2
X3,8 Curb box adjustment with excavation 3.28 × ­10–2
X3,9 Meter issue due to interior dirt 3.65 × ­10–3
X3,10 Meter issue due to parts failure 9.12 × ­10–3
X3,11 Meter needs to be replaced 1.09 × ­10–2
X3,12 Dole valve issue 1.19 × ­10–2
X3,13 Stainless steel rod failure 1.28 × ­10–2
X3,14 Gate valve issue 6.40 × ­10–3
X3,15 Galvanized nipple issue 2.70 × ­10–3
X3,16 Service line blocked or needs to be replaced 2.28 × ­10–2
X3,17 Issue on homeowner side 5.66 × ­10–2
X3,18 Other reasons, e.g., ice melt and surface drainage 9.10 × ­10–3
H. Haider et al.
 Table 1  (continued)

X Symbol Basic event (root cause) Probability of

basic event (­ Pf)

X3,19 Service line damaged 1.55 × ­10–2

X3,20 Pressure gauge failed 1.82 × ­10–3
X3,21 Water hydrant failure 4.56 × ­10–3
X3,22 Issue on homeowner side 2.19 × ­10–2
X3,23 Other reasons, e.g., ice melt and surface drainage 1.73 × ­10–2
Customer‑Driven Water Supply Systems: Synergizing System…
3491

13
3492 H. Haider et al.

ratio between number or failure days (i.e., complaints) and the total number of days in the
assessment period, i.e., 1095 days for three years of complaint data.

2.3.2 Event Tree Analysis

In Fig. 5a, ETA evaluates the utility response to pressure complaints using various func-
tional events (columns). The upper part of Fig. 5 presents the ETA for ‘no water’ com-
plaints waiting for an emergency response. For a HS issue, the customer needs to be imme-
diately informed and put in touch with a private contractor to resolve the problem. For
US issues, immediate response and resolution on the same day leads to a high customer
satisfaction level (CSL). Resolving a major issue involving deep excavation, sophisticated
equipment, or the replacement of expansive parts results in medium CSL if the initial
response comes on the same day and is resolved within two weeks. In the lower part of
Fig. 5a showing ETA for low pressure complaints, the utility provider needs to respond
within 4 h; failure to do so results in low CSL. Complaints about a minor issue that are
responded to within 4 h are best resolved on the same day to attain high CSL. Low pres-
sure complaints caused by a major issue have to be resolved within two weeks to attain
medium CSL. The ETA shown in Fig. 5b shows that all the complaints were nonemergen-
cies because of the overall customer acceptance of lower source WQ.
The upper part of the ETA in Figure 5c evaluates the response to SC on/off and inspec-
tion requests. A complaint responded to and resolved at the scheduled time without iden-
tifying any structural issues results in high CSL. If utility personnel immediately inform
the customer about a minor repair required on the homeowner side, the complaint also
ends with high CSL. A minor issue found on the US needs to be resolved on the same day;
otherwise, the CSL declines accordingly until the issue is completely resolved. The mid-
dle part of Figure 5c evaluates emergency structural complaints, which the utility provider
needs to immediately respond to. For a major issue or special equipment need, a next-day
resolution leads to medium CSL. For nonemergency complaints, where field staff identify
issues on both the utility and homeowner sides (e.g., minor leak on service line), the home-
owner needs to be informed within 4 hours to obtain high CSL.
The following equation assesses the frequency (λi) of a given functional event:
n
∏
𝜆i = 𝜆 × Pi , i = 1, 2, 3, … , n (2)
i=1

The results given in the last two columns of Fig. 5 are discussed in Sect. 3.2.

2.4 Reliability Analysis, Risk Prioritization, and Risk Mitigation

The SR is complementary to the failure probability (Mencik 2016). The following equation
assessed the reliability of various system components:
Rx = 1 − Px (3)
where Rx is the reliability and Px is the probability of failure of a subcomponent or
component.
The overall system reliability (Rs) was estimated using the following equation that is the
final outcome of FTA:

13
Customer‑Driven Water Supply Systems: Synergizing System… 3493

n
∏
Rs = Rx (4)
i=1

where Rx is the overall reliability of pressure, water quality, and structural components.
The following equation calculates the risk of customer dissatisfaction (RCD), which is
the final outcome of ETA:

Resolved Resolved
Emergency Utility Major Resolved Resolved Customer 
Responded Responded within 8 within 24
response side issue/missing within 1 within 2 satisfaction (per
immediately within 4 hrs hrs (same hrs (next
required issue equipment week weeks level (CSL) day)
day) day)
Yes High -
Yes 0
0 No very low -
Yes 0
0.57 Yes High 0.0036
Yes No 1.0 Yes Medium -
1.0 1.0 No 0 Yes Low -
Yes 0 No 0
0.07 0 No Very low -
0
No Very low 0.0027
0.43
Pressure
High -
complaint Yes
0.0958 No 0
0 No Very low -
0
Yes Medium 0.0018
Yes 1.0
0.04 No very low -
Yes 0
0.73 Yes High 0.0484
Yes No 1.0 Yes Medium -
0.77 0.96 No 0 Yes Low -
0 No 0
0 No Very low -
No 0
0.93 Yes Medium 0.0027
Yes 0.75
0.20 No very low 0.0009
No 0.25
0.27 Yes Medium 0.0064
No 0.44 Yes Low 0.0082
0.80 No 1.00
0.56 No Very low -
0.00
Yes Very high 0.0146
No 0.70
(a) 0.23 No Low 0.0064
0.30

Emergency Utility Source Responded

Responded Unplanned Responded 
response side water within the CSL
immediately maintenance next day (per year)
required issue issue same day

Yes Low -
0.00 Yes Low -
Yes 0.00 Yes
Yes 0.00 No 0 Very low -
0.00 0 No Very low -
No 0
Yes 0.00 Yes Medium -
0.00 No 0.00 Yes Low -
0.00 No 0
0 No Very low -
0
Yes High -
No 0.00
WQ complaint 0.00 No Low -
0.0885 0 0.00

Yes Very low 0.0657

0.78 Yes low 0.0128
Yes 1.00 Yes
Yes 0.70 No 0 Very low -
0.95 0 No Very low -
No 0
0.22 Yes Medium 0.00547
No 1.00 Yes Low -
No 0.30 No 0
1.00 0 No Very low -
0
(b) Yes Very high 0.0046
No 1.00
0.05 No Low -
0.00

Fig. 5  Event tree analysis (ETA) for (a) pressure, (b) water quality, and (c) structural complaints

13
3494 H. Haider et al.

Resol.
Resp. Issue Major Resol.
ON/ Minor Emer. Utility Resp. in 24 Res Res. 
at on Resp. issue/ in 8 hrs
OFF repair Resp. side within hrs . in in CSL (per
asked both Imm. miss. (same
req. req. Req. issue 4 hrs (next 1W 2W day)
time sides equip. day)
day)
Yes High 0.0274
Yes 0.75
0.73 Yes Medium 0.0091
Yes No 1.00 Yes Low 0.0
0.16 0.25 No 0.0
Yes 0 No Very low 0.0
0.87 No 0.0 Very high 0.0137
0.27
Yes No Very high 0.2710
0.46 0.84
Yes Low -
Yes 0.00
Yes 0.00 Yes Low -
0.00 No No 0.0 Yes Very Low -
No 0.00 0.0 No 0.00
0.13 0.0 No Very low -
No 0.0 Low 0.0474
1.00 Yes Medium 0.0100
Yes 1
0.46 No very low 0.0100
Structural
Yes 0
complaint
0.796 1.0 Yes High -
Yes No 0.0 Yes Medium 0.0118
0.36 0.54 No 1 Yes Low -
1.0 No 0
0 No Very low -
Yes 0
0.14 No Very low -
0.00
Yes Very high 0.0191
No 0.49
0.64 No Very low 0.0200
0.51
Yes Medium 0.0182
Yes 1.00
0.09 No very low -
No Yes 20 0.00
0.54 0.69 Yes High 0.1679
Yes No 0.89 Yes Medium -
0.82 0.91 No 0.00 Yes Low 0.0182
0.11 No 0.87
1.00 No Very low 0.0027
0.13
Yes Low 0.0630
Yes 1.00
0.68 No very low -
No No 0.00
0.86 0.31 Yes Medium 0.0301
No 1.00 Yes Low -
0.32 No 0.00
0.00 No Very low -
0.00
Yes Medium 0.0009
Yes 1.00
0.10 No very low -
Yes 0
1.00 Yes High 0.0082
No 1.00 Yes Medium -
0.90 No 0 Yes Low -
0 No 0
Yes 0 No Very low -
0.14 0
Yes Medium -
Yes 0.00
0.00 No very low -
No 0.00
No 0.00 Yes Medium -
0.18 No 0.00 Yes Low -
0.00 No 0.00
0.00 No Very low -
0.00
Yes Very high 0.0547
(c) No 0.97
0.86 No Low 0.0018
0.03

Fig. 5  (continued)

RCD = 𝜆i × Ci (5)
where λi is the frequency a given functional event estimated by Eq. (3), while Ci is the con-
sequence of low CSL in Table 2.
The mitigations on fault tree reduce the Pf of root causes and thus the number of
complaints, while an improved response on the event tree reduces CDR. Table 2 defines
the priority levels of different types of failures. In the event tree, both the promptness
and efficiency are functions of time. Table 2 presents the exact and linguistic terms for

13
Customer‑Driven Water Supply Systems: Synergizing System… 3495

Table 2  Failure frequency, customer satisfaction level, and consequence

Failure probability (­ Pf) Frequency or likelihood (L) Consequence “C” Customer
(Customer dissatisfaction
dissatisfaction) risk (CDR)
Pf Priority levela λi Linguistic Score Linguistic Score Score RPLb

< 0.0015 L < 0.005 Very low 2 Very High 10 < 10 L

0.0015–0.003 M 0.005–0.01 Low 4 High 8 10–19 M
0.003–0.005 H 0.01–0.03 Medium 6 Medium 6 20–39 H
0.005–0.01 VH 0.03–0.05 High 8 Low 4 40–59 VH
> 0.01 EH > 0.05 Very High 10 Very ­low1 2 > 60 EH
a
Low priority level (excluding ON/OFF requests) is acceptable, and no major improvement action is
needed; low (L), moderate (M), high (H), very high (VH), extremely high (EH).
b
Risk priority level (RPL)

the frequency and consequence. As both λi and consequence range between 1 and 10,
Eq. (5) estimates risk scores between 1 and 100. Table 2 linguistically defines the risk
priority level (RPL), where the outcomes with high CSL correspond to acceptable risk
while others are prioritized. Table 3 presents five risk mitigation actions, their impacts
on root causes and response frequencies to improve SR and CDR, and the prioritiza-
tion of mitigation actions based on cost and priority levels. Detailed calculations are
given in Appendix A. The cost estimates are based on the performance assessment
data, the information of past (source change) projects and future (filtration plant) plans
obtained from the utility personnel of the study area, and published literature (Ferreira
and Carrico 2019).

3 Results

3.1 System Reliability Assessment using FTA

Using failure probabilities, Eq. (1) analyzed the root causes, and Eq. (3) calculated the
SR for pressure, WQ, and structural components for the fault trees shown in Fig. 4.
Figure 6a describes the percentage contributions of different SR failures, and Fig. 6b
presents Px for each component. More than 90% of failures resulted in low pressure
­(Px = 9.31 × ­10–2), while the remaining ­(Px = 0.82 × ­10–2) left customers with no water.
Due to the meter-related issues having the highest frequency (­Px = 5.5 × ­10–2), the fre-
quency of low-pressure failures on the US (­Px = 7.2 × ­10–2) was much higher than that
on the HS ­(Px = 2.1 × ­10–2). Unplanned maintenance activities (­X1,1 and ­X1,2) left the
customers with no water twice a year (­ Px = 0.64 × ­10–3); hence, they need to be given a
very high priority (VHP). Finally, the reliability of the pressure component ­(Rx1) was
found to be approximately 0.9 (90%) (Fig. 6c). Figure 6b illustrates that economic loss
­(Px = 1.37 × ­10–2) and health issues ­(Px = 1.46 × ­10–2) caused 30% of WQ complaints,
while the remaining 70% of complaints originated from aesthetic issues ­(Px = 6.2 × ­10–2).
Color problems (­Px = 3.92 × ­10–2) caused the largest number of complaints, followed
by odor ­(Px = 1.55 × ­10–2) and taste ­(Px = 0.73 × ­10–2) issues. The reliability of the WQ
component ­(Rx2) was also found to be approximately 90% (Fig. 6c).

13
 3496

13
Table 3  Risk mitigation actions to improve system reliability and customer satisfaction
No Mitigation action Description Impact on system reliability Impact on customer dissatisfaction Tentative
risk additional
cost
(CAD
million)a

A1 Automation of booster station and Booster stations are installed to Reduces Pf for ­X1,4 and ­X1,5 Reduce λ for emergency response 0.12
proactive maintenance of PRV maintain pressure in the hilly
terrain of the study area
A2 Planned proactive maintenance and Planned and proactive maintenance Reduces Pf for ­X1,1, ­X1,2, ­X1,3, and Reduce λ for both emergency and 0.15
asset management and asset management reduce the ­X3,1 nonemergency response
number of complaints
A3 Proactive inspection of service Newly hired staff can perform Reduces Pf for ­X3,4, ­X3,5, ­X3,6, ­X3,7, Reduce λ for both emergency and 2.5
connections with hiring of proactive inspection of service ­X3,8, ­X3,9, ­X3,10, ­X3,11, ­X3,12, ­X3,13, nonemergency response
additional staff connections and improve ­X3,14, ­X3,15, ­X3,16, ­X3,19, ­X3,20
complaint response mechanism
A4 Improvement of water source Augmenting creek water with lake Reduces Pf for ­X2,1, ­X2,2, ­X2,3, ­X2,4, Reduce λ for nonemergency 15
water in 2013 improves the quality ­X2,5, ­X2,6, ­X2,9, ­X2,12, ­X2,13 response
of source water
A5 Installation of filtration system for Installation of complete filtration Reduces Pf for ­X2,1, ­X2,2, ­X2,3, ­X2,4, Reduce λ for nonemergency 35
source water treatment system can significantly minimize ­X2,5, ­X2,6, ­X2,9, ­X2,12, ­X2,13 response
the water quality complaints
Cumulative risk mitigation cos 52.77
a
Tentative cost includes operation and maintenance cost for 10 years. Past cost data have been updated based on a cumulative inflation rate of 16.8% for the period between
2014 and 2021
H. Haider et al.
Customer‑Driven Water Supply Systems: Synergizing System… 3497

No water low pressure Economic loss Health concerns

Aesthetic issues Major leak Minor leak SC and meter

Structural failures

Water quality failures

System pressure failures

0% 20% 40% 60% 80% 100%

(a) Percentage contribution

Major leak
Structural
failure

Minor leak

Service connection

Socil concern
Pressure Water quality
failure

Aesthetic failure

Economic loss

Low pressure
failure

No water

0.0E+00 5.0E-02 1.0E-01 1.5E-01

(b) Failure probability (Px)

System pressure
reliability

Water quality reliability

Structural reliability

Overall reliability of WSS

0.0 0.2 0.4 0.6 0.8 1.0

(c) Reliability (%)

Fig. 6  System reliability analysis results for structural, water quality and pressure components: (a) percent-
age contribution of different failures, (b) failure probabilities, and (c) overall reliability

In the FTA for the structural component, the utility received 185 work orders
annually for SC seasonal on/off or inspection requests. Some on/off requests
­(Px = 5.02 × ­10–2) turned into standard complaints when previously hidden issues were
identified. As the utility repaired SCs, meters, dole valves, minor and major leaks, and
curb box adjustments before failure, the requests essentially enhanced SR. Excluding
these work orders, Fig. 6b illustrates that 38% of complaints registered minor water
leaks ­(Px = 11.0 × ­10–2), 21% major (­Px = 6.11 × ­10–2), and the remaining reported

13
3498 H. Haider et al.

numerous service connection-related issues. Damaged service lines mainly caused

major leaks (­Px = 2.19 × ­10–2) on both the utility and homeowner sides. Excluding SC-
related requests, the reliability of the structural component (­Rx3) was also found to be
approximately 71% (Fig. 6c) and required major intervention.

3.2 Customer Satisfaction Assessment Using ETA

Detailed CDR assessment results for ETAs of all components described in Fig. 5 are attached
in Appendix A. For complaints of no water (λi = 3.6 × ­10–3/day), the utility responded imme-
diately and resolved the issue on the same day, and field personnel left the customer happy
(P1 in Appendix A). In some cases (λi = 2.7 × ­10–3/day), a late response to an emergency
complaint (P2) resulted in a dissatisfied customer. For a large number (λi = 6.7 × ­10–2/day)
of nonemergency complaints on the US (P4-P8), either high likelihood (P4) or consequence
due to late resolution (P6-P8) led to high CDR.
The source water intake drew water from a creek at a depth where turbidity mainly
remained within the acceptable limits (1–2 NTU). Occasional water fluctuations raised
turbidity and chlorine levels, resulting in both color and odor complaints. A very high
likelihood (λi = 6.6 × ­10–2/day) and customer dissatisfaction levels were noted for the
WQ complaints originating from inadequate source WQ (W1 and W3). Although they
responded the same day and medium frequency (λi = 1.28 × ­10–2/day), increased turbidity
due to unplanned maintenance (W2) seriously (high CDL) upset customers. The occasional
(λi = 4.6 × ­10–3/day) failure of storage and treatment systems on the consumer side (W4)
resulted in low CDR when conveyed to the homeowner in a timely manner.
On/off requests (S1-S5, Appendix A) improved SR when a minor or major repair was
identified and resolved before a failure occurred; nevertheless, the utility provider needs
to respond to these complaints at a convenient time for the customer. S1 (λi = 2.74 × ­10–2/
day) to S3 (λi = 1.37 × ­10–2/day) achieved high RPL due to its medium frequency, while S5
with a late response resulted in high CDR due to high (λi = 4.74 × ­10–2/day) frequency and
consequences. A water main break requiring an emergency response (S6 and S7) gener-
ally takes some time to resolve. Proactive measures, e.g., inspection of service connec-
tions and risk-based repair and replacement of mains, can avoid such complaints. A late
response can result in high CDR, even if the problem was on the HS (S9). For nonemer-
gency complaints, a response within 4 h followed by early identification and resolution
improves the CS (S10-S13). A late response to such complaints leads to very high CDR
(S14 and S15).

3.3 Risk Management

Appendix B describes the failure events, failure probabilities, risk priority levels, and pro-
posed mitigation actions. Overall, 32 failure modes define CDR in three event trees for
pressure, WQ, and structural components (Fig. 5). Figure 7a shows the decline in the num-
ber of high-risk failure modes with improvement actions. Action 5 achieved low risk prior-
ity (acceptable risk) for 60% (18) of failure modes as well as the absence of failure modes
with extreme priority (RS ≥ 60). Figure 7b correlates the cumulative cost of improvement
actions with SR and CDR.

13
Customer‑Driven Water Supply Systems: Synergizing System… 3499

L M H VH EH

A5

A4

A3

A2

A1

A0

0 5 10 15 20 25 30 35
No of failure modes in the event tree analysis
(a)

100 A0
A1
System reliability / Risk reduction (%)

90
A2
A4
80 A3 100
System
70 reliability 80
A5

60 60

Customer 40
50
dissatisfaction
risk 20
40 0 20 40 60

30

20
0 5 10 15 20
(b) Cummulative cost (million USD)

Fig. 7  Outline of bowtie analysis results, (a) risk priority levels of 32 failure modes in three event trees for
pressure, water quality, and structural complaints, (b) cumulative cost of improvement actions vs. system
reliability and customer dissatisfaction risk. A1: Automation of booster station and proactive maintenance
of PRV; A2: Planned proactive maintenance and asset management; A3: Proactive inspection of service
connections with hiring of additional staff; A4: Improvement of water source; and A5: Installation of filtra-
tion system for source water treatment

4 Discussion

To implement a customer-driven approach, utilities need a well-maintained complaint

record for the previous few years. FTA requires that the data be categorized into differ-
ent root causes (Table 1) under the three main components. Readers can see the fishbone
diagram in Haider et al. (2016a) identifying root causes, which further helps in understand-
ing the fault trees. A careful effort to classify the complaints in different functional events
(columns of Fig. 5) for ETA could be challenging for utility providers. Comprehending and

13
3500 H. Haider et al.

interpreting the complaints’ information using natural language processing and deep learn-
ing techniques can help in this regard (Tian et al. 2022). Certainly, minimizing the prob-
ability of root causes is a more desirable option. Most (88% pressure, 98% WQ, and 87%
structural) of the complaints resulting from US issues imply the utility’s responsibility for
ensuring SR and CS. Data showing 70% structural complaints (or work orders) from ser-
vice connection issues are consistent with past findings (Ferreira and Carrico 2019; Farley
and Trow 2003). Over 70% of complaints highlight inadequate source WQ as the primary
customer concern.
Once a complaint registers after a failure event, utility response efficiency brings sev-
eral possible CSLs. Multiple safety barriers reduce or eliminate consequences in indus-
tries (Shahriar et al. 2012), while in WSS, a prompt response, quick problem identification,
effective communication, and the efficient resolution of complaints essentially reduce CDR
(see functional events at the top of Fig. 5). Essentially, the availability of equipment and
supplementary parts minimizes the resolution time.
Booster station automation is an inexpensive action that can improve 1% SR and mini-
mize 4% of cumulative CDR. Proactive asset (mains and hydrants) management can reduce
the complaints resulting in high CDL (A2, Table 3). The proactive inspection of meters and
service connections (A3) with available resources drastically improves both SR and CS.
A recent study of Lisbon’s metropolitan area (Portugal) reported an annual investment of
€100,000 (135,000 $CAD) for 1.5% of the 31 km of long mains, 5% of 2180 service con-
nections repaired, and new meter installation (Ferreira and Carrico 2019). The estimated
cost of $265,000 CAD/year (Table 3) for rehabilitating 2% of the 152 km of mains (A2)
and 10% of service connections (A3) cumulatively improves SR by 28% and CDR by 63%.
Shifting the water source to Okanagan Lake (A4) reduced the number of WQ com-
plaints (Haider et al. 2016a), increased SR by 5%, and reduced CDR by 6.5%. Fewer com-
plaints reporting socioeconomic concerns increase customer acceptance to low WQ. On
the homeowner side, customers can keep their filters and in-house storage reservoir clean.
Nevertheless, source water improvements need to be supported with a filtration process
(A5) to avoid high turbidity and corresponding chlorine doses, instigating odor complaints.
Although a small incremental improvement does not justify the 66% increase in cumulative
costs in Fig. 7, the results are based on fewer complaints subject to customer perceptions
about inadequate source quality and an allied acceptance of the prevailing rates. The utility
provider needs to assess customer preferences and acceptance using willingness to pay and
the willingness to accept the changes with improved WQ after the installation of a conven-
tional filtration facility (MacDonald et al. 2010).

5 Conclusion

A water utility receives complaints as a result of hydraulic, WQ, and structural failures.
Following a careful effort to arrange and analyze complaint records, the FTA helps improve
the SR by reducing failures, and the ETA enhances the CS by working on the complaint-
response mechanism. BTA integrating FTA and ETA provided a novel customer-driven
decision-support tool for WSS. Utilities mostly receive structural complaints related to SC,
and meter issues need proactive maintenance. Late responses and the nonavailability of
equipment result in high CDL even for regular service connection ON/OFF requests, while
poor source WQ was found to be the primary reason for WQ complaints. BTA simulating

13
Customer‑Driven Water Supply Systems: Synergizing System… 3501

five improvement actions revealed that cost-effective proactive maintenance and efficient
response improve the overall SR from 58 to 86%, which reduce the cumulative CDR by
62%. Spending an additional 30% of resources further improves the SR to 90% and reduce
the CDR up to 68%. The feasibility of the remaining 66% being spent on filtration plants
achieving 94% reliability and CDR reduction needs to be evaluated based on customers’
willingness to pay and acceptance. Water utilities, municipalities, and industries can adopt
the proposed approach for the long-term sustainability of their operations, processes, and
products.

Abbreviations BTA: Bowtie analysis; CDR: Customer dissatisfaction risk; CS: Customer satisfaction;
CSL: Customer satisfaction level; ETA : Event tree analysis; FTA: Fault tree analysis; HS: Homeowner
side; RPL: Risk priority level; SC: Service connection; SR: System reliability; US: Utility side; WQ: Water
quality.

Supplementary Information The online version contains supplementary material available at https://​doi.​org/​
10.​1007/​s11269-​022-​03200-6.

Acknowledgements The authors acknowledge the water utility in the Okanagan Basin (Canada) for sharing
its complaint data. The authors are also thankful to the utility provider’s field engineers for sharing their
knowledge and experience.

Author Contributions H. Haider developed the concept, methodology, analysis, and original draft. M. Alinizzi
assisted in the analysis and preparation of the original draft. M. Shafiquzzaman, S. AlSaleem, and M. Alresheedi
assisted in data preparation, validation, and review of the original draft. R. Sadiq supported the data collection,
concept, and review of the original draft.

Funding The authors declare that no funds, grants, or other support were received during the preparation of
this manuscript.

Data Availability All the shareable data are presented in the manuscript. Detailed complaint records can-
not be publicly shared due to the confidentiality agreement between the researchers and the data sharing
organizations.

Declarations
Ethical Approval Not applicable.

Consent to Participate All authors confirm their coauthorship.

Consent to Publish The authors give their permission to publish this work.

Conflicts of Interest The authors declare that they have no competing interests.

References
Amano I, Kurisu K, Hanaki K (2016) Evaluation of consumers’ preferences on drinking waters considering
the information provision effect. Water Science Technol: Water Supply 16(4):1057–1067. https://​doi.​
org/​10.​2166/​ws.​2016.​019
Anh NT, Dung NH, Thu DT (2022) Privatization in Rural Water Supply and Customer Satisfaction: An
Empirical Case Study in Vietnam. Sustainability 14(9):5537. https://​doi.​org/​10.​3390/​su140​95537
Beygi S, Tabesh M, Liu S (2019) Multi-Objective Optimization Model for Design and Operation of Water
Transmission Systems Using a Power Resilience Index for Assessing Hydraulic Reliability. Water Resour
Manage 33:3433–3447. https://​doi.​org/​10.​1007/​s11269-​019-​02311-x

13
3502 H. Haider et al.

Dey SS, Thommana J, Dock S (2015) Public Agency Performance Management for Improved Service Deliv-
ery in the Digital Age: Case Study. J Manage Eng 31(5):05014022. https://​doi.​org/​10.​1061/​(ASCE)​ME.​
1943-​5479.​00003​21
Farley M, Trow S (2003) Losses in water distribution networks. IWA Publishing, London
Franceschini F, Galetto M, Turina E (2010) Water and sewage service quality: a proposal of a new multi-
questionnaire monitoring tool. Water Resour Manage 24(12):3033–3050. https://​doi.​org/​10.​1007/​
s11269-​010-​9593-0
Ferreira B, Carriço NJ (2019) Urban water infrastructure asset management plan: Case study. Open Eng
9(1):459–467. https://​doi.​org/​10.​1515/​eng-​2019-​0058
Haider H, Sadiq R, Tesfamariam S (2014) Performance indicators for small and medium sized water supply
systems: A review. Environ Reviews 22(1):1–40. https://​doi.​org/​10.​1139/​er-​2013-​0013
Haider H, Sadiq R, Tesfamariam S (2016a) Risk-based framework for improving customer satisfaction
through system reliability in small-sized to medium-sized water utilities. J Manage Eng 32(5):04016008.
https://​doi.​org/​10.​1061/​(ASCE)​ME.​1943-​5479.​00004​35
Haider H, Sadiq R, Tesfamariam S (2016b) Intra-utility Performance Management Model (In-UPM) for the
sustainability of small to medium sized water utilities: Conceptualization to development. J Clean Prod
133:777–794. https://​doi.​org/​10.​1016/j.​jclep​ro.​2016.​06.​003
Han S, Chae M, Hwang H, Choung Y (2015) Evaluation of Customer-Driven Level of Service for Water
Infrastructure Asset Management. J Manage Eng 31(4):04014067. https://​doi.​org/​10.​1061/​(ASCE)​ME.​
1943-​5479.​00002​93
Hanson JJ, Murrill SD (2013) South Tahoe public utility district 2012 customer satisfaction and perceptions
survey report of results. Meta Research Inc, California, US
Kassa K, Chernet M, Kelemework G, Zewde B, Woldemedhin A (2017) Customer satisfaction survey: the
case of urban water supply services in Southern Ethiopia. Water Practice Technol 12(4):1009–1017.
https://​doi.​org/​10.​2166/​wpt.​2017.​105
Li H, Lv L, Zuo J, Bartsch K, Wang L, Xia Q (2020) Determinants of public satisfaction with an Urban
Water environment treatment PPP project in Xuchang. China Sustain Cities Society 60:102244. https://​
doi.​org/​10.​1016/j.​scs.​2020.​102244
MacDonald DH, Morrison MD, Barnes MB (2010) Willingness to pay and willingness to accept compensa-
tion for changes in urban water customer service standards. Water Resour Manage 24(12):3145–3158.
https://​doi.​org/​10.​1007/​s11269-​010-​9599-7
Maziotis A, Villegas A, Molinos-Senante M (2020) The cost of reducing unplanned water supply interrup-
tions: A parametric shadow price approach. Sci Total Environ 719:137487. https://​doi.​org/​10.​1016/j.​
scito​tenv.​2020.​137487
Mencik J (2016) Concise Reliability for Engineers. IntechOpen. https://​doi.​org/​10.​5772/​62009
Mesalie RA, Aklog D, Kifelew MS (2021) Failure assessment for drinking water distribution system in the
case of Bahir Dar institute of technology. Ethiopia Appl Water Sci 11:138. https://​doi.​org/​10.​1007/​
s13201-​021-​01465-7
OFWAT (The Water Services Regulation Authority) (2012) Key performance indicators guidance. www.​
ofwat.​gov.​uk. Accessed on 13 May 2022
OFWAT (The Water Services Regulation Authority) (2020) The economic regulator of the water sector
in England and Wales. https://​www.​ofwat.​gov.​uk/​house​holds/​your-​water-​compa​ny/​chang​esupp​lier/.
Accessed on 13 May 2022
OFWAT (2022) https://​www.​ofwat.​gov.​uk/​regul​ated-​compa​nies/​compa​ny-​oblig​ations/​custo​mer-​exper​ience/​
servi​ce-​incen​tive-​mecha​nism/. Accessed on 13 May 2022
Ohwo O, Agusomu TD (2018) Residential Customers Satisfaction with Public Water Provision in Ojota.
Nigeria Eur Sci J ESJ 14(23):117
Rausand M, Hoyland A (2003) System reliability theory: models, statistical methods and applications. John
Wiley & Sons
Sayles R, Smith HM, Jeffrey P (2021) Customer priorities for water and wastewater services: a compara-
tive evaluation of three elicitation methods. Water Environ J 35(1):55–66. https://​doi.​org/​10.​1111/​wej.​
12575
Shahriar A, Sadiq R, Tesfamariam S (2012) Risk analysis for oil & gas pipelines: A sustainability assess-
ment approach using fuzzy based bow-tie analysis. J Loss Prevent Process Indus 25(3):505–523.
https://​doi.​org/​10.​1016/j.​jlp.​2011.​12.​007
Sivo SA, Saunders C, Chang Q, Jiang JJ (2006) How low should you go? Low response rates and the valid-
ity of inference in IS questionnaire research. J Assoc Info Sys 7(1):17. https://​doi.​org/​10.​17705/​1jais.​
00093

13
Customer‑Driven Water Supply Systems: Synergizing System… 3503

Smith MG, Witte M, Rocha S, Basner M (2019) Effectiveness of incentives and follow-up on increasing sur-
vey response rates and participation in field studies. BMC Medical Res Methodol 19(1):1–13. https://​
doi.​org/​10.​1186/​s12874-​019-​0868-8
Stitzel B, Rogers CL (2022) Residential Water Demand Under Increasing Block Rate Structure: Conserva-
tion Conundrum? Water Resour Manage 36:203–218. https://​doi.​org/​10.​1007/​s11269-​021-​03022-y
Tian X, Vertommen I, Tsiami L, van Thienen P, Paraskevopoulos S (2022) Automated Customer Complaint
Processing for Water Utilities Based on Natural Language Processing-Case Study of a Dutch Water
Utility. Water 14(4):674. https://​doi.​org/​10.​3390/​w1404​0674
USEPA (2017a) Effective Utility Management, A Primer for Water and Wastewater Utilities, The EUM
Utility Leadership Group, United States Environmental Protection Agency, USA
USEPA (2017b) Designing Customer Complaint Surveillance For Water Quality Surveillance and Response
Systems. United States Environmental Protection Agency, USA
Vloerbergh II, Fife-Schaw C, Kelay T, Chenoweth J, Morrison G, Lundéhn C (2007) Assessing consumer
preferences for drinking water services Methods for water utilities. TECHNEAU - An Integrated Pro-
ject Funded by the European Commission. https://​sswm.​info/​sites/​defau​lt/​files/​refer​ence_​attac​hments/​
FIFES​CHAW%​202006%​20Con​sumer%​20Pre​feren​ces.​pdf. Accessed on 13 May 2022
WATERTODAY.CA (2020). https://​www.​water​today.​ca/​map-​graph​ic.​asp. Accessed on 13 May 2022
WUG (Water Utilities Group) (2014) Key performance indicators. Strategic Report. http://​annua​lrepo​rt2014.​
unite​dutil​ities.​com/​strat​egic-​report/​key-​perfo​rmance-​indic​ators. Accessed on 13 May 2022
Yannopoulos S, Spiliotis M (2013) Water Distribution System Reliability Based on Minimum Cut – Set
Approach and the Hydraulic Availability. Water Resour Manage 27:1821–1836. https://​doi.​org/​10.​
1007/​s11269-​012-​0163-5

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

13