Introduction

1. In 2013–14 the Department of Human Services (Human Services) delivered $159.2 billion payments to customers and providers.¹ The department delivers these payments and related services on behalf of the Australian Government through a variety of channels² including telephone, on-line, digital applications and face-to-face through some 400 service centres located across Australia. While there has recently been strong growth in digital channels and the department is actively encouraging the use of these self-service channels, demand for telephone services remains strong. In 2013–14, the department handled 59.5 million telephone calls about Centrelink, Child Support and Medicare services.³ The majority of calls, 43.1 million annually or more than 800 000 per week, related to Centrelink services.

2. Human Services manages over 50 Centrelink-related telephony lines, each with its own ‘1800’ or ‘13’ telephone number.⁴ Each of the main payment types such as the aged pension and employment services has its own line, and calls are managed and distributed nationally through a virtual network. Previously referred to as call centres, the department now provides telephone services through a network of 29 Smart Centres, with some $338 million expended on Centrelink telephone services in 2013–14.

3. The department’s shift to Smart Centres began in 2012–13 and is still being implemented. Previously, telephony and processing work were undertaken separately and were arranged by payment type. Smart Centres are intended to blend telephony and processing work and reorganise work around the complexity and frequency of customer transactions rather than by payment type, so as to improve customer service and allow staff to be deployed more flexibly.

Audit objective, criteria and scope

4. The audit objective was to assess the effectiveness and efficiency of the Department of Human Services’ (Human Services) management of Smart Centres’ Centrelink telephone services.

5. To form a conclusion against the audit objective, the ANAO adopted the following high-level criteria:

• Human Services offers customers effective telephone services in relation to a range of quality indicators, for example, wait times and the accuracy of the information provided;
• Centrelink call services in Smart Centres are managed efficiently; and
• Human Services effectively monitors and reports on the performance of Smart Centres’ Centrelink telephone services.

6. The audit scope did not include an examination of Smart Centres’ processing services other than the processing that is done as part of the telephone service; or Smart Centres’ Medicare and Child Support telephone services.

Overall conclusion

7. Telephone services provided by the Department of Human Services (Human Services) through Smart Centres⁵ are an integral part of the Australian Government’s delivery arrangements for welfare services and income support provided through the Centrelink program. In 2013–14, the department handled 43.1 million telephone calls for Centrelink services—an average of around 800 000 calls per week—at a cost of some $338 million. The large volume of calls handled by the department is unique in comparison with other Australian call centres in either the public or private sectors.⁶ Many of the calls made by Centrelink customers are technically complex—relating, for instance, to the application of various income and asset tests—and may also involve support for customers with complex needs.⁷ Since 2012–13, the department’s key performance indicator (target KPI) for all telephone services to customers has been an average speed of answer of less than or equal to 16 minutes.⁸

8. The department faces the challenge of managing significant call volumes for Centrelink services while also transitioning to revised service delivery arrangements, through Smart Centres and self-service options. Smart Centres are undergoing a major reorganisation of work focused on achieving efficiencies and improved customer service by deploying staff more flexibly, introducing new technology and using a different system for distributing telephone calls. The department also has a long term strategy to move most customer transactions from a personal service basis (conducted by telephone or face-to-face) to a self-managed basis (conducted mainly over the internet) so as to focus customer service staff and telephone support on more complex services and customers most in need. However, while this transition is underway (for instance, mobile app transactions increased from 8.6 million in 2012–13 to 36.1 million in 2013–14), it will take time for customer behaviour to change and to realise expected benefits. In the interim the telephone remains a significant channel for customers seeking access to Centrelink services and assistance with online service channels, as digital services can vary in their ease-of-use and reliability.

9. Overall, the Department of Human Services is making progress in its transition to revised delivery arrangements for Centrelink services through its Smart Centre and self-service initiatives, while continuing to face challenges in managing a significant volume of telephone calls from Centrelink customers. The department is pursuing a number of useful reforms under its transformation program for Smart Centres—including the reorganisation of work, the introduction of new telephony technology and a digital strategy⁹—with the aim of improving overall efficiency and customer outcomes, including call wait times. Human Services has also established a soundly-based quality assurance framework for Centrelink telephone services, which should be extended to all relevant staff to improve the overall level of assurance. While Human Services’ data indicates that the department has met its overall target for all customer telephone services in the last two years¹⁰, the more detailed results for Centrelink telephone services show an increase in average speed of answer from well under 16 minutes in 2012–13 to over 16 minutes in 2013–14.¹¹ From a customer perspective, the 16 minute average speed of answer target for Centrelink telephone services is much higher than targets recently set for other telephony services provided by the department¹² as well as those set by other large Australian call centres. Further, the current target does not provide a clear indication of the wait times Centrelink telephone customers can generally expect, due to the distribution of actual wait times around the ‘average’. Centrelink customers also continue to experience high levels of call blocking¹³ and call abandonment¹⁴, which can further impact on the customer experience.

10. When customers call a Smart Centre to access Centrelink services they can have a variety of experiences. Of the 56.8 million calls made to Centrelink 1800 or 13 telephone numbers in 2013–14, 43.1 million calls were able to enter the network while 13.7 million calls were unable to enter the network, that is, the calls were blocked and the callers heard the ‘busy’ signal. Of the 43 million calls in 2013–14 that were able to enter the network, around 45 per cent were answered by a Service Officer (SO) and around a quarter were resolved in the Interactive Voice Response (IVR) system.¹⁵ The ANAO estimates that the remaining calls, around 30 per cent, were abandoned; that is the customer hung-up without resolving the reason for their call. The 45 per cent of calls that resulted in access to a SO waited for an average of 16 minutes and 53 seconds prior to talking to a SO and callers who abandoned their call after entering the queue to talk to a SO waited an average of 9 minutes and 42 seconds before hanging up. Reflecting these access issues, call wait times have been the largest single cause of complaint regarding Centrelink services in each of the past three years, with customer satisfaction with Centrelink telephone services falling to 66.6 per cent in 2013–14 from 70.4 per cent in 2012–13.¹⁶

11. Further, from the customer perspective, the department’s target KPI relating to average speed of answer does not clearly indicate what service standard customers can expect, due to the distribution of actual wait times around the ‘average’. In 2013–14, for example, for the top 10 Centrelink telephone lines¹⁷, 36 per cent of customers waited less than 10 minutes while some 30 per cent waited for more than 30 minutes.¹⁸ Other large customer service organisations, such as the Australian Taxation Office (ATO), express their call metric in a way that provides customers with more helpful information in this regard. By way of example, the target for the ATO’s general enquiry line is 80 per cent of calls answered within five minutes¹⁹, meaning that ATO customers can expect that most times their call will be answered within five minutes.

12. The department’s overall target for all customer telephone services—average speed of answer within 16 minutes—has been agreed with government. From 2014–15, the department will report separately on the performance of the Centrelink, Medicare and Child Support programs. While separate reporting is a positive development, there is no documented rationale as to why the revised targets set for average speed of answer for Medicare and Child Support services are significantly lower than the target for Centrelink services. Similarly, there is no clear reason why the department’s quality assurance mechanism for telephone calls—Quality Call Listening (QCL)—is not applied to all staff handling calls, including staff with less experience who may therefore be at greater risk of making errors. To improve the level of assurance provided by QCL, the department should apply the framework to all relevant staff.²⁰

13. It is a matter for the Government and the department to set service standards taking into account the resources available, systems capability and longer term strategies to shift customers to self-service. In this context, appropriate regard should be given to relevant industry benchmarks and the customer experience. The ongoing target for average speed of answer for Centrelink telephone services is very much at the upper end of the range of organisations and benchmarks examined. One consequence of high average wait times is that around 30 per cent of calls are abandoned by customers before the reason for the call is addressed. As mentioned, customer satisfaction with Centrelink telephone services is falling and ‘access to call centres’ is the largest cause of customer complaints about Centrelink services. Community stakeholders interviewed by the ANAO for this audit also drew attention to problems for customers relating to lengthy call wait times.²¹ Against this background, the department should review performance measures for Centrelink telephone services to clarify the service standards that customers can expect and to better reflect customers’ feedback and experience.

14. In a resource constrained environment, it is necessary to make choices about the allocation of limited resources. The department is pursuing a transformation program for service delivery, with a focus on realising efficiencies by transitioning customers to self-service where possible, and reserving telephone services for more complex cases and those customers most in need. Nonetheless, while this transition is underway, in the short to medium term the telephone remains a key access channel for Centrelink services and a way of providing assistance to those experiencing difficulties using digital channels, as evidenced by the very large volume of calls handled. Consistent with international trends, Centrelink customers are often using Smart Centres as a ‘help desk’ when accessing digital channels²² and while it is not known, at this stage, whether this trend is likely to be transitional or ongoing, there is a need to appropriately manage the telephony channel so as to provide a reasonable customer experience while also developing a viable pathway to the planned state.

15. The ANAO has made three recommendations focusing on: the implementation of a channel strategy to help deliver improved services across all customer service channels and a more coordinated approach to the management of call wait times; the application of the department’s quality assurance mechanisms to all relevant staff in Smart Centres; and the review of target KPIs to better reflect the customer experience and to clarify the service standards that customers can expect.

Key findings by chapter

Managing Customer Wait Times (Chapter 2)

16. Wait times for Centrelink telephone services have increased significantly in recent years from an average of 3 minutes and 5 seconds in 2010–11 to an average of 16 minutes and 53 seconds in 2013–14. Key factors underlying the increase include: reductions in the number of staff answering telephones; the performance and reliability of other customer service channels; and the more limited use of call blocking from late in 2011. While Human Services wishes to reduce the demand for telephone services by encouraging customers to use self-service channels, difficulties with using digital channels can lead to customers making telephone calls to seek assistance with online transactions. Comparisons with other large customer service organisations that deliver telephone services indicate that the department’s average wait times for Centrelink telephone services are very much at the upper end of contemporary service delivery standards.²³ Average wait times for Centrelink telephone services are also significantly above the separate targets recently set for the department’s Medicare and Child Support telephone services.²⁴ Further, Centrelink customers continue to experience call blocking and call abandonment.

17. Customer satisfaction with Centrelink telephone services is falling, ‘access to call centres’ is the largest cause of customer complaints about Centrelink services, and community stakeholders interviewed by the ANAO, as well as a 2014 Commonwealth Ombudsman’s report, drew attention to problems for customers relating to lengthy call wait times. For instance, Indigenous people in remote communities who may share the one telephone available in an Indigenous Agent’s premises sometimes take more than one day to access telephone services. In these circumstances, people may need to queue to use the telephone and may miss out if others’ calls are lengthy, as the Agent’s premises may only be open for a limited time each day.

18. In 2011–12, when call wait times began to increase significantly, the department responded by establishing a Call Improvement Taskforce (CIT). While the CIT did not operate within an overarching channel strategy it did have a useful five point strategy, providing a coordinating framework for its activities. Since the CIT was wound-up in 2013, the department has pursued a range of initiatives to reduce wait times, and recently advised that work was underway to scope the development of a channel strategy, consistent with sound practice²⁵, to help coordinate these initiatives within the context of the department’s transformation program for service delivery.²⁶

Managing Quality and Efficiency (Chapter 3)

19. A sound quality control framework can help provide assurance that Smart Centres deliver consistently high quality Centrelink telephone services. Further, the more efficiently such services can be delivered, within available resourcing, the greater the potential for improving productivity and/or reducing costs.

20. The department has two quality assurance frameworks in place to measure the accuracy and quality of its Centrelink telephony services: Quality On Line (QOL) and Quality Call Listening (QCL), both of which are underpinned by ‘check the checkers’ processes. The QOL framework is soundly-based and results for Smart Centres have been good with the target correctness rate of 95 per cent generally being met.

21. While the department has a well-established QCL framework for measuring and monitoring the quality of SOs’ customer interactions, which formally applies to all staff answering telephones, there are a number of gaps in the QCL framework’s implementation. Intermittent and Irregular employees (IIEs)—who will generally have relatively lower levels of experience and may therefore be at greater risk of making errors—are not currently included in QCL processes in Smart Centres.

22. The number of calls per SO required to be monitored in the QCL framework falls within the range of other call centres examined by the ANAO. However, the implementation of the required four calls per experienced SO and eight calls per new SO is low. Participation rates in calibration exercises for QCL evaluators are also relatively low. To maintain the integrity of the Quality Call Listening (QCL) process, the department should review the potential impact of these gaps in the implementation of QCL.

23. A number of measures have been used at various times to help the department assess the efficiency of Centrelink telephone services, including Average Handle Time²⁷ and First Call/Contact Resolution.²⁸ In recent years, the department’s Average Handle Time for Centrelink telephone services and First Contact Resolution have been broadly within the range experienced by other call centres. For instance, in 2013–14 the Average Handle Time for Centrelink telephone services was 8 minutes and 20 seconds compared to 9 minutes and 32 seconds for the ‘education and government’ sector of a global survey of call centres.²⁹ However, the department’s performance against both Average Handle Time and First Contact Resolution declined in 2013–14 relative to the previous year.³⁰ While transitional issues associated with the implementation of Smart Centres appear to have contributed to these declines in performance, continued analysis by the department of its performance against these key efficiency measures would help confirm the factors behind these trends. Analysis would also assist in evaluating delivery of the expected benefits from Smart Centres, new technology capabilities, and structural changes to the department’s telephony workforce.

Performance Measurement, Monitoring and Reporting (Chapter 4)

24. The department internally monitors the performance of Smart Centres against a range of useful, albeit traditional call centre metrics. A number of internal targets have also been established. While these metrics are an aid to assessing performance information at an operational level, they provide a more limited basis for assessing customer outcomes and the success or otherwise of the Smart Centre concept. There would be value in examining existing metrics and their fitness for purpose in the Smart Centre environment. In particular, the department could usefully focus on:

• First Call/Contact resolution—as improving resolution rates is a key goal of the Smart Centre concept;
• the IVR system—to assess its effectiveness in resolving calls without having to speak to a SO; and
• the interpretation of Average Handle Time—which has changed following the blending of telephony and processing work.

25. Human Services currently uses the average speed of answer as the single KPI for its public reporting on telephony services.³¹ This target KPI is both relevant and reliable for its intended purpose, as it addresses a significant aspect of the department’s telephone services and can be quantified and tracked over time. However, it is not complete as it does not provide insight into the range of customer telephony experiences, including the service levels that most customers can expect and the incidences of call blocking and abandoned calls.

Summary of entity response

The proposed audit report was provided to the Department of Human Services. The department’s summary response to the audit report is provided below. The formal departmental response is included at Appendix 1.

The Department of Human Services agrees with ANAO recommendations 1 and 2 and agrees with qualifications to recommendation 3. The department considers that implementation of recommendations 1 and 2 will further assist in the management of Centrelink telephony services.

The department has already commenced documenting its channel strategy to reflect the work already in place. Additionally, the department is developing a new Quality Call Framework which will enhance the informal and formal mechanisms already in place.

The department currently meets its agreement with Government through its Key Performance Indicator (KPI) for telephony which is answering calls with an average speed of answer of 16 minutes. The KPIs that the department has across all its services and channels are dictated by the funding available for the department to meet its obligations. The department has estimated that to reduce the KPI to an average speed of answer of 5 minutes, it would need an additional 1000 staff at a cost of over $100 million each and every year. The transformation of the department’s services over the past three years has been profound. In accordance with the Government’s agenda on digital services, customers now have a greater range of ways to do their business with the department. For some customers, the predominant method of contact with the department is now online. This trend will increase as the department builds more online capability and the Welfare Payments Infrastructure Transformation (WPIT) programme will shape the future of delivery for the department for the next decade.

Recommendations

Background

1.1 The Department of Human Services’ (Human Services) Centrelink program delivers a range of government payments and related services to Australians.³² In 2013–14 the department processed 3.7 million Centrelink program claims and paid over $108 billion to individual customers for Centrelink related transactions on behalf of the Australian Government.

1.2 One of the most common ways that customers contact Human Services is via the telephone. In 2013–14 around 43 million calls were handled by the department relating to the Centrelink program. Human Services manages over 50 Centrelink-related telephony lines, each with its own 1800 or 13 telephone number.³³ Each of the main payment types has its own line. There are also a number of ‘boutique’ lines for smaller groups of clients, for example, a line for agents³⁴ and a number of lines used for disaster relief such as that established for the Queensland floods in 2010–11.

1.3 Table 1.1 sets out the ten telephone lines which accounted for over 70 per cent of answered calls in 2013–14. The telephone line which had the most answered calls was the Families and Parenting line followed by the Employment Services line.

Table 1.1: Top ten telephone lines in 2013–14

Source: Department of Human Services data.

Note:

a Participation Solutions is a telephone line that interacts with customers who receive activity tested payments who have not met their activity obligations.

1.4 Calls are managed and distributed nationally through a virtual network. Customers are asked to identify the reason for their call and the call is then distributed to staff located across Australia according to ‘skill tags’³⁵ which identify staff with the required skills. Generally telephone lines are open from 8am to 5pm, Monday to Friday (except public holidays). Some lines, including the Families and Parenting line and the BasicsCard, are open for longer.³⁶ The emergency lines (to handle disaster recovery), which operate from the Geelong Smart Centre, are generally open 24 hours, 7 days per week.³⁷ Some telephone enquiry lines are prioritised differently. For example, callers using emergency lines receive priority.³⁸ Calls to the tip-off line, where people can report possible fraud, also receive a higher priority.

Smart Centres

1.5 Previously referred to as call centres, the department now provides telephone services through Smart Centres. Human Services operates 29 Smart Centres that provide Centrelink services. All of these centres also provide Medicare Public services and nine of these locations provide multi-lingual services. In 2013–14 a total of 56.8 million calls were made to Centrelink 1800 or 13 numbers—the processing of these calls is illustrated in Figure 2.1 in Chapter 2.

1.6 The shift to Smart Centres, which began in 2012–13 and is still being implemented, is changing the way work is organised in a number of ways. Previously, telephony and processing work were undertaken separately and both telephony and processing work were arranged by ‘customer segment’ and payment type. Staff specialised in one or more of these customer segments, which included Older Australians, Disabilities, Carers, and Families. In the call centre environment, there was some flexibility to move staff to cope with peaks in telephony workloads in particular customer segments but very limited flexibility to deploy staff across telephony and processing work.³⁹

1.7 One of the key objectives behind the move to Smart Centres is a desire to blend telephony and processing work. The new arrangements involve staff being cross-trained in both types of work and being used flexibly depending on work priorities and their skills. For instance, Service Officers (SOs) taking calls will also perform some limited processing work while the customer is on the telephone, to assist in achieving more timely outcomes for customers. Therefore the distinction between call and processing is becoming increasingly blurred as the implementation of Smart Centres proceeds.

1.8 Another key objective is moving away from work organised around payment types and customer segments to organising work around the complexity and frequency of customer transactions. Customer transactions and enquiries have now been categorised into a ‘Skills Pyramid’, with lower complexity
and/or high frequency transactions/enquiries making up the ‘general’ base layer of the pyramid. Higher complexity and less frequent transactions are found further up the pyramid in the ‘assisted’ and ‘managed and intensive’ layers.⁴⁰

Service delivery transformation

1.9 The Department of Human Services’ Strategic Plan 2012–16 outlines a high level strategy to transform service delivery:

• when appropriate, move transactions from a personal service basis (face-to-face or telephone) to self-managed mechanisms; and
• customer service staff are to focus on more complex services and helping those most in need rather than dealing with simple transactions. ⁴¹

1.10 Smart Centres’ Centrelink telephone services are being managed in the context of this high level strategy and within a dynamic service delivery environment of changing technology, customer expectations and preferences, and the pursuit of efficiencies in service delivery. Human Services’ customers can access services through an increasing variety of channels⁴² and performance in one channel increasingly affects other channels.⁴³ More recently available channels include the trialling of virtual face-to-face services through video conferencing, online transactional services, and a growing range of mobile apps which allow customers to view and update their personal information, fulfil their reporting obligations, apply for advance payments and read their online letters at a time and place that suits them.⁴⁴

1.11 Table 1.2 presents data on Centrelink self-managed transactions from 2011–12 to 2013–14. Consistent with the strategy for service delivery transformation, there has been growth in online transactions and very strong recent growth in mobile apps. However, the number of telephone self-service transactions have remained largely unchanged over this same period.

Table 1.2: Transactions for digital and online services 2011–12 to 2013–14

Source: Department of Human Services, Annual Report 2013–14, p. 110.

Funding

1.12 Table 1.2 summarises expenditure on Centrelink Telephony Services from 2010–11 to 2013–14. In 2013–2014 the department directed $337.9 million to providing Centrelink telephony services.

Table 1.3: Expenditure on Telephony services between 2010–11 and 2013–14

Source: Human Services data.

Note a: The largest component of total expenditure on Centrelink telephony is the staffing costs of the SOs answering telephones. However, since the introduction of Smart Centres from 2012 a proportion of SOs’ time has been spent on processing work thus making an accurate comparison of total expenditure on Centrelink telephony between years difficult.

1.13 From 2003–2004 to 2011–12, there was supplementary budget funding for Centrelink call centres to improve technology in call centres and to ‘ensure Centrelink is able to meet the demand arising from customers making increased use of call centres’.⁴⁵ The additional funding ranged from $15.4m in 2003–04 to approximately $50–60m from 2006–2007 onwards. In the 2012–13 Budget, the supplementary funding (of just over $50m per annum) was rolled into ongoing base funding. The 2013–14 Budget provided an additional $30million ($10m in 2012–13, $20m in 2013–14), on top of the increase in base funding from the previous Budget, to ‘allow additional staff to be employed to act as a ‘surge capacity’ in call centres’.⁴⁶

Recent reviews

1.14 The ANAO reported on the administration of the Australian Taxation Office’s contact centres in November 2014.⁴⁷ Where pertinent, the ANAO has drawn on that performance audit to provide an Australian public sector context.

1.15 In 2014, the Commonwealth Ombudsman conducted an investigation into access and service delivery complaints about Centrelink services, including Centrelink telephone services.⁴⁸ The Ombudsman’s report is discussed at times in this audit report.⁴⁹

Audit approach

Audit objective and criteria

1.16 The audit objective was to assess the effectiveness and efficiency of the Department of Human Services’ (Human Services) management of Smart Centres’ Centrelink telephone services.

1.17 To form a conclusion against the objective, the audit examined the management of Smart Centres’ Centrelink telephone services against the following high-level criteria:

• Human Services offers customers effective telephone services in relation to a range of quality indicators, for example, wait times and the accuracy of the information provided;
• Centrelink call services in Smart Centres are managed efficiently; and
• Human Services effectively monitors and reports on the performance of Smart Centres’ Centrelink telephone services.

Audit scope

1.18 The audit scope did not include:

• Smart Centres’ processing services other than the processing that is done as part of the telephone service; and
• Smart Centres’ Medicare and Child Support telephone services.

Audit methodology

1.19 The audit methodology included:

• an examination of the department’s files and documentation relating to the management of Smart Centres’ Centrelink telephone services;
• interviews with Human Services managers and staff involved in the management of Smart Centres, both in Canberra and in the call network;
• interviews with the department’s managers involved in the coordination and governance arrangements for Smart Centres and Human Services proposed channel strategy;
• interviews with external stakeholders including the Commonwealth Ombudsman’s Office and peak bodies such as People With Disability Australia and the Welfare Rights Network;
• comparisons with other large service delivery organisations; and
• site visits to a number of Smart Centres.

1.20 The audit was conducted in accordance with the ANAO’s auditing standards at a cost to the ANAO of approximately $459 000.

Structure of the report

1.21 The structure of the audit report is shown in Table 1.4.

Table 1.4: Structure of the Audit Report

Source: ANAO.

Introduction

2.1 The Department of Human Services’ Smart Centres are an integral component of the Australian Government’s delivery arrangements for Centrelink services. The operation of Centrelink telephone services can directly influence customer experiences of Commonwealth service provision, and wait times in particular can have a negative impact on individuals. In April 2014, the Treasurer, the Hon Joe Hockey MP, commented that the extended wait times for Centrelink telephone services are ‘… hugely expensive and undermine productivity. It undermines the capacity of people to get on with their own lives.’⁵⁰ The department has acknowledged that highly vulnerable, elderly, Indigenous and disabled income support customers use the call channel as their primary means of accessing the department, and that it needs to improve its performance in this area.⁵¹ Further, the department has made a commitment in its Service Charter to its customer: ‘Easy access to services: We will give you quick and easy access to the right services’.⁵²

2.2 In this chapter, the ANAO examines:

• the customer experience of Centrelink telephone services;
• data relating to call demand and wait times including variations in wait time across different Centrelink telephone lines;
• levels of customer satisfaction and complaints relating to Centrelink call services; and
• departmental strategies to improve call wait times.

The Customer experience of Centrelink telephone services

2.3 Figure 2.1 shows the variety of experiences customers can have when calling the various 1800 and 13 numbers for Centrelink services.

2.4 Figure 2.1 indicates that of the 56.8 million calls made to Centrelink 1800 or 13 telephone numbers in 2013–14, 43.1 million calls were able to enter the network while 13.7 million calls were unable to enter the network, that is, the calls were blocked and the callers heard the ‘busy’ signal. While this is a large number of blocked calls, and likely to have caused frustration and inconvenience for customers, it should be noted that customers whose calls were blocked do not necessarily miss out on accessing Centrelink services as they could have: made multiple calls until they were able to enter the network; rung back at a later time; or used another channel to access Centrelink services (such as online services).⁵³

Figure 2.1: Pathway of calls through Centrelink Telephony Services in 2013–14

Source: ANAO analysis of Human Services’ documentation, based on 2013–14 data.

Notes to Figure 2.1:

a The number of blocked calls, 13.7 million, does not necessarily represent the number of customers who received the ‘engaged’ or ‘busy’ signal. The number of customers is lower because customers will often make multiple calls attempting to enter the network.

b In 2013–14, 12.1 per cent of answered calls used the place-in-queue option which allowed them to be called back rather than waiting on the line.

c This is a maximum figure, based on the assumption that no transferred calls go back into the IVR system, that is, they are all transferred into a queue to talk to a SO.

d The total number of calls entering the queue is equal to the total number of abandoned calls plus the total number of answered calls. Transferred calls are a subset of answered calls. If it is assumed that all transferred calls return to a queue rather than to the IVR system, there would be 25.3 million calls entering a queue for the first time with a further 3.3 million calls re-entering the queue.

2.5 Figure 2.1 also indicates that of the 43.1 million calls that were able to enter the network, 20.8 million calls—some 45 per cent of the calls that entered the network—were answered by a SO. Of the remaining 55 per cent of calls:

• 17.8 million calls (around 41 per cent of calls entering the network) end in the Interactive Voice Response (IVR) system⁵⁴ either because:
  • the customer’s enquiry was answered by the IVR message (no data available);
  • the customer chose an IVR option that transferred them into a self-service application, for example, to report their latest fortnightly income. In 2013–14, 5.5 million transactions were completed in IVR self-service—comprising roughly one-third of calls that end in the IVR system; or
  • the customer abandoned the call for a variety of reasons including that, after being advised by the IVR system of the estimated wait time, they decided it would take too long to access services (no data available);
• 7.8 million calls were abandoned after they had entered a queue to speak to a SO, after waiting an average of 9 minutes and 42 seconds in the queue. As indicated above, data on the number of calls that are abandoned in the IVR system is not available, therefore the proportion of total calls abandoned cannot be precisely calculated. An estimate made by the ANAO⁵⁵ suggests that around 30 per cent of calls that enter the network are abandoned either in the IVR system or after the customer has entered the queue to talk to a SO.

2.6 In summary, around 45 per cent of calls made by customers that are able to enter the network are answered by a SO, an estimated 30 per cent of calls are abandoned, and the remainder are resolved in the IVR system either because the recorded messaging answers their query or the customer utilised IVR self-service options.

Demand for Centrelink telephone calls

Table 2.1: Telephone calls to Centrelink 2009–10 to 2013–14

Source: ANAO analysis of departmental data.

Notes:

a The annual data from 2011–12 onwards (unshaded) is affected by the department’s decision in late 2011 to make less use of call blocking to manage demand. Therefore comparisons with earlier years (shaded) need to take account of the different use of call blocking. See paragraph 2.8 for more detail.

na Not available.

2.7 Table 2.1 indicates that based on the number of calls entering the network, demand for Centrelink phone services has stabilised in the past two years after strong growth in the preceding years. Calls entering the network grew very strongly over the three years to 2011–12, and then stabilised at around 42 to 43 million in 2012–13 and 2013–14. Calls answered by SOs have also stabilised in 2012–13 and 2013–14 at around 20 million calls, down from a high of 28.8 million calls in 2010–11.

2.8 Table 2.1 also indicates that the number of blocked calls was very large until relatively recently, falling from 39.9 million in 2010–11 to 13.7 million in 2013–14. The department has advised that in the past, the then separate Centrelink used call blocking as an operational strategy to prevent wait times becoming too long. While call blocking provides a reasonably clear message to ‘try again’ at another time and can save a customer’s time in waiting for their call to be answered, there are disadvantages in the use of call blocking as customers may need to call again repeatedly or have to ring back at a time that is less convenient. The department further advised that since late 2011, it has made more limited use of the engaged signal to manage demand so that: it could more accurately gauge total demand; more callers could access telephone self-service; and callers could have the option of being advised of estimated wait times via the IVR system and then make the decision to wait or call back another time. The department indicated that this change of operational strategy in relation to call blocking was a major factor in the increase in average wait times at that time.⁵⁶ Human Services advised the ANAO that currently it only blocks calls when the telephony system would otherwise become overloaded or at the end of a day when it is clear that calls will not be answered within the remaining business hours for that day. This approach more than halved the number of blocked calls in 2013–14 (13.7 million) compared to 2012–13 (32.5 million).⁵⁷

Call wait times and abandoned calls

2.9 Table 2.2 presents data relating to call wait times, abandoned calls, and staffing from 2009–10 to 2013–14.

Table 2.2: Metrics relating to call wait times, staffing levels, and abandoned calls (2009–10 to 2013–14)

Source: ANAO analysis of Department of Human Services data.

Notes:

a The KPI used by the then Centrelink in relation to customer wait times in 2009–10 and 2010–11 was the percentage of calls answered within 2 minutes and 30 seconds. The target was 70 per cent. The target of 16 minutes for average speed of answer used by Human Services since 2012–13 is for all telephony services, not just Centrelink telephone services.

b The number of SOs on a Full-Time Equivalent basis paid out of the telephony budget. However, since the introduction of Smart Centres a proportion of SOs’ time is spent on processing work thus making an accurate comparison of FTE staff answering calls between years difficult. The department adjusted the FTE staffing numbers downwards in 2012–13 and 2013–14 for the intensive commencement training undertaken by IIEs as these staff were not available to answer phones. The data for 2013–14 has also been adjusted downwards for the department’s estimate of the processing work undertaken by telephone staff (10 per cent of ongoing staff and all weekend hours for IIE staff).

c The number of abandoned calls (calls where the caller hangs up after they have entered the queue to talk to a SO) as a percentage of the calls that enter the network. This data does not include calls abandoned in the IVR system.

d The annual data from 2011–12 onwards (unshaded) is affected by the department’s decision to make less use of call blocking since late 2011 to manage demand. Therefore comparisons with earlier years (shaded) need to take into account the different use of call blocking. See paragraph 2.8 for more detail.

na Not available.

2.10 Table 2.2 indicates that the average speed of answer—a metric that Human Services uses to measure call wait times⁵⁸—deteriorated significantly from three minutes and five seconds in 2010–11 to 11 minutes and 45 seconds in 2011–12. As indicated earlier the department advised that the more limited use of call blocking was a significant factor in the increase in call wait times. The number of full-time equivalent SOs also fell by 19 per cent in 2011–12. The reduction in frontline staff answering calls in 2011–12 coincided with a threefold increase in call wait times. The number of SOs again fell significantly by 8.3 per cent in 2013–14, coinciding with an increase in average speed of answer from 12 minutes and 5 seconds to 16 minutes and 53 seconds. The data indicates that staff reductions have impacted on call wait times.

2.11 Table 2.2 also shows that the rate of abandoned calls, once a caller has gone through the IVR system and entered the queue to talk to a SO, was similar in 2011–12 and 2012–13, at around 15 to 16 per cent of all calls that entered the network—but increased to 18 per cent in 2013–14 as average waiting times increased by more than 30 per cent. As discussed in paragraph 2.4 and in the box below, ANAO analysis indicates that in 2013–14 around 30 per cent of calls made to access Centrelink services were abandoned when account is taken of calls abandoned in the IVR system.⁵⁹

Note a: The ANAO based its estimate of time lost in person-years on the average time customers waited on the phone prior to abandoning the call after entering a queue to talk to a SO. The estimate did not include the time taken to navigate the IVR system and did not including any calls abandoned within the IVR system. See Appendix 3 for more details on the methodology.

2.12 The target KPI used by the department in its public reporting of telephony services—average speed of answer of less than or equal to 16 minutes—gives limited information on the variability in the wait times experienced by customers. Table 2.3 presents data on answered calls by time intervals for the ten telephone lines that accounted for around 70 per cent of answered telephone calls in both 2012–13 and 2013–14.

Table 2.3: Answered calls by telephone line and time interval

Source: Department of Human Services data.

2.13 The 2013–14 data indicates that just over one third of calls (36 per cent) were answered within 10 minutes, with around half of all calls answered within 20 minutes. Thirty per cent of callers waited over 30 minutes. This data is a significant deterioration from the previous year when only 15 per cent of callers waited more than 30 minutes. Callers on different lines also had different experiences, for example, in 2013–14 over half of customers calling the Participation Solutions line waited over 30 minutes whereas nearly all calls on the Income Management and Tip-off lines were answered within 20 minutes. The department advised that for certain telephone lines, such as ‘Youth and Students’, decisions have been taken to distribute resources to allow relatively longer wait times on these telephone lines to act as an incentive for this cohort to move to self-service using other channels such as apps and online.

Comparison of call wait time targets

2.14 To gain some perspective on the department’s approach to measuring its performance on call wait times, the ANAO reviewed the timeliness performance measures used by a selection of large government and non-government service delivery organisations. While there was variation in the targets adopted by organisations, a typical target for response times was a majority of calls answered within two or three minutes.

2.15 Since 2012–13, Human Services has used an average speed of answer of less than or equal to 16 minutes as the single KPI for external reporting purposes for all telephony services including Medicare and Child Support services. The department advised that it developed this KPI based on internal information and modelling, including consideration of the department’s funding and projected resourcing. However, departmental documentation about the decision to establish this KPI was not available to the ANAO. From 2014–15 the department will report Centrelink, Medicare and Child Support telephony performance separately against the KPI of average speed of answer. From the separate 2014–15 targets listed in Table 2.4, it can be seen that targets for Medicare and Child Support customers are significantly lower than the 16 minute target for Centrelink phone services, at 7 minutes and 3 minutes respectively. There is no documented rationale as to why these targets differ.

Table 2.4: Timeliness performance measures of a selection of organisations

Source: ANAO analysis.

2.16 A global benchmarking report⁶⁰ of contact centres indicates that the overall average speed of answer in 2013 for the 817 organisations included in the report was 30 seconds; and for the Education and Government sub-sector, the overall actual achievement was 49 seconds in the same year. These comparisons suggest that both the actual wait time for Centrelink telephone services and the department’s relevant performance target (less than or equal to 16 minutes) are very much at the high end compared to a number of other government and private sector organisations that deliver services via the telephone including for the Medicare and Child Support telephone services provided by Human Services.

Customer satisfaction and complaints

2.17 Lengthy call wait times and call blocking impact on customer satisfaction and are a major source of complaints. The global benchmarking report on contact centres noted above observes that customer satisfaction and complaints are the top indicators of the operational performance of telephone services.⁶¹

Customer satisfaction

2.18 Human Services measures customer satisfaction of telephony services provided by Smart Centres on a regular basis.⁶² The overall satisfaction rate with Centrelink telephony services was 70.4 per cent in 2012–13 and dropped to 66.6 per cent in 2013–14. This result is significantly below the average customer satisfaction rate for 2013 in the global benchmarking report of call centres of 79.2 per cent.⁶³

Complaints

2.19 Table 2.5 presents data on complaints to Human Services about Centrelink services from 2009–10 to 2013–14.

Table 2.5: Complaints (per cent of total complaints about Centrelink services) (2009–10 to 2013–14)

Source: Data provided by Human Services.

2.20 The department’s complaints data indicates that the largest single cause of complaint regarding Centrelink services over the past three years has been ‘access to the call centre’. In 2013–14, this category comprised 23.5 per cent of complaints, of which 18.8 per cent concerned wait times and four per cent related to the use of the busy signal or call blocking. The impact of the reduced use made by the department of call blocking since late in 2011 can be seen in the data in Table 2.5—since 2011–12, complaints about hearing the ‘call busy’ signal reduced while complaints about call wait times increased.

2.21 In 2014, the Commonwealth Ombudsman conducted an investigation into access and service delivery complaints about Centrelink services, including Centrelink telephone services.⁶⁴ Typical complaints reported included:

• waits of up to one hour in telephone queues;
• wait times far in excess of the estimated wait times customers were told via a recorded message at the start of their call;
• costs of call wait times when using mobile telephones;
• calls transferred between telephone lines compounding wait times; and
• customers being told to call another number once the initial call was answered after a considerable wait time.

2.22 A number of community stakeholders interviewed for this audit⁶⁵ also drew attention to problems relating to lengthy call wait times; for instance: people with a disability often access telephone services with the assistance of a carer and in these circumstances both the person with the disability and the carer experience lengthy wait times; and Indigenous people in remote communities who may share the one telephone available in the Indigenous Agent’s premises sometimes take more than one day to access telephone services. In these circumstances, people may need to queue to use the telephone and may miss out if others’ calls are lengthy as the Agent’s premises may only be open for a limited time each day.

Call Improvement Taskforce

2.23 In July 2012, the department established a Call Improvement Taskforce (CIT) in the context of significant increases in call wait times for Centrelink telephone services. The establishment of the CIT followed a review which recommended ‘a prioritised action plan for improvements and investment within the telephony processing channel’.⁶⁶

2.24 The CIT operated under two phases. The first phase, from July to November 2012, focused on identifying, implementing and monitoring short term projects that could result in ‘quick wins’ in terms of reducing call wait times. The second phase, from December 2012 to June 2013, focused on the ongoing monitoring of a small number of projects from phase one, along with the implementation of further short term improvement opportunities, combined with analysis of issues requiring more in-depth study or more complex implementation arrangements.

2.25 In both phases the membership of the CIT involved the department’s senior leadership and was chaired at the Associate Secretary level, with representatives at the Deputy Secretary, General Manager (SES Band Two) and National Manager levels (SES Band One) from a range of relevant areas. In both phases the CIT met regularly: in phase one, the CIT met twice a month; moving to around monthly in phase two. A Senior Responsible Officer at the SES Band Two level was made responsible for implementing Taskforce outcomes and to oversee both phases of the CIT.⁶⁷

2.26 The CIT developed a five point strategy which covered:

• reducing the need for customers to call, for example, improving the quality of letters—unclear letters generate telephone calls;
• increasing the use of self-service, for example, online services and mobile apps;
• improving call centre technology, for example, the introduction of
  
  place-in-queue call back (see paragraph 2.30);
• increasing the efficiency of the call network, for example, better balancing the rostering of employees and call demand; and
• adjusting resources to cope with demand, for example, use of Smart Centres to flexibly utilise staff across telephony and processing work to assist in dealing with periods of peak demand.

2.27 While the CIT did not operate under the framework of an overall channel strategy⁶⁸, the five point strategy covered the key factors that impacted on call wait times including the inter-relationships between channels. The five point strategy was used to organise the initiatives that were implemented under the CIT.⁶⁹

2.28 Overall, the CIT had effective processes for monitoring and reporting on initiatives.⁷⁰ The framework for reporting progress on each initiative and stream included an overall status report using a traffic light reporting system and a benefits realisation component.⁷¹ The reporting on each initiative contained an estimate of one or more of the following issues:

• potential reduction on call wait times (expressed as minutes of calls);
• projected reduction in call volume (as minutes of calls);
• diversion of staff to telephones (in full-time equivalent staff numbers); and
• other measures such as number of claims to be completed.

Management of call wait times after the Call Improvement Taskforce

2.29 At the end of June 2013, the CIT ceased operating and the initiatives that were ongoing or incomplete were transferred to a line area within the department for management on a ‘business as usual’ basis. Central monitoring and reporting of these initiatives as a package of inter-related measures no longer occurred. The department has since pursued a range of initiatives to manage call wait times in the context of its service delivery transformation strategy to transfer, where possible, customers to self-service channels. These measures are outlined in the following paragraphs.

Place–in–Queue call back

2.30 Human Services began trials of place-in–queue call back technology on a number of its telephone lines in 2012 to reduce the impact on customers of long call wait times. Access to place-in-queue call back has been progressively expanded over subsequent years. Place-in-queue allows customers to choose to be called back instead of staying on the line waiting for their call to be answered.⁷² Place-in-queue is automatically offered to eligible customers when the estimated wait time in a queue is more than five minutes. It is available on nine Centrelink telephone lines although, for two of those lines, it is only offered to customers calling from mobile telephones.⁷³ In 2013–14, 12.1 per cent of answered calls used the place-in-queue option.

Smart Centres

2.31 The shift from a traditional call centre to the Smart Centre concept is a key initiative in the department’s efforts to reduce call wait times. In the previous call centre environment, there was limited flexibility to move staff to cope with peaks in workload for particular customer segments/payment types or between telephony and processing. This lack of flexibility contributed to longer wait times in peak times; customers were also transferred between SOs if their call dealt with issues crossing payment types.

Blending telephony and processing work

2.32 One of the key components of the Smart Centre concept is to blend telephony and processing work to increase flexibility in the deployment of staff. This approach involves staff being cross-trained in both types of work and being used flexibly depending on work priorities and their skills. Enabling technology, once fully functional⁷⁴, will allow call and processing work to be prioritised and managed centrally, with workflows distributed across all work stations in Smart Centres. SOs taking calls also perform some limited processing work while the customer is on the telephone to assist in achieving more timely outcomes for customers. Therefore, as previously mentioned, the distinction between telephone work and processing work is becoming increasingly blurred.

2.33 During the ANAO’s site visits for the audit, it was clear that the blending of telephony and processing work was still in the transitional stage. While telephony staff were commonly undertaking processing tasks during the call, and were being offered overtime to complete processing work on the weekends, it was less common that processing staff were undertaking telephony work. The ICT equipment to allow work to be blended across all telephony and processing staff was not yet in place at all workstations. However, many of the SOs and managers interviewed by the ANAO in Smart Centres were positive about the benefits already achieved from blending, including the opportunity for more varied work for staff and more First Contact Resolution⁷⁵ for customers.

Source: ANAO analysis of Human Services documentation.

Skills Pyramid

2.34 Another key component of Smart Centres is moving away from work organised around payment types and customer segments to work being organised around the complexity and frequency of customer transactions. Customer transactions and enquiries have been categorised into a ‘Skills Pyramid’. Lower complexity and/or high frequency transactions/enquiries make up the ‘general’ base layer of the pyramid while higher complexity and less frequent transactions are found further up the pyramid in the ‘assisted’ and ‘managed and intensive’ layers. Figure 2.2 illustrates the Skills Pyramid.

Figure 2.2: Skills pyramid

Source: ANAO analysis of Human Services’ documentation.

2.35 It is intended that all SOs are to be trained in handling calls involving transactions and queries in the ‘general layer’ of the pyramid, while calls categorised as being in the ‘assisted’ or ‘managed and intensive’ layers would be dealt with by a subset of SOs with skill tags indicating more specialised skills.⁷⁶

2.36 In recent years, the department has increasingly employed Intermittent and Irregular employees ⁷⁷ (IIEs). At June 2014, 902 full-time equivalent IIEs were employed comprising 26 per cent of the front-line staff providing Centrelink telephone services. The IIE workforce, employed at the APS 3 classification, has assisted with the implementation of the Skills Pyramid as they are trained, and undertake work, in the ‘general’ layer. The ongoing workforce, generally at the APS 4 classification, while also able to complete transactions in the ‘general’ layer, focuses on the more complex work in the ‘intensive’ and ‘assisted’ layers of the pyramid.

2.37 ANAO visits to Smart Centres during the audit indicated that the Skills Pyramid was still in the implementation phase. The ANAO observed that many IIE staff were in the process of being trained in transactions in the ‘general’ layer (putting considerable strain on training and technical support resources) while some ongoing staff were still only skill tagged in their traditional specialist area based on customer segment/payment type. A majority of Smart Centre managers interviewed were of the view that the Skills Pyramid had the potential to organise call distribution more efficiently and therefore to assist in reducing call wait times. Some ongoing staff interviewed by the ANAO, however, pointed to problems that can arise if IIE staff do not have a broad contextual understanding of the interrelationships between various payments and allowances.⁷⁸ Ongoing staff also commented on their experience of receiving high rates of transfers of calls from IIE staff, who did not have the skills or delegations necessary to complete certain calls.

Source: ANAO analysis of Human Services documentation.

Managed Telecommunications Services

2.38 In October 2012, the department entered into a five year contract with Telstra for telephony services—the Managed Telecommunications Services Contract (the MTS). The department advised the Senate at that time that the MTS had the potential to reduce call wait times by, amongst other things, improving capacity to distribute workloads, enabling the use of speech analytics to assist in analysing why customers call, and better queue management at the national level.⁷⁹

2.39 The MTS is yet to deliver these benefits for Centrelink calls due, in large part, to the department’s higher priority to migrate Medicare and Child Support programs to the new telecommunications platform. However, the department advised that this migration work and many of the other capabilities enabled by the MTS will become available over the course of 2015. Several of these capabilities have significant potential to assist in reducing call wait times or alleviating the impact of wait times on customers. These capabilities include: a new workforce management system (which will replace the existing tool for forecasting call demand and matching rostering of staff); call recording and speech analytics (this can assist in investigating the factors driving call demand); voice biometrics which aims to reduce the effort for customers to authenticate when they call the department; an enhanced call back facility where the customer may request a call back for a specified time; the capacity to deliver any call to any appropriately skilled staff member anywhere across the department; and a more sophisticated IVR system which, for example, could provide a ‘progress of claim’ self-service option. ⁸⁰

Digital Strategy

2.40 A key plank in the department’s approach to reducing call wait times is to shift customers from telephone services to self-service digital channels. Both online transactions and the use of mobile apps have increased strongly over the past three years (see Table 1.2 in Chapter 1). However, the growth in digital transactions has not reduced the demand for call services as anticipated. The department’s experience is consistent with wider international trends, and a recent global benchmarking report observed that ‘the emergence of multichannel customer management is fuelling growth within the contact centre industry, not curtailing it’. ⁸¹ Part of the explanation for this trend is the unreliability and difficulty of using some digital applications. International experience indicates that some 40 per cent of calls to contact centres are now made because of a failed self-service interaction; a noteworthy trend given the expectations of improved service delivery and efficiencies offered by self-service channels.⁸² In effect, customers are using Smart Centres as a ‘help desk’ for digital channels, and at this stage it is not known whether this trend is likely to be transitional or ongoing.

2.41 The department is currently developing a draft digital strategy. The draft strategy acknowledges the linkages with other channels and the need to align overall strategy with the Smart Centres operating model. There is a focus in the draft strategy on improving the reliability and ease of use of digital transactions and expanding the range of transactions that can be completed digitally. Improvements in digital service channels can be expected to assist in reducing call demand, call wait times and related costs for customers and the department.

A coordinated channel strategy

2.42 In April 2006, the Australian Government Information Management Office (AGIMO) released the guide Delivering Australian Government Services: Managing Multiple Channels. The guide emphasised the importance of entities establishing a channel strategy to manage service delivery to their clients through the most appropriate channels. A channel strategy is a ‘set of business driven choices aimed at delivering services to customers efficiently and effectively using the most appropriate mix of channels for the customer and the agency’.⁸³ AGIMO identified the following benefits of a channel strategy:

• the alignment of customer needs, services, channels and agency priorities;
• improved cost efficiency of service delivery across multiple channels;
• seamless, integrated and consistent delivery of services across channels; and
• informed and prudent future channel investments.⁸⁴

2.43 The AGIMO guide also encouraged organisations to forecast future channel usage patterns as part of their channel strategies.

2.44 At the time of the audit fieldwork, Human Services did not have a detailed channel strategy (covering all channels including digital channels) but has advised that work is underway to scope the development of a strategy. The department has developed a high level Service Delivery Operating Model with more detailed channel operating models to underpin it (comprising a draft Digital Strategy, a Smart Centre Operating Model and a Front of House Operating Model). The purpose of the Service Delivery Operating Model is to assist staff to understand how the separate channel operating models can work together to achieve government outcomes and will assist in the development of the detailed channel strategy. To help progress its work, the department should establish a pathway and timetable for the development of a coordinated channel strategy, and also consider options for learning further from the experience of other large service delivery organisations which operate multiple customer channels.

Recommendation No.1

2.45 To help deliver improved services across all customer channels and a more coordinated approach to the management of call wait times, the ANAO recommends that the Department of Human Services establish a pathway and timetable for the implementation of a coordinated channel strategy.

Human Services’ response:

2.46 The department agrees with the recommendation.

2.47 The department has a Service Delivery Operating Model (SDOM) that outlines the four levels of services that it provides to its customers based on their individual needs (naturally connected, self-managed, assisted and managed and intensive). A coordinated channel strategy is being documented to reflect the work that the department has put in place.

Conclusion

2.48 Human Services’ Smart Centres play a key role in the department’s capacity to deliver payments and services efficiently and effectively on behalf of the Australian Government. The factors underlying the significant increase in wait times in recent years (and the related issues of high rates of call blocking and abandoned calls) are complex and inter-related. Reductions in staff, the performance and reliability of other channels, including digital channels, and changes in the department’s use of call blocking, have all played a role. However, the growth in digital transactions has not reduced the demand for call services proportionately, consistent with wider international experience which indicates that some 40 per cent of calls to contact centres are now made because of a failed self-service interaction.

2.49 Comparisons with other organisations that deliver telephone services indicate that the department’s average wait times for Centrelink telephone services are very much at the upper end of contemporary service delivery standards, and are significantly above the separate targets recently set for Medicare and Child Support telephone services. One consequence of high average wait times is that around 30 per cent of calls are abandoned by customers before the reason for the call is addressed. The department’s records indicated that customer satisfaction with Centrelink telephone services is falling and ‘access to call centres’ is the largest cause of customer complaints about Centrelink services. Community stakeholders interviewed by the ANAO for this audit and a 2014 Commonwealth Ombudsman’s Report, also drew attention to problems for customers relating to lengthy call wait times.

2.50 In 2011–12, when call wait times began to increase significantly, the department responded by establishing a CIT. While the CIT did not operate within an overarching channel strategy it did have a useful five point strategy, providing a coordinating framework for its activities. Since the CIT was wound-up in 2013 the department has pursued a range of initiatives to reduce wait times. Human Services has advised that it has started to develop a channel strategy, consistent with good practice, which will consider these initiatives and the department’s transformation program for service delivery—transitioning customers to self-service where possible, and reserving telephone services for more complex cases and those customers most in need. The blending of telephony and processing and the introduction of the Skills Pyramid in Smart Centres, the Digital Strategy and the MTS all have the potential to contribute to reducing call wait times as further progress is made with their implementation. To help deliver improved services across all customer channels and a more coordinated approach to the management of call wait times, Human Services should establish a pathway and timetable for the implementation of a coordinated channel strategy, and also consider options for learning further from the experience of other large service delivery organisations operating multiple customer service channels.

Introduction

3.1 A sound quality control framework underpins the provision of effective client service for customers using telephone services, and can help minimise the risk of payment errors occurring in telephone transactions. In addition, measuring and monitoring efficiency, including through the use of key metrics, can assist in identifying and assessing potential efficiency gains and cost savings.

3.2 The ANAO examined the department’s two quality control mechanisms operating in Smart Centres: Quality Call Listening (QCL) and Quality On Line (QOL).

3.3 To assess Human Services’ overall efficiency in providing Centrelink telephone services, the ANAO examined:

• two efficiency measures—Average Handle Time and First Contact/Call Resolution; and
• two cost measures—‘cost per call’ and ‘cost per minute’.

Quality Call Listening

QCL process and results

3.4 The use of QCL is consistent with standard global practice for contact centres to assess the quality of calls. QCL is guided by the department’s Quality Call Listening Framework85 (the framework), which outlines the expected level of quality services to be provided by Smart Centres. The framework is expected to be applied to all SOs answering telephones including non-ongoing staff and Intermittent and Irregular employees (IIEs). The results from QCL are used by individual sites to monitor SO performance and identify key systemic issues to improve the accuracy of information and quality of service delivered to customers.

3.5 Under the department’s framework, four calls are selected for QCL per experienced SO in each four week settlement period. New SOs have eight calls selected each settlement period for the first six months of employment. Calls are assessed side-by-side with an evaluator listening to the call live.⁸⁶ Upon completion of a call, the evaluator provides immediate feedback to the SO by identifying strengths and areas which require further improvement.

3.6 Evaluators assess each call against 22 quality indicators outlined in the department’s Quality Checking Guide. The overall result determines whether it is a ‘quality’ or ‘non-quality call’. While the department does not have a clear scoring guide for ‘quality’ and ‘non-quality calls’, failure to satisfy eight key quality indicators will result in an automatic ‘non-quality’ call.⁸⁷ In addition to assessing performance against these eight criteria, evaluators use their professional judgement against the criteria as a whole to determine whether the call is a ‘quality’ or ‘non-quality’ call. The reliance on professional judgement highlights the importance of evaluators participating in calibration exercises, which help reduce subjective assessments and promote consistent application of the framework standards.

Comparisons with other organisations

3.7 The ANAO compared the sampling approach used by the department for QCL with two government and two large private contact centres and with a global benchmarking report, as summarised in Table 3.1.

Table 3.1: Comparison of the sampling approach used for QCL

Source: ANAO analysis.

Note:

A Dimension’s Data 2013/14 Global Contact Centre Benchmarking Report.

3.8 The number of calls evaluated by the department as part of QCL falls within the range of the other four Australian organisations compared in Table 3.1. The distinction made by the department between experienced and new SOs in call selection for QCL follows standard global call centre practice to evaluate more calls for less experienced SOs. The department has adopted a risk-based approach which monitors less experienced SOs on a more regular basis, to help promote quality.

QCL results

3.9 Table 3.2 summaries departmental data on the number of quality and non-quality calls as assessed under the QCL framework from 2009–10 to 2013–14. Overall, the rate of quality calls has increased from 93.85 per cent in 2009–10 to 96.30 per cent in 2013–14, with the results indicating that departmental SOs achieved high rates of quality calls in QCL.

Table 3.2: QCL – quality and non-quality calls (2009–10 to 2013–14)

Source: Human Services data.

Implementation of the QCL Framework

3.10 The department’s ability to maintain high rates of quality calls depends, in part, on the consistent and rigorous application of the QCL framework.

Ongoing and non-ongoing employees

3.11 The QCL framework requires that the QCL checking rates of four calls per experienced SO and eight calls per new SO should be completed at an average rate of 95 per cent across all relevant staff.⁸⁸ Smart Centres failed to meet this target in 2013–14. Table 3.3 sets out the rate of calls reviewed in accordance with the QCL framework and the subsequent calls that were rated as quality calls over the 13 settlement periods in 2013–14.

Table 3.3: Implementation of QCL Framework (2013–14)

Source: Human Services data.

Note:

a The QCL framework provides for an average checking rate of 95 per cent across all relevant staff.

3.12 Table 3.3 (column 2) indicates that the department did not implement the requirement that QCL is applied to an average of 95 per cent of SOs at the expected rate (four calls per experienced SO and eight calls per new SO) in any settlement period in 2013–14. The department advised that the very low results for settlement periods one to three in particular resulted from a decision that QCL was not mandatory during this peak period. While overall quality may not be affected by short-term or periodic pauses in the application of the QCL framework, there is a risk that longer-term pauses may result in the framework not providing the intended level of quality assurance. Table 3.3 (column 3) further indicates that the calls that did undergo QCL generally met the target of 95 per cent being rated as ‘quality calls’.

Intermittent and Irreguar employees (IIEs)

3.13 The QCL framework was not applied to IIEs at the Smart Centre level.⁸⁹ While Smart Centre managers advised the ANAO that calls taken by IIEs were generally subject to ad hoc quality evaluation these employees were not subject to the formal requirements of the QCL framework. Rates of quality evaluation for IIEs interviewed by the ANAO varied: from four calls evaluated per settlement period; to no calls evaluated over the course of their employment. Not applying the QCL framework to IIEs risks weakening the department’s quality assurance processes for Centrelink telephone services. There would be merit in reviewing the current approach given recent growth in this employment category and the likelihood that it is the department’s least experienced cohort.

Recommendation No.2

3.14 To maintain the integrity of the Quality Call Listening (QCL) process and improve the level of assurance on the quality and accuracy of Centrelink telephone services, the ANAO recommends that the Department of Human Services applies the QCL framework to all staff answering telephone calls, and reviews the potential impact of gaps in the implementation of QCL.

Department of Human Services’ response:

3.15 The department agrees with the recommendation.

3.16 The department has a number of quality measures both informal and formal to assure quality outcomes. All staff, including intermittent and irregular employees, are subject to a mix of both informal and formal quality measures to ensure accuracy of outcomes. In addition, the department is developing a new Quality Call Framework which will apply to all staff who have contact with customers in the telephony environment.

Monitoring and reporting of QCL results

3.17 Performance monitoring and reporting is used at both the operational and strategic levels within Human Services. Smart Centre managers use site level reports to monitor the implementation of QCL at their site, and to identify and address issues of non-performance of individual SOs. At the strategic level, the national Quality Listening Team produces a regular report on QCL with a breakdown of national results reported by each Smart Centre site and a summary of the top five errors contributing to a non-quality call. The ANAO examined the reports of 13 settlement periods from August 2013 to September 2014; the key error identified in each report was ‘transitioning customers to
self-service options’. A departmental report analysing six settlement periods between 2011–12 and 2012–13 also confirmed this issue was an ongoing factor accounting for 30 per cent of non-quality calls.

3.18 The department has implemented various initiatives to address the issue of SOs not transitioning customers to the self-service option. In 2013–14, the department: re-designed calibration exercises to include self-service; reviewed resource documents; updated training resources; and established quarterly QCL stakeholder meetings to discuss results. In light of the long term and persistent nature of this issue, and its contribution to the department’s strategy to shift customers to self-service, the department advised the ANAO that it will have a continued focus in this area.

Calibration of QCL

3.19 The QCL Step by Step Calibration Guide states that QCL evaluators should participate in national calibration exercises by assessing two mock calls every second settlement period; although the department advised that participation is not mandatory. The purpose of calibration activities is to ensure that evaluators are consistently and accurately applying the framework when assessing calls, which is particularly important given the emphasis on professional judgement in classifying calls as ‘quality’ or ‘non-quality’ calls. The department uses calibration to measure the integrity of QCL responses, identify gaps or improvements to the QCL evaluation process and determine the skills and capabilities of evaluators.⁹⁰

3.20 Table 3.4 shows the participation rate of QCL evaluators in calibration exercises and the percentage of calls that were correctly evaluated for the available settlement periods in 2013–14.⁹¹ For example, in settlement period three, 62 per cent of evaluators participated in calibration exercises, and 91 per cent of their calls were evaluated as correct. Overall, the participation rate in calibration exercises varied between 57 and 68 per cent, a relatively low level of participation. There would be benefit in the department reviewing the reasons for this trend.

Table 3.4: QCL calibration results (2013–14)

Source: Human Services data.

Notes:

a The national calibration exercises were redesigned during 2013–14. Instead of conducting national calibration exercises each settlement period, they occurred on a quarterly basis. Hence, there are no calibration results from settlement periods 7–11.

b Due to the redesign of the national calibration exercises, settlement period 12 does not have an overall correctness figure as it does not follow the previous calibration assessment template.

Quality On Line

3.21 Quality on line (QOL) is a national quality control mechanism used by the department across a number of customer channels including Smart Centres. QOL aims to detect and rectify errors by SOs before any payment is made. It provides an assurance mechanism for payment correctness and recording and verification requirements.⁹²

QOL process

3.22 QOL is guided by the department’s National Quality Online Instructions, which sets out procedures and relevant standards for performing quality checking under QOL, to facilitate consistency and integrity. QOL checking of Smart Centre SOs is conducted by Checkers from Smart Centre sites.⁹³ QOL applies to all employment types—ongoing, non-ongoing and IIEs. The rate of activities sampled for each SO is based on their level of proficiency—learner (100 per cent), intermediate (25 per cent) and proficient (2 per cent) as outlined in Table 3.5.

Table 3.5: Quality online proficiency levels

Source: Human Services data.

3.23 SOs can progress or regress between proficiency levels, depending on the level of correctness attained. The QOL ‘sampling engine’ determines which activities will be sampled for checking, using a weighted algorithm that takes into account the number of transactions since each SO’s last check. Smart Centre staff interviewed by the ANAO confirmed the experience that the specified frequency of QOL checking did occur in practice.

3.24 Under the department’s QOL, sample activities are checked against two quality standards: ‘Getting It Right—Minimum Standards’; and ‘Payment Correctness’. If an activity does not meet either of the quality standards, it is recorded as a critical error and must be returned to the SO to be amended, with feedback from the Checker included.

3.25 Under the National Quality Online Instructions, QOL Checkers are required to complete QOL certification training and receive a 100 per cent result on their skillset test, as well as be re-certified every 12 months. Checkers must also: possess technical knowledge in relevant payment areas; undertake consistent assessment activities in QOL; ensure integrity of payments; provide consistent and constructive feedback to SOs; and maintain a personal QOL correctness rate of 95 per cent or higher.

Monitoring and reporting of QOL results

3.26 Table 3.6 summarises the QOL results for all ongoing, non-ongoing and IIEs from 2009–10 to 2013–14. QOL results show ‘correctness rates’—that is, the percentage of activities processed and released without critical error. The overall correctness target is set as 95 per cent. Overall, the QOL correctness target has been consistently met since 2009–10, with the exception of IIEs in 2013–14.

Table 3.6: Quality online results (2009–10 to 2013–14)

Source: Human Services data.

3.27 The analysis of QOL results is a potentially valuable source of information for the department, and an aid to identifying systemic issues, reviewing training material, and improving current processes. The department advised that since July 2014 it has initiated a weekly review of QOL results. Prior to that time there was no analysis of QOL results at a national level.

Impact of introduction of Skills Pyramid and IIEs

3.28 ANAO interviews with departmental staff identified that the introduction of the Skills Pyramid and growth in the IIE cohort has been challenging for some of the QOL Checkers⁹⁴ and their managers. The challenges have been twofold:

• moving to the Skills Pyramid has required the QOL Checkers (and TPSOs) to be able to perform these roles across the whole general layer of the Skills Pyramid (consistent with the blending of telephony and processing work in the Smart Centre concept) for some processing work. However at the time of the audit, these employees generally have been trained and have work experience in telephony for a specific payment type;⁹⁵ and
• providing technical support to the growing number of IIEs has increased the workload of TPSOs considerably, as IIEs are relatively inexperienced and require more support.

Checking the Checker

3.29 Smart Centres monitor the quality of QOL (that is, checking the checker) by rechecking a sample of QOL checked activities to provide assurance and confirm their accuracy.⁹⁶ Presently, local Smart Centres are responsible for checking the checker, and the frequency of re-checking varies between Smart Centre sites. The department advised the ANAO that Smart Centres are in the process of implementing a consistent check the checker process called Aim for Accuracy, to help overcome the current ad hoc arrangements.

Efficiency and cost measures

3.30 Human Services is exploring a number of strategies to improve the efficiency of its Smart Centres’ Centrelink telephone services, including:

• new telephony technology under the Managed Telecommunications Services contract;
• the employment of IIEs to assist, among other things, with implementing the Skills Pyramid and the blending of processing and telephony work in Smart Centres; and
• a potential partnership arrangement with Telstra for the delivery of less complex, high volume telephone services.⁹⁷

Efficiency of telephony services

3.31 The department regularly monitors efficiency measures such as the Average Handle Time of a telephone call and whether a customer’s issue is resolved in the one telephone call (referred to as ‘First Call Resolution’).

Average Handle Time

3.32 Average Handle Time (AHT) measures the average time a SO engages with a customer on the telephone (including any ‘hold’ time) plus the average time spent on after-call work such as entering data or completing forms. Table 3.7 presents comparative data for the department and data from a contact centre benchmarking report based on self-reported results from 817 contact centres across 79 countries.⁹⁸ The results for a subset of 76 contact centres providing services for the education and government sectors, contained in the benchmarking report, are also presented in Table 3.7 as these contact centres are likely to be more comparable than the results for all industry sectors.

Table 3.7: Average Handle Time (minutes:seconds)

Source: Data provided by Human Services and Dimension Data Global Contact Centre Benchmarking Report 2013–14.

Note:

a Data from the Contact Centre Benchmarking report is for calendar rather than financial years.

3.33 Table 3.7 indicates that the department’s AHT for Centrelink telephone services falls between the average for global call centres and those operating in the education and government sectors. However, while such comparisons can provide useful context, it is difficult to draw accurate conclusions about the relative performance of contact centres operating in different environments. Differences in AHT can indicate differences in efficiency (reflecting differences in, for example, the skill level of employees, the timely availability of technical support and the efficiency and reliability of ICT systems), but they can also reflect differences in the complexity of the calls (that is, simple calls generally take less time than more complex calls).

3.34 The data presented in Table 3.7 also indicates that AHT in the department has been increasing in recent years. Evidence collected from staff in Smart Centres during the audit indicated that this is likely to be, at least in part, due to: the increasing complexity of calls as customers with more simple transactions are increasingly using self-service digital channels; and the additional processing work that staff are completing while engaging with the customer during the call—consistent with the Smart Centre concept of blending telephony and processing to achieve more timely outcomes for customers. Further analysis by the department would help establish the underlying causes for the increase in AHT.

First Contact/Call Resolution

3.35 First Contact Resolution is a situation where an issue can be resolved by the first SO taking the call; providing an efficient and effective service for the customer which avoids the time and cost involved in transferring the call to another SO. A key principle for the operation of Smart Centres is ‘first point of contact resolution for work [is favoured] over handing work between different staff members.’⁹⁹ If a call does need to be transferred, the goal becomes First Call Resolution, that is, a situation where the customer’s issue is resolved by the time the customer ends the call.

3.36 Human Services has no direct metric for First Contact Resolution. However, it does collect data on (but does not generally include in performance reports), the number of calls transferred to another SO. The measurement of Transferred Calls, as a proxy indicator for First Contact Resolution, gives some indication of the proportion of calls dealt with by the original SO taking the call.

3.37 The department also uses a proxy measure for First Call Resolution which is based on repeat callers within a seven day period. The logic is: if a caller rings back within seven days, then it is likely that the reason for their initial call was not fully resolved; conversely, if a caller does not ring back their call purpose was satisfied. However, this proxy measure has significant limitations and does not necessarily indicate that an issue was not resolved. For instance, on some telephone lines, such as the BasicsCard enquiries line, customers may call several times per week to regularly update their knowledge of the balance on their card after making purchases. Table 3.8 presents data on repeat callers and transferred calls.

Table 3.8: Proxy measures of First Call/ Contact Resolution

Source: ANAO analysis of data provided by Human Services.

Note:

a Measures the number of callers from the same telephone number who are repeat callers in a seven day period as a percentage of all calls entering into the call network (excluding blocked calls). Multiple calls from PABX services, Indigenous agents etc. are unable to be excluded and are counted as repeat calls.

3.38 Table 3.8 indicates that the department’s performance in resolving issues during the first contact with customers deteriorated somewhat in 2013–14. The percentage of calls dealt with by the first SO taking the call fell, with the rate of transferred calls increasing from 11.9 to 15.9 per cent over the same period. By way of context, ATO data for 2013–14 on the percentage of calls escalated or transferred was 21 per cent for outsourced centres and 12 per cent for ATO managed centres¹⁰⁰, indicating that Human Services’ performance in relation to the rate of transferred calls is broadly comparable.

3.39 The deterioration in First Contact resolution in the department in 2013–14 occurred notwithstanding the progress made with implementing the Smart Centre concept during the year and some progress in implementing the Skills Pyramid in the latter half of the year. Key goals of both these initiatives are to reduce the need to transfer customers and reduce customer call back. That said, it is still relatively early in the implementation phase of these initiatives, and transitional issues such as the need to train staff may have affected the results. For example, in the first half of 2013–14 IIEs transferred an average of 40 per cent of their answered calls as they did not have the skills and/or delegation to deal with customer issues, many of which had multiple dimensions. In the same period in 2014–15 the transfer rate by IIEs fell to 29 per cent as the experience of this employment cohort increased.

3.40 There would be benefit in the department monitoring the proxy measures for First Call/Contact Resolution, including disaggregated data for IIEs and ongoing employees, to assist in the assessment of the benefits realisation of the Smart Centre concept and Skills Pyramid.

Cost of telephony services

3.41 The department monitors its overall expenditure on Centrelink telephone services, which averaged some $324 million per year from 2010–11 to 2013–14. Table 3.9 shows total expenditure for this period.

3.42 The department advised the ANAO that it does not routinely use metrics such as costs per call. However, Human Services undertook a one-off exercise to construct a cost efficiency metric of ‘cost per minute’ in 2011–12. Table 3.9 presents data on total expenditure, ‘cost per call’ and ‘cost per minute’ from 2010–11 to 2013–14.¹⁰¹

Table 3.9: Cost measures for Centrelink telephone services

Source: ANAO analysis based on departmental data.

Notes:

a The largest component of total expenditure on Centrelink telephony is the staffing costs.

b Cost per call is calculated as follows: total expenditure/number of answered calls.

c Cost per minute is calculated as follows: total expenditure/(number of calls x Average Handle Time in minutes).

3.43 Table 3.9 indicates that costs per call and costs per minute rose significantly in 2012–13 as the number of answered calls fell. In 2013–14, both cost measures reduced. However, a large component of the calculation of cost per call and costs per minute relates to the staffing costs of the SOs answering telephones. Since the introduction of Smart Centres in 2012 a proportion of SOs’ time is spent on processing work, complicating the comparison of cost measures between years.

Conclusion

3.44 Sound quality controls can help provide assurance that Smart Centres deliver consistently high quality Centrelink telephone services. The department has two quality assurance frameworks in place to measure the accuracy and quality of its Centrelink telephony services: QCL and QOL, both of which are underpinned by ‘check the checkers’ processes. While the department has a well-established QCL framework for measuring and monitoring the quality of SOs’ customer interactions, there are a number of gaps in its implementation.

3.45 The QCL framework covers all staff answering telephones—ongoing, non-ongoing and IIEs. However, IIEs—as a group of employees with a generally higher risk of making errors given their relatively lower level of experience—are not currently included as part of the QCL framework in Smart Centres; although some are evaluated under ad-hoc arrangements. The non-application of QCL to IIEs risks weakening the effectiveness of QCL as a quality assurance mechanism and QCL should be applied to all relevant staff.

3.46 In addition, the requirement in the QCL framework that four calls per experienced SO and eight calls per new SO are quality assured is not being achieved. Participation rates in calibration exercises for QCL evaluators are also relatively low. There would be benefit in the department reviewing the reasons for these gaps in implementation.

3.47 The additional quality assurance mechanism of QOL is being completed by Smart Centres and all employment types are evaluated. Smart Centres have high QOL results and have generally met the target correctness rate of 95 per cent.

3.48 A number of measures have been used at various times to help the department assess the efficiency of Centrelink telephone services, including Average Handle Time and First Call/Contact Resolution. In recent years, the department’s Average Handle Time for Centrelink telephone services and First Contact Resolution have been broadly within the range experienced by other call centres. However, the department’s performance against both Average Handle Time and a proxy measure of First Contact Resolution declined in 2013–14 relative to the previous year. While issues associated with the implementation of Smart Centres appear to have contributed to these declines in performance, continued analysis by the department of its performance against these key efficiency measures would help confirm the factors behind these trends. Analysis would also assist in evaluating delivery of the expected benefits from Smart Centres, new technology capabilities, and structural changes to the department’s telephony workforce.

Introduction

4.1 Performance measurement, monitoring and reporting encourages ongoing advancements in service delivery and effectiveness, by highlighting improvement and innovation.¹⁰² External performance reporting informs the Parliament and the public of a program’s performance, enhancing program transparency and accountability.¹⁰³

4.2 To assess how Human Services measures, monitors and reports on Smart Centres’ Centrelink telephone services, the ANAO reviewed:

• the range of performance metrics used by the department and their alignment with call centre industry standards and the Smart Centre concept;
• internal systems for performance monitoring and reporting; and
• external performance monitoring and reporting in Human Services’ Portfolio Budget Statements and Annual Reports.

Performance metrics

4.3 Human Services uses a range of performance metrics for Smart Centres’ Centrelink telephone services. Table 4.1 sets out the department’s most commonly reported metrics including targets where relevant.¹⁰⁴ Other metrics used by the department are set out in Appendix 5. Table 4.1 also reports on the consistency of these metrics with contact centre industry definitions.¹⁰⁵

Table 4.1: Telephony performance metrics and targets

Source: ANAO analysis of Human Services’ data.

4.4 Table 4.1 indicates that the department’s commonly reported performance metrics are consistent with industry definitions. The department has set internal targets for a number of these metrics.

Consistency of traditional performance metrics with Smart Centre concepts

4.5 ANAO analysis indicated that the current metrics used by the department are largely traditional call centre metrics.¹⁰⁶ Such metrics generally have a quantitative, process-driven focus and, while useful, provide a limited basis to assess customer outcomes and the success or otherwise of the Smart Centre concept.

4.6 As discussed in Chapter 2, a key component of the Smart Centre concept is to blend telephony and processing work and achieve better, more timely outcomes for customers including improved First Call/Contact Resolution.¹⁰⁷ As outlined in Chapter 3, the department does not directly measure First Call/Contact Resolution, although it can calculate some proxy measures. There would be benefit in Human Services reviewing its metrics and reporting on First Call Resolution outcomes given the importance of this process in achieving improved customer outcomes and efficiencies for the department.

4.7 Another key component of the Smart Centre operating model is to assist in shifting customers to self-service channels, including digital channels where possible and appropriate. The department’s draft Digital Strategy¹⁰⁸ states that Smart Centres should ‘ensure key performance indicators for average call handling times reflect the requirement for staff to stream customers to the digital channel and assist them with their transactions online if required’. Some SOs interviewed by the ANAO commented that they were still managed to a target for Average Handle Time that had not been reviewed, even though they were performing more processing while the customer was on the line. They were also taking time during the call to encourage and assist customers utilise self-service. These staff were of the view that the target should be reviewed so that it better reflected Smart Centre objectives.¹⁰⁹

4.8 The Interactive Voice Response (IVR) system is an important component of Centrelink telephone services, with all calls able to enter the network directed through the IVR system. Around 40 per cent of calls entering the network ended in the IVR system in 2013–14 (see Figure 2.1 in Chapter 2). A global benchmarking report¹¹⁰ of call centres in 2013–14 commented that despite IVR being the most longstanding and widely used self-service channel, most call centres do not monitor and report on its performance. The report states that maximising the benefits of IVR will only occur if the performance of self-service IVR is carefully monitored and made the subject of ongoing iterative improvements.¹¹¹ The department has limited data on the performance of the IVR system other than the number of self-service transactions, and there would be benefit in monitoring how many callers receive the information they sought from the pre-recorded messages, as a basis for assessing the IVR system’s performance.¹¹²

4.9 The department advised it has recently commenced a review of Smart Centres’ performance measurement and reporting. The review will consider changes to reflect the objectives of Smart Centres and the implementation of the Skills Pyramid. It is anticipated, for example, that separate data may be available for the ‘general’, ‘assisted’ and ‘intensive’ layers of the Skills Pyramid. Over time, these new data sets would allow the department to identify appropriate performance expectations for each layer.

Internal performance monitoring and reporting

4.10 A structured and regular system of performance monitoring and reporting allows for the timely consideration of performance information, positioning the department to respond and make appropriate adjustments to service delivery arrangements.¹¹³

Performance reporting arrangements

4.11 Within Human Services, the Service Delivery Performance and Analysis Branch is responsible for the internal reporting of performance outcomes for Smart Centre’s Centrelink telephone services. The branch produces over 30 performance reports that include metrics for Smart Centre’s Centrelink telephony channel performance. Detailed performance reports have been developed for relevant departmental governance committees. The department also produces operational reports which are specific to a particular queue or payment type; for instance, the Indigenous Agents Silver Service Report. A number of reports are produced with individual site level results available.

4.12 These reports generally provide a structured and regular system of performance monitoring and reporting and are appropriately targeted to each report’s audience.

Responses to performance reports

4.13 At the operational level, the performance of Smart Centre’s Centrelink telephone services is overseen and managed by the Smart Centres Channel Operations Branch. Smart Centre managers and leadership teams are responsible for ensuring their centre’s performance against the metrics and associated targets.

4.14 The ANAO’s review indicated that Smart Centre managers responded to the performance information in the performance reports, and had devised strategies tailored to their own locations and staff to improve performance. For example, local strategies were developed to help reduce unscheduled absences and to manage variation between SOs in Average Handle Time.

4.15 Consistency between the Smart Centres is promoted through a Performance Managers’ meeting, which meets every six weeks to share good practice and discuss issues across the network.

External performance reporting

4.16 Ongoing performance monitoring and external reporting provide visibility of public sector activities so that the Parliament and other stakeholders can assess performance against planned outcomes.¹¹⁴

4.17 Centrelink’s telephone services are a core part of Human Services’ business and are of ongoing interest to both Parliament and the community. The effectiveness of Centrelink’s telephone services has been discussed in Senate Estimates hearings with Senators regularly requesting data on the performance of Centrelink’s telephone services. A community stakeholder group interviewed by the ANAO also commented on the desirability of having access to more performance information on telephone services.

Current Key Performance Indicators arrangements

4.18 Since 2012–13, Human Services has used an average speed of answer of less than or equal to 16 minutes as its single KPI for external reporting purposes for telephony services.¹¹⁵ In both 2012–13 and 2013–14, Human Services has reported against this KPI using data averaged across all its programs (Centrelink, Medicare and Child Support). In 2013–14, Human Services reported achieving its KPI, with an average speed of answer across its programs of 14 minutes and 26 seconds. However when the 2013–14 data is disaggregated by program, the average speed of answer for Centrelink telephone services was 16 minutes and 53 seconds. Human Services advised the ANAO that from 2014–15 it will report Centrelink, Medicare and Child Support telephony performance separately in its Annual Report.¹¹⁶Table 2.4 sets out the separate KPIs for average speed of answer for Centrelink, Medicare and Child Support Program. This development will provide improved transparency of the performance of Centrelink’s telephony services.

4.19 The ANAO developed criteria to evaluate the appropriateness of key performance indicators used by entities to report to Parliament as part of a pilot project conducted between 2011–12 and 2012–13.¹¹⁷ The criteria identify that KPIs should be relevant, reliable and complete, and provide a balanced examination of the overall program performance, both quantitatively and qualitatively.

4.20 The KPI of an average speed of answer of less than or equal to 16 minutes is relevant as it addresses a significant aspect of telephone services and is largely understandable to the Parliament and other stakeholders.¹¹⁸ The KPI is also reliable, as it is can be quantified and the results can be used to show trends over time. However, as a single measure of a service user’s telephone experience, the KPI is far from complete, as it does not provide an overall picture of Centrelink’s telephony performance. For instance the average speed of answer KPI provides limited information on the variability in wait times experienced by customers. As Tables 2.2 and 2.3 indicate, while the average speed of answer in 2013–14 was 16 minutes and 53 seconds, 36 per cent of callers waited less than ten minutes, while 30 per cent of callers waited more than 30 minutes for their call to be answered.¹¹⁹ Anecdotal reports of customers waiting much longer than average wait times have also been reported by the Commonwealth Ombudsman. As an alternative, some organisations, such as the ATO, use the percentage of calls answered within a specified time frame as a measure of performance.¹²⁰ The ATO indicator provides a clearer picture of the proportion of customers that experienced the desired service delivery target and thus the service standard that most customers can expect.

4.21 In addition, the average speed of answer only measures one aspect of the program’s overall performance. As indicated in Figure 2.1, average speed of answer relates only to calls that are answered by a SO—around 45 per cent of calls able to enter the telephone network. To gain a more complete picture of the performance of Smart Centres’ Centrelink telephone services, it is desirable for Parliament and other stakeholders to have information about the broader customer experience. For instance, the ANAO estimated that in 2013–14 around 30 per cent of calls that entered the network were abandoned and 13.7 million calls were blocked.

4.22 A larger range of KPIs relating to Centrelink telephony services were previously reported by the then separate Centrelink in its Annual Reports. In 2006–07, $115 million was allocated to Centrelink telephone services over two years to provide additional capability to ensure a ‘reasonable’ service level in response to growing customer demand. A ‘reasonable level’ of service was measured by Centrelink as achievement against the following KPIs:

• unmet demand¹²¹ less than 9 per cent in peak demand periods;
• unmet demand less than 8 per cent in non-peak demand periods;
• 80 per cent of callers would be successful on their first attempt;
• 70 per cent of calls would be answered within 2 minutes and 30 seconds; and
• less than 22 000 customer complaints concerning poor access to Call Centres.¹²²

4.23 Table 4.2 presents the then separate Centrelink’s target and actual performance against the KPIs from 2006–07 to 2010–11.

Table 4.2: Centrelink call key performance indicators targets and actuals (2006–07 to 2010–11)

Source: ANAO analysis of Centrelink Annual Reports from 2006–07 to 2010–11.

Notes:

a Unmet demand measures the proportion of calls that had been unable to enter the network over a consecutive seven day period (Monday to Sunday). For a more detailed explanation see footnote 51.

b For 2008–09, the Key Performance Indicator for unmet demand had targets for 13 settlement periods – 9 per cent in peak periods and 8 per cent in non- peak periods. Centrelink reported how many targets were achieved for these periods.

4.24 While some of the performance targets in Table 4.2 changed over time, the KPIs provided an indication of the level of service considered acceptable for customers. Overall, the KPIs provided a more complete understanding of the customers’ experience of Centrelink telephone services including by providing a target for the proportion of customers answered within a specified timeframe and therefore a measure of the service standard that most customers can generally expect.¹²³

4.25 The department has moved over time to a single target KPI which, as discussed above, does not provide insight into the range of customer telephony experiences. There would be merit in reviewing the KPIs for Centrelink telephone services to clarify the service standards customers can reasonably expect to receive when they ring a Centrelink telephone number, having regard to other relevant benchmarks.

Recommendation No.3

4.26 To clarify the service standards that customers can expect and to better reflect customer experience, the ANAO recommends that the Department of Human Services review Key Performance Indicators for the Centrelink telephony channel, in the context of the implementation of a coordinated channel strategy.

Department of Human Services’ response:

4.27 The department agrees to the recommendation with qualifications.

4.28 The department believes that the current KPIs provide a sound and comprehensive set of metrics for Centrelink telephony and it closely monitors service levels including average speed-of-answer. The KPI for average speed-of-answer of 16 minutes was met in 2013–14. To reduce average speed-of-answer there would need to be either an increase in resources or an increase in call blocking which would prevent access into the department’s telephony systems. The department has estimated that to reduce the KPI to an average speed-of-answer of 5 minutes, it would need an additional 1000 staff at a cost of over $100 million, each and every year. The department provides callers with information on call wait times and as much access to telephone self-service transactions and information as possible, including how to access digital services. By directing customers to self-service, resources are prioritised to calls from customers that need more assistance. That said, a review of KPIs may need to be considered in the context of longer term changes underway with the recently announced Welfare Payments Infrastructure Transformation (WPIT) Programme.

Conclusion

4.29 The department internally monitors the performance of Smart Centres against a range of useful, albeit traditional call centre metrics. A number of internal targets have also been established. While these metrics are an aid to assessing performance information at an operational level, they provide a more limited basis for assessing customer outcomes and the success or otherwise of the Smart Centre concept. There would be value in examining existing metrics and their fitness for purpose in the Smart Centre environment. In particular, the department could usefully focus on:

• First Call/Contact resolution—as improving resolution rates is a key goal of the Smart Centre concept;
• the IVR system—to assess its effectiveness in resolving calls without having to speak to a SO; and
• the interpretation of Average Handle Time—which has changed following the blending of telephony and processing work.

4.30 Human Services currently uses the average speed of answer of less than or equal to 16 minutes as the single KPI for its public reporting on telephony services. While, the ANAO considers this target KPI to be relevant and reliable, it is not complete and merits review, having regard to other relevant benchmarks. In particular, the KPI does not provide insight into the range of customer telephony experiences, including the length of wait time that most customers can expect to experience and the common customer experience of receiving a busy signal, that is, call blocking or rates of call abandonment.

Appendices

Please refer to the attached PDF for the Appendices:

• Appendix 1: Entity Response
• Appendix 2: Department of Human Services’ Response to Ombudsman’s Recommendations on Centrelink Telephone Services

Abbreviations