Scottish Government

BENCHMARKING AND DEFINITIONS WITHIN CONTACT CENTRES

SECTION 2: QUANTITATIVE BENCHMARKING

2.1 INTRODUCTION

Call centre benchmarking can be defined as the measurement and comparison of practices and standards, within an organisation and/or with other operations externally, which can enable the adoption of effective management practice and drive decision making and improvements.

The most comprehensive form of benchmarking is one that involves all aspects of a call centre: operations, standards, physical characteristics, IT and telephony infrastructure, HR, recruitment and training practices and financial information. However, smaller studies can also provide a very valuable insight into key metrics and how your centre measures up to others.

Many of the areas are included in the CCA Standard Framework (Section 5 Service Performance and Organisational Efficiency), e.g:

5.1 Standards have been set for key activities related to the centre and are measured

5.2 Standards are understood by employees and plans are in place to achieve/maintain the standards agreed.

Most call centre environments can be rigorously measured and controlled, and so lend themselves well to quantitative benchmarking. IT and telephony systems such as the ACD (Automatic Call Distributor) are able to easily extract required data.

Anyone involved in either the design, direction or management of call centres can gain invaluable information from this activity.

2.2 THE BENEFITS OF BENCHMARKING

Considerable benefits can be gained from benchmarking a call centre. Selecting areas to focus on and gathering required data can be a valuable management exercise in its own right. Indeed, it may highlight areas which have not been previously considered on a regular basis.

Comparisons can then be made with other centres, which is a valuable way to evaluate the operation, justify improvements and bring about changes. Actions taken as a result of the information can then result in improvements to productivity, customer satisfaction and cost and revenue.

There are 4 steps to the benchmarking process; below is shown the benefits at each stage:

2.3 TYPES OF BENCHMARKING

The most accessible form of benchmarking is within the centre or organisation. There are considerable benefits to be gained from comparing different call centres within the same organisation, which are often managed quite differently and could learn from each other. The other main strength of this is that it is the most relevant comparison in terms of the work undertaken, the processes used and the culture. Confidentiality should not be an issue, and it may be possible to get to know the other centre on an in depth, one-to-one basis.

Whilst internal benchmarking is recommended, it is equally important to benchmark against the industry as a whole. Business and consumer expectations are influenced over time by their experience of a whole range of products, services and organisations.

There are a number of industry bodies, research organisation and call centres companies who offer organisations the chance to participate in benchmarking surveys, of varying complexity, size and cost.

2.4 SETTING BENCHMARKS

Within a centre, quantitative benchmarks should be selected using the SMART principle:

S Specific - clear and precise, easy to understand and communicate to others

M Measurable - tangible, measurable and easy to monitor

A Agreed - accepted as the desired focus by all parties in the call centre

R Relevant - relevant, given the activity of the operation

T Time bound - specific in terms of the timescales for measurement

Where benchmarks are to be set for comparing individuals, it is advisable to make them consistent with the average of the best current performers.

2.5 KEY QUANTATIVE MEASURES

There are hundreds of measures which could be used for benchmarking in a call centre. When making comparisons, a range of measures should be selected:

2.6 HUMAN RESOURCE BENCHMARKS

Absenteeism

This is the average number of agent days lost through sickness and unauthorised absence as a percentage of contracted days annually, e.g. 15 ÷260 + 5.8%, and should include all forms of unauthorised absence including long term sick. Ideally, if systems allow, this should be measured on a rolling month basis.

This is a key benchmark as it is an indicator of underlying motivational and stress problems, and affects productivity and potential revenue. It is important to measure agent absenteeism separately from that of the managers and team leaders as there is usually quite a difference.

A best practice figure to use for comparison purposes is 5-6%.

Attrition

This is the staff turnover annually, expressed as a percentage. It is calculated by taking the number of people that have left the call centre for whatever reason as a percentage of the total number of staff. It is important to use the actual number of heads, rather than the full time equivalent (FTE) as the impact of attrition is the cost of re-recruitment and re-training of replacements, and this is on a per head basis.

This benchmark can show just those who leave the organisation (external attrition) or all leavers which would include those who move on to other roles either within the call centre or the organisation.

There are many factors affecting attrition; as well as internal factors such as the salary offered, type of contract, job satisfaction and career progression there are external factors such as the make up of the local labour market and whether there are other competing call centres nearby.

Attrition levels differ by industry and geographical location. A best practice figure would be between 15 and 25%. It is healthy to have a certain level of turnover, and therefore possible to have a level which is "too low".

2.7 TECHNOLOGY BENCHMARKS

Core system % downtime

All call centres are reliant on technology, and a key technology benchmark is the amount of systems downtime, calculated as the percentage of time that the call centre is operational, and includes telephone as well as computer systems.

This indicator can highlight unreliability due to lack of investment in technology, or times at which there is non availability due to development or testing work being undertaken.

As a benchmark, systems should be available the whole time that the centre is operational including weekends, and the downtime percentage should be no more than 2%. The appropriateness of the benchmark should be viewed in light of the kind of work being undertaken, for example emergency services would require certain systems to be available 100% of the time.

Systems support response times

Another technical benchmark is the speed at which technical support respond to individually reported problems in the call centre, and this would be measured in hours. It may also be appropriate to categorise the different types of request, as there will be differences in expected time of resolution.

% contacts handled by automated services

With the advent of automated technology, where customers can resolve queries through "self service", either through entering information on the touch pad of the telephone or speaking to a voice recognition system, this benchmark is a useful indicator of the take up of these facilities. This information is available through the telephone systems.

When comparing this information, it is important to compare with organisations that do calls of a similar type of complexity. Examples where full automation is used are: utilities for collecting meter readings, financial services for giving basic account information and card activation, travel and event companies for giving standard information about timings and locations.

2.8 TELEPHONE BENCHMARKS

Service level/Grade of Service

The service level benchmark is one most commonly used to demonstrate the service which a call centre provides to its customers. It is the percentage of calls received by the centre that are answered by a human agent within a certain time frame, usually 10 or 20 seconds.

This information is available from the ACD, and the configuration of the phone system should be taken into account to ensure that the measurement is consistent, as there are different stages which could be included:

Average speed to answer (by human agent)

The speed to answer is related to the service level, but it also takes account of all the calls that are not answered within x seconds, producing the average overall.

Calls abandoned

Calls abandoned is the number of calls that are abandoned, and includes those which abandon before the IVR/VPS and those which abandon whilst queuing for a human agent. This is then expressed as a percentage of all the calls received.

The propensity for customers to abandon will vary considerably depending on the expectations that they have of the service being provided.

A guideline benchmark is between 3-8%, and best practice would be 3-5%. In many situations a consistent abandoned rate of less than 2% is likely to indicate low staff utilisation.

Number of calls answered per hour

A key measure of productivity is the number of calls taken per hour. This is affected by the skill and motivation of the agents, the speed of the systems, the type of calls, the amount of after call work and follow up calls, and how closely matched the resourcing levels are to the volume of calls coming into the centre.

This measure usually just includes the number of inbound calls answered. If the agents are doing outbound or follow up calls as well, this will need to be taken into consideration.

Average time to abandon

The average time to abandon is the length of time in seconds that a customer waits in the queue (from the time that the call is answered by the system) until they abandon. This is obtained from the ACD.

An average benchmark is 40-60 seconds. Customers tend not to queue for an agent to answer their call for more than a minute, except in exceptional circumstances, and customers who wait two minutes or longer are either in real need of assistance or have come to expect this level of service from an organisation. Much depends on whether the caller has any alternative - whether they can go elsewhere to satisfy the query.

2.9 EFFICIENCY BENCHMARKS

Accuracy of call forecasting

Within a call centre, the accuracy of forecasting of volume, type and length of call is vital to enable the right level and mix of resources.

This information can be obtained from the call forecasting package such as Blue Pumpkin or CMS, or calculated manually.

If the actual calls are more than predicted, there will not be enough people to answer them and there will be pressure on the service level. If the calls are less, the agents will be under-utilised, having a negative impact on costs.

The target to aim for should be 100%, and +/- 5% is acceptable.

Adherence to roster

Call and workload forecasts are developed into rosters for all members of the call centre. A key efficiency benchmark is the adherence to these, i.e. are agents on time for their shift, do they go for lunch and breaks at the specified time, are they able to finish on time. This is particularly important in smaller call centres as there is not as much room for manoeuvre in terms of answering calls. It is also important to ensure that the agents are working their contracted hours.

This information is obtained from a workforce management package, or from manual timesheets.

Utilisation

The main measure of productivity in call centres is utilisation, i.e. what are the agents doing in their in work time.

The main categories are:

• Talk time - the time spent speaking to customers on the telephone.
  This usually only includes inbound calls in a service centre.
• Wrap Up - After Call work. This is the time that an agent takes after the call has finished to complete the case and may include updating the system, completing forms etc associated with the call.
• Available - this is the amount of time that an agent is waiting for a call. This is often referred to as "idle" time but in practice agents often spend the time catching up on reading, emails etc whilst they are still logged in waiting for a call.
• Other - this can include team meetings, coaching sessions, admin, breaks and any other activity that is not involving call taking.

Agents should log out when it is their lunch break, as this is not paid for time.

This information is available from the ACD, but care should be taken as different systems will be configured in different ways, and have varying definitions. The amount of time in each activity will vary depending on the size of centre, service levels, systems used, type of activity, and amount of training and coaching undertaken. As a general rule, talk plus wrap should be no less than 65% and available should be less than 10% (or less in a very large centre).

% occupancy of workstations

This benchmark shows how well used the workstations are, and will change at different times of the day, e.g. at peak times, during the evening shift, etc depending on the agent rotas and the extent to which the centre uses hot desking.

It is an important measure to highlight under-utilisation of capital expenditure in workstations and computer hardware and premises space, or at the other end of the spectrum, over-utilisation resulting in inadequate desk-space which affects productivity.

2.10 SERVICE BENCHMARKS

Query resolution

Query resolution can be measured in two ways:

• The amount of time it takes to fully resolve a query, including any call backs, research or internal calls
• The percentage of all calls that are resolved on the first attempt, without the agent needing to refer the customer on to a colleague, their manager, or calling the customer back. On average this benchmark would be in the region of 70%, but this depends on the complexity of the problem and the way in which the centre is organised, for example if the agents are multi-skilled or if they have a system which provides the required information most of the time.

Ratio of inbound to outbound/onward calls

The extent to which outbound or onward calls are required is a measure of first time resolution in a service call centre.

This information should be available from a call logger or from the ACD. However, this often does not include internal calls within the call centre, so the number could be underestimated.

Turnaround time

As well as call handling, many call centres process other types of media such as post, emails and faxes.

It may be possible to get this information from a document logging/ workflow system or an email handling system. If this is not possible, a representative sample should be measured at regular intervals.

2.11 COST BENCHMARKS

Cost per contact or call minute

The cost per contact or call minute is a useful measure to compare the cost effectiveness of the call centre channel compared with other means of communicating with customers, for example face-to-face through a field force or walk in centre.

The cost per call should take account of all costs but exclude capital investment:

• Agent costs (salary, all on-costs, overtime etc)
• Management, supervisory and support staff costs
• Telephone bills
• Technology, systems and telecom costs (licences, support and running costs)
• Premises and facilities
• Contribution to overheads
• Other operational costs.

Economies of scale are usually much greater for larger call centres, so this should be taken into account when comparing cost data.

This data is usually quite a challenge to obtain in its entirety as in practice a lot of budgets are held outside of the call centre.

Cost to serve

The cost to serve is another way of comparing the worth of the call centre compared with other channels.

It is calculated by taking the total call centre costs and dividing it by the number of customers. Two versions are possible - one for all customers and the other for "live" (i.e. people that have made some form of contact with the organisation in the last period of time) customers only.

Cost of complaints and compensation

In certain organisations, the cost of complaints and compensation can be a significant expense. This is an important benchmark as it is a direct cost indicator of when things go wrong, and can often be used to put together a business case to make improvements.

There should be a mechanism in place to monitor the amount of compensation awarded, and who authorised it.

2.12 SALES BENCHMARKS

The benchmarks used in call centres which are focussed on sales will be quite different from those concentrating on service. Measures such as contact rates (the number of customers that the agent has made contact with to make a sales presentation), conversion rates (initial enquiry to lead, outbound call to lead, lead to sale, lead to appointment/visit, etc), value per order and the extent to which cross- and up-sales have been achieved become important.

The benchmarks become more focussed towards results and less about quantity. This in turn, requires a mind-shift within the centre.

The difference between sales and service benchmarks

As we see a further convergence of sales and marketing in call centres with integrated functions and products, in the future call centre benchmarking will encompass measures linking business outcomes with the call operations, such as:

• Customer profitability
• Customer loyalty and retention
• Customer share (of potential or product holding)
• Customer lifetime value.

With technology becoming more prevalent, e.g. CRM systems, it is only a matter of time before comprehensive benchmarking can be undertaken in these areas.

2.13 AREAS TO CONSIDER

Inevitably there are some areas which you need to be aware of and take into consideration when taking part in benchmarking exercises. On the whole, use of common sense is invaluable!

• Survey size - ensure that the survey is sizeable enough to produce a reliable benchmark
• Confidentiality - where data is being given to third parties or external research organisations, be aware of any specific information that will be released into the public domain
• Relevance - there can be marked variations in values for call centres of different sizes, sectors and locations, so it is important to consider these factors and note variations to come up with the most appropriate comparison for your centre
• Definitions - the call centre industry is full of jargon and definitions which vary (even within the same organisation!). Make sure that definitions are clearly explained and understood, and if in any doubt then ask the organiser of the benchmarking
• Accuracy - even with all participants making every effort to provide data as accurate as possible, there will invariably be some inaccuracies and inconsistencies (for example due to shortcomings in management information systems) which should be read with caution when analysing such information
• Focus - different studies will have a different focus, e.g. technology, people, industry specific, and it is important to keep this in mind
• Interpretation - benchmarks should be carefully interpreted. You should know exactly how the information was made up and how it relates to the specific workload and situation in your call centre.

Quantitative benchmarking is a useful starting point to assess call centre standards, and can be used to determine how a call centre is performing at the present time, as well as looking towards aspirations for the future. To gain a fuller picture, it is recommended that some form of qualitative benchmarking is also considered, and this is covered in the next chapter.