A Short Guide To Service Measurement

By Graham Williams, Centre-ing Services

At a conference on customer satisfaction measurement in Atlanta several years ago, I was struck by the way delegates fell into two categories:

Both were right in their own way. Certainly information is food to the organization, a prerequisite for understanding and improvement—if properly interpreted and used. But not if proliferated and wrongly applied.

Service measurement is an area of much confusion. There is a scramble after the latest best-practices and benchmarks. There is disillusionment and disenchantment when measurement policies don’t work. It’s an area deserving of attention. Is there a sound "middle route" we can begin to develop in our organizations—a route that balances the mathematics and the psychology to best advantage?

How Measurement Goes Wrong

Consider an example drawn from the petroleum industry.

A service station operator experiences a petrol pump breakdown. The basic measurement standard used by the petroleum company’s engineering services unit is 24 hours for a normal repair, 8 hours for an urgent repair, 2 hours for emergencies.

SERVICE AGENT A: Dispassionately logs and references the call, makes a ‘seriousness judgement’ of ‘normal’, passes the incident onto the maintenance people. This agent wraps the incident up quickly, is now free to take the next call, and scores the maximum on efficient performance according to the measurement system.

SERVICE AGENT B: Takes the call, understands and responds to the customer’s anxiety-state, correctly recognises the incident as being urgent, logs and references, ‘goes the extra mile’ for the customer, telephones an engineering focal point authority, gets permission to make direct contact with a standby pump technician, actions the repair, and follows-through. In this case, ‘wrap up’ takes longer, there is a delay in taking the next call, more operating cost money is spent as a result of the higher seriousness rating, the service agent does not score the maximum, and is effectively penalised in terms of performance measurement.

In the case of A the short term result is seen to be "better," even if the service agent makes a wrong judgement-call and the customer is unhappy. This is because the measure is about cost; and not also about relationship, lost sales during downtime, and takes no account of the relative importance of the customer to the company. Service agents, in their own interests over time, will invariably follow route A. However, route B is clearly in the organization’s best interests, certainly in the longer term.

Situation: Pump breakdown. Anxious customer. Task measure is 24/8/2 hours based on cost implications.





24hr "normal" judgement

8hr "urgent" judgement


Dispassionate, cold

Deals with customer feelings


Quick "wrap-up"

Makes special arrangements

Outcome: Company

Cost efficient, lost sales

Higher cost, saved sales

Outcome: Customer



Outcome: Agent


Penalised (time, # calls)

Possible solution:

  • Measure "Process efficiency cost and speed" + Quality of relationship + Relative value of customer to the company
  • Ensure optimum outcome occurs consistently


Common Mistakes

What can we learn from this simple example which occurs frequently and in different forms for different processes in different industries? Few would argue against the notion that service measurement is an indispensable means to implement service vision and strategy by:

So why so many mistakes?

Many organizations get their measures and applications wrong in four areas:

  1. The measurement is used for the wrong reason. It is used for "command and control" purposes. Resource International reports that in over 70 studies, not one of the organizations studied was able to get their "galley slave" (that is, productivity-loaded and cost-savings) models to work. This is a bit like the example of the Russian shoe factory where productivity was measured by the weight of the shoes produced—but the shoes were not saleable! Or the case of the security guard/vehicle operator whose basic performance measure was the number of miles patrolled per shift. Guess what he did. Or, service agents measured and remunerated on the basis of the number of calls taken.
  2. The measures are not appropriate to the organization’s culture. Mindlessly adopting benchmarks and "best-practices" without due regard to their applicability to the organization and customer service center, especially the internal organizational culture, doesn’t make a lot of sense. Organizational culture is an extremely powerful force. A strong, dominant cultural pattern within an organization expresses implicit values. It conditions and determines how people behave. So for example, in a "command and control" culture, service agents are very likely just to work to rule and constrain the organization from reaching it’s goals. Adopting best practice measures from a variety of sources without consideration of cultural factors can render the entire exercise quite useless. Instilling a company-wide service oriented culture requires a lot more than introducing new measures.
  3. The objects of the measures, employees, do not properly relate to the measures. The inclusion of measures over which staff members have no real control, is self-defeating. In the petroleum industry, for example, it’s silly to measure service center agents on the quality of a customer’s physical fuel delivery experience. The delivery vehicle driver may spill product, not have the right couplings, behave badly—and although this represents poor customer service, the customer service center agent cannot be held responsible, other than for receiving and handling a complaint and correctly reporting the incident.
  4. Applying unique measures to the customer service center in isolation from the rest of the organization. The customer service center is not a "tack on," but an integral part of the wider organization. Measures applied in the customer service center should foster service co-operation between the center and other departments, should be congruent with appraisal and reward mechanisms applied elsewhere so that employees perceive fairness. Service Center measures should be designed in a way that enables the service staff to act as catalysts to service level improvements throughout the organization.

The results of these bad measures and practices are:

How do we avoid these common errors and get it right?

Measure for the right reasons. First, it’s about attitude. If we earnestly believe that customer satisfaction drives business results, and that customer satisfaction is driven by process-efficiency, supported by the quality and effort of our people: then a properly constructed measurement framework will work for us at the behaviour-level. We will be far more inclined to control operating efficiency, achieve lower unit costs per service provider, reduce the number of rings before phone answered, improve response times, reduce customer queue lengths/waiting time—by instilling passion and commitment in carefully-selected, well-coached agents; as an alternative to "galley-slave management." This way, we avoid "working to rule," undue stress, constant tinkering with wrong measures in a futile attempt to make them work.

Part of getting it right is an acceptance that customer service is as much about relationship building as it is about task performance. A good balance between task (productivity) and relationship (quality of moments of truth) is absolutely essential.

Consider a trip on an aircraft flying from Bangkok to Kuala Lumpur: the flight attendants are courteous, helpful, competent, respond quickly to your in-flight needs. However, you arrive late and upon disembarking discover that you’ve arrived in Singapore and your baggage has been lost: The relationship aspects were good and the task aspects were bad. This equals poor service.

Alternatively, everything works smoothly. You depart and land on time and in the right place. However, during the flight, the flight attendants failed to acknowledge you, did not respond when you rang the buzzer, carelessly spilt soup on your clothing and mopped up without a real apology or proper recovery action. In this instance – the travel task was carried out but the relationship aspect was poor. This also equals poor customer service. Task and relationship must support each other in order to provide good customer service, and both task and relationship elements must be properly included in your measures.

Ensure a measurement and culture "fit." It holds true that visionary, forward looking service providers work back from where they want to be. Thus, using the Treacy/Wiersema model (from The Discipline of Market Leaders by Michael Treacy and Fred Wiersema, Harper Collins, 1995) as an example:

And an organization aiming at service excellence will determinedly tackle cultural constraints along with the introduction of new service measures. This may well entail:

Facilitate employee ownership of the measures. If you want employees to accept performance measures, then allow them to participate fully and as equal partners in the selection, design and application of service measures. If you want them to make sound decisions during service "moments of truth," then provide decision-making assistance. A good rating to include in the information system is a "relative value of this customer." This provides service agents with a fair means to prioritise their responses. Take care when using across-department measures that people are appraised only on what they can actually control or strongly influence. This requires a clear understanding of which measures are fully shared, partially shared or not shared.

Apply performance measures congruently across the organization so that the service center is not isolated. There is no point in the call center achieving excellence in isolation if at the same time the overall organizational levels, especially in customer satisfaction, do not improve significantly. There is also no point in the call center taking responsibility for areas where it has no real authority, nor has accountability. The achievement of fully integrating the service center with the remainder of the organization entails a lot more than just measurement. An excellent start-point however is to introduce the eloquent logic of a balanced scorecard design (*1). This can be used very effectively to:

Incidentally, the cause and effect connections found in the service profit chain (*2) dovetail very closely with the balanced scorecard.


The messages are clear. A right attitude to measures leads to right behaviours. A solid organization-wide service orientation is essential to a measurement system that works. Service measures that assist in the realisation of vision and strategy should:

And remember: the best service providers involve customers in the design of their service measures.

*1 The Balanced Scorecard: translating strategy into action, by Robert S Kaplan & David P Norton (Harvard Business School Press, 1996).

*2 The Service Profit Chain, by J L Heskett, W E Sasser & L A Schlesinger (The Free Press, N.Y., 1997).

Graham Williams is a management consultant and founder of Centre-ing Services, an international consultancy focused on dissolving customer stress points in organizations. He is based in Cape Town, South Africa. He can be reached at http://www.centre-ingservices.com.


Back to the Top of The Page, The Overview Page or The Managers Files or Home Page.