Value Demand in Call Centers

Value demand' implies understanding call types in customers' terms. Typically, this leads to a quite different classification than that which call centers have established for themselves.

For example, a call center for a multi-national computer maintenance company was established to deal with customers' calls. An IVR (Interactive Voice Response) system was set up for customers to route themselves to specialist groups.

From the customers' point of view their first interaction with the organization was a choice of telephone 'menu' options that related to the internal structure of the organization ('Key 1 for hardware, 2 for software.' Then 'Key 1 for printers etc.').

The statistics from the 'phone system would tell managers the number of calls that went to the different groups, but they were incapable of describing the calls from an outside-in (customer) point of view.

When agents from the call center studied the nature of the customer demand by listening to customer calls, it became apparent that a different classification of call types would make sense; customers predictably wanted different things from the choices assumed by the original call-routing rules. Customers' demands were better classified by such descriptions as: 'I want some advice'.

Before this work was done, managers of the Center had no data that told them anything about why customers were calling. All they had was the volume of calls in total and the volume of calls into each specialist group.

The new 'customer classification' of type led people in this call center to optimize the different processes ('sending an engineer', 'shipping a part') so that the processes were designed from the customers' point of view.

As an aside, the company disabled the IVR system - the better solution was to have an expert take the call, agree what would be the right actions ('choices') for the customer and then execute them.

Almost without exception, in the organizations we have worked in, managers have adopted an 'internal' perspective when considering types of calls. Internal distinctions are no match for understanding demand and value from the customers' point of view, which provide the means for improving the way a call center works.

Example: IT sales & support
An IT organization has separate divisions for 'service' work and parts sales. Any call to the parts sales call center that required technical support and problem solving ('service' work) had to be responded to by telling the customer to ring another number.

Similarly callers for parts arriving in the 'service' call center were advised to call parts sales. It mattered little to the divisional managers whether customers did as they were advised - each was focused on making his own numbers. An analysis of demand and flow showed that the majority of these callers did not follow the advice.

Putting resources into the places where demand occurred increased revenue immediately; but the solution first required managers to ignore their traditional functional demarcation.

When we first discuss the ideas of understanding demand with managers, they invariably tell us they have a method of classifying call type. Generally, they ask service agents to classify a call once it has been completed. We have never found such methods to produce reliable information.

The classification of calls has usually been provided to agents by their managers, and agents generally find it doesn't work very well - the classifications being hard to apply because they have been based on internal or 'management' view of the work. Moreover, agents, being focused on taking calls - 'making their numbers' - do not complete such tasks reliably.

We have yet to find a classification used in a call center that coincides with customer classification derived from listening to customers of that call center.

Customers make demands for service (in this sense sales is sometimes service too). To take an operational view of how a call center deals with customers, you need to know what happens to customers at the points of transaction; what demands they make and how the system responds to those demands.

You should take the view that you want to know how well customers 'pull value' from the call center. If a customer demand is met with a response that fulfils it, and only does that, the consequence is good service at the lowest cost.

Taguchi
It was Taguchi's work that first helped us think in this way. Taguchi is one of the leading thinkers in the quality movement. Most of the methods associated with quality thinking were originally developed in manufacturing organizations. In the design of production methods, Taguchi challenged the idea of working to 'standards' or 'blueprints' which meant 'working within tolerances' and showed that setting any (nominal) value and working to continually reduce variation around it resulted in better quality and lower cost.

When manufacturing goods, working to standards or tolerances means tolerating variation. In simple terms, the more variation, the more likely something will go wrong.

Taguchi explained his ideas like so: The further any item was from the 'nominal value', the greater the economic loss to the system; the more things go wrong, break down or take longer to deal with. Doing more than is required, for example over-specification, is another potential loss.

Too few organizations have understood how to apply this thinking in manufacturing, although the number is growing. Fewer have understood how to apply the same idea to all transactions with customers.

Think of any service you regularly encounter. If the organization understands and responds to what matters to you (your nominal value), you experience good service and the organization is likely to be delivering it in the most economic way. If, for any reason, the organization does not recognize and respond to what matters to you, your service experience is poorer and the chances are the organization consumes more resources in resolving the situation.

Unfortunately, call centers are, as we have seen, designed as though they were factories, using mass-production or 'traditional' thinking. Staff in customer service functions are told how they are to behave. When such prescriptions ignore or interfere with the 'nominal value' of customers, as they so often do, sub-optimization occurs.

Failure to respond to the nominal value(s) of the customer is just one form of sub-optimization. Worse than that, by treating work as 'units of production', mass production ignores the nature of work. And a major cause of sub-optimization in call centers is the volume of 'failure' demand.

Failure demand
In our experience, call centers have at least 30%, and in some cases as many as 75% of calls that can be classified as failure demand. Customers might call in to say...'where is it?'...'you've sent this but it's wrong'...'I've already called about...', and so on.

In every case we have studied we found calls into the centers that customers should not have had to make. The typical response was to manage the call as a normal part of the call center operation, instead of seeing the need to treat such demand as something to be removed. In some organizations we have worked in we found IVR menus giving options for failure demand (for example, progress chasing) - it does not occur to managers that this is institutionalizing waste.

A public service organization's call center was found to have 65% 'failure demand'. The largest type was caused by the practice of attaching all cases to specified case owners. Most calls relating to cases could not be dealt with, as the owners were not available at the time of the calls. The caller had to call back, increasing the 'costs of production' and worsening the service experience.

Call centers should not be 'designed for failure'. Management's job should be to remove the cause(s) of failure from the delivery process.

Of course, adoption of this idea has dramatic implications for how call center managers - and, for that matter, all managers - spend their time. To remove the causes managers need to be able to react across organizational boundaries. But all-too-often, managers find themselves impeded by functional structures, roles and measures.

Managers are often reluctant to give up their resources (people); sometimes managers are blatant about the need to keep their resources to maintain their job grading or make their numbers. Functions may succeed but the system loses.

Example: telecommunications
A telecommunications service provider improved its bottom line by millions of dollars in six months. One of the actions which made a significant contribution was to find out what was predictable in terms of customer demand and re-structure the organization to put all of the necessary skills into the call center.

In practice it meant moving whole departments into the front-line. The purpose was clear to all and very simple; they set out to learn how to answer all calls at the first point of contact. The leader recognized that good service would be cheaper - and it would keep more customers.

The secret to the above organization's success was first establishing what was predictable about customer demand. The only reliable way to establish predictability of demand is to take measures of every type of demand over time (and it may be important to use a control chart). It is important NOT to act without first establishing predictability, for to make changes when demand is unpredictable might makes matters worse.

Establishing the nature and predictability of customer demand is the first step for transforming call center operations.

Demand, value and flow
In every call center we have had experience of, we have found the work design to be based on the principle of functional specialization. When what actually happens to the work - i.e. what happens to the customers' demands - is studied, sub-optimization becomes evident; sometimes it is extreme. In the telecommunications example (above), the volume of work being passed to the 'back office' inevitably caused delays, caused customers to call back and both of these things added to costs (and worsened service).

Typically, in such circumstances, managers establish procedures for calls that need to go across boundaries. The work of the service agent becomes filling in forms, which is not 'value' work but waste, as it does not solve the customer's problem, but merely passes it to someone else for resolution. Further waste is caused when the service agent does not (for whatever the reason) give the recipient the information they need to solve the problem, which can occur for a variety of reasons.

Procedures only work well when the nature of demand is predictable and the 'value work' - the work that needs to be done to service the customer - is always the same. This is why 'simple' call centers (for example retail enquiries) are able to exploit standard procedures - the customer demand also has 'standard' characteristics.

Instead of being concerned with establishing and managing procedures - 'what to do when you get one of these' - managers should start by getting a thorough understanding, from the customers' point of view, of why customers are calling in.

In some call centers, functions are established for handling different types of calls. Sometimes this takes the form of specialist groups, sometimes specialist levels and sometimes a mixture of both.

Example: leasing company
The management of a leasing organization decided that customers would prefer to have one central number to call rather than have to call their local branch, the rationale being that customers required longer 'opening hours'. Management believed that it would only make sound economic sense if the initiative was funded centrally.

As management believed they understood the kind of things customers would ring up about they recruited inexperienced (i.e. low cost) staff and trained them (gave them prescriptions) in what to say to customers. This, in turn, they believed would allow them to release people from the branches (make them redundant), and free up their technical (high cost) experts to spend more time on the more complex work.

Almost immediately it was found that the new hires did not have the necessary knowledge and experience to deal with the types of questions and problems customers were calling about. Managers then set about establishing second-level support functions. These second-level functions were defined in terms of customers' status - new, lease in progress, end of lease, arrears and, moreover, were split by lease value.

At the time of our investigation there were nine level 2 support functions and level 3 was in the process of being designed and implemented. The amount of 'value' that a level 1 service agent could create for a caller was limited to providing a statement of current lease details, one of the least frequent types of call.

This type of situation is more common than one might imagine. Many managers have pursued a strategy of putting 'low cost' agents as the first point of contact with their customers. Managers can sometimes point to 'savings' in terms of costs of resources (people), but they cannot see the costs they are causing. The costs only become apparent when one studies demand, value and flow - the demands customers make, the value work associated with those demands and the flow of work responding to those demands across the organization.

Example: IT support
In 1996, an IT organization's Enterprise Software support division was redesigned using the principles of demand, value, flow. What mattered to customers was getting through to someone who could adequately and quickly understand their problem. In previous organization design the 'specialists', being the most expensive resource, were kept at the back-end of the workflow and the customers experienced a number of hurdles in getting through to them.

Moving the specialists to the front of the flow improved the speed with which 'value to the customer' was understood. In effect, the redesign meant that each customer demand created its own process.

In a matter of weeks, service, efficiency and morale improved. The specialists had control of the work, they continually monitored demand, value and flow and took immediate action on opportunities for improvement.

The specialists, as discussed above, had clarity of purpose, measures which tracked performance against purpose and freedom to experiment with method. In the previous design the specialists were treated as resources in functional silos, and measured by activity. Improvements in service efficiency and morale go hand-in-hand when the system - the way work works - is changed.

The weakness of functional specialization is exposed when call centers always receive calls that do not fit into the pre-determined categories. It is better to think about the design of call center work in terms of demand, value and flow rather than functional specialization.

Here are some other inappropriate structural 'solutions' we have seen:

Example: tail chasing
Customers of a call center were given an option to route themselves to a chase group ('press 4 if you are progress chasing a previous call'). The chase group became overloaded. Managers decided the solution to the problem was to get service agents to call customers to give them an update, the idea being that this would cut down customer chase calls.

Managers started measuring how many up-date calls agents made and compared these numbers to missed deliveries. Incentives were introduced for individual service agents to achieve up-date targets.

What the managers were not doing was solving the problem in their system. To do so would have meant working across functional boundaries as the failure lay somewhere in the service delivery processes. As the problem grew worse, managers put more resource into up dating, without which they would have failed to meet their up-date targets.

Management's focus on internal issues blinds them to the customers' experience and, hence, the auses of costs in their organization.

Example: financial services The managers of a financial services organization had established a number of call centers and, as part of the justification, had cut staff numbers in the branches. To limit the work returned to branches by call center service agents, strict procedures were introduced.

If an agent were found returning work to a branch which, in a manager's view, should have been completed at the call center, there would be serious consequences. Yet, often the customer wanted to talk to someone with whom they had a relationship; these people were based locally. The call center process set up a direct conflict between managers' rules and customer demands, with the agents caught in the middle. The impact on customers can easily be imagined.

Conclusion
The focus of a call center should be creating value for customers. The better way to design call center work is to design according to demand - why customers predictably call in- and flow - how to ensure customers get what they need in the most efficient manner. As we saw in both the telecommunications example and the software support example, flow can only be optimized once demand is understood from the customers' point of view .