The first step in the QFD process is to identify the exact expectations of the customers. We can see immediately why this approach brings dis- tinct advantages. To obtain customer input, we have to listen. This calls for a reassessment of the listening devices that we already have, and usually for their replacement with more effective methods. The mechanisms do not necessarily have to change, but they often have to begin to be man- aged actively rather than passively.
Passive monitoring of customer comments is rarely likely to yield the amount of detail that is needed, but many organizations rely on this as their only means of collecting information about customers. Even when sophisticated tools are used in the marketing department, their results often fail to percolate to the other departments, and especially to the prod- uct developers.
Among the tools that fall into this category are customer service desks that only handle sales or specific complaints. The telephone numbers of many of these are conveniently hidden on the label or in the small print of a document. Levels of activity are usually piecemeal.
In the 1980s, a television advertising campaign run by a lager brewery played on the lack of use of these telephone numbers. It showed a man walking along a high-tech corridor when he heard a telephone ringing.
He rushed around until he found a door marked ‘Customer Complaints’.
Upon opening it, he discovered a room furnished with old wooden desks and a captain’s chair. There were cobwebs everywhere. On the desk was an old-style mechanical dial telephone. Answering the telephone, he found that the person at the other end of the line had dialled the wrong number!
Other response mechanisms are likewise passively managed, and as such yield very limited quantities of information. Most hotels have a customer satisfaction survey form. The way in which these forms are managed ranges from the totally passive to the highly proactive. On arriving at some hotels, you are handed the form with your key and the receptionist encourages you to complete it. Some forms have incentives to complete them, such as free weekend stays. At other hotels the form
is to be found hidden between a crumpled bit of hotel stationery and an old magazine in the top drawer of the desk in the bedroom. Completion rates vary tremendously. Only rarely will you be approached personally by a manager or member of staff and asked about your stay. If we are serious about obtaining customer feedback, then we can do so.
In the past few years there has been a lot of criticism of airport ter- minal areas. Holiday makers awaiting the departure of their flight have complained bitterly about the lack of comfort and facilities. As the air- port operators began to see the (largely financial) benefit of providing extensive shopping and better quality catering, franchise operators have moved in to enhance these environments. Even so, you will have trouble finding a suggestion box.
We checked the opportunity for customers to complain at three major British airports. Only one had suggestion boxes. Supplies of comment cards had run out and there was no facility for writing on the card. The boxes were mounted in obscure places and well out of sight (and mind) of most passengers.
At the other two airports boxes were not provided; instead, we were told that passengers could always speak to a member of staff, particu- larly at the customer information desks. But many people find com- plaining, especially in person, emotionally very difficult, so it is unlikely that this approach will yield many responses. In any case, only one of the information desks had any kind of procedure for recording and subsequently analysing customer feedback, and two were unmanned.
Compare this with one European airport, where multimedia com- ment points enable passengers to write their concerns on a card, type them onto a screen or record them by voice.
Gathering analysable customer feedback is so rare that the places where it is solicited are much easier to recall. As with all the tools and techniques of quality, it is not what you do, but the way that you do it.
Most of us have heard of the survey conducted in the USA in the late 1970s that demonstrated that only about one in ten dissatisfied customers bother to complain. It is hardly surprising that such passive mechanisms do not provide us with enough information.
Among the more successful attempts at gathering genuine customer feedback, we should include the growth in the use of 0800 freephone numbers in the 1990s. One petrol company which decided to display its 0800 number in very large signs around its forecourts experienced a
phenomenal growth in the number of people calling in. They were sur- prised at how consistent the causes of the complaints were: most were related to a sales promotion that they were running. It used tokens, and many people felt that they were being short-changed.
During the 1990s, 0800 numbers gave way to 0845 ones: no longer were customers given a free call, instead they had to pay the local call rate. All was fine until British Telecom finally lost its grip on the tele- phone system and many alternative providers emerged. Today, to people on residential telephone payment packages, 0845 numbers are among the most expensive. Do we really want to create yet another barrier to our customers?
There is still a psychological barrier to overcome if you are going to use a telephone number, whether it is free or not. For this reason many other approaches can be used.
The Nationwide Anglia Building Society introduced focus groups as a way of collecting data from its customers. Customers were encour- aged to participate in Saturday morning group discussions, conducted by a trained facilitator. The results from these groups were fed back to both branch and senior management so that immediate practices and policy decisions could be reviewed.
Many companies find that one of the powerful triggers to their qual- ity improvement process is a survey, either of customers or of their own staff. The external survey can be also used as a starting point for QFD.
Organizations vary in how they choose to conduct a survey. A govern- ment agency, the Transport Research Laboratory, chose to use external consultants. The reasons were three-fold. It was felt that consultants would act in a more objective fashion, being less inclined to become defensive under criticism. This, in turn, was expected to make outsiders more likely to respond evenly, whereas most staff would resist being critical directly to the person from the organization. The consultants’ objectivity also meant that their report and recommendations would have greater cred- ibility with the organization’s sponsor, in this case the Department of Transport. Their findings were also more credible to the scientists in the laboratory, because they were perceived as experts in a distinct field.
Surveys, whether they are conducted by the company’s own staff or by external consultants, can be very specific or too general. We have all experienced the lengthy questionnaire sent by a company on the pretext of getting to know its customers, which actually only consists of lifestyle
questions about the consumer. This has little to do with our expectations of the product or service being developed, and a lot to do with creating a saleable database! Worse still is when these have been distributed using a firm’s name on the assumption that customers share an affinity with that business and will therefore respond, whereas enquiries from a mar- ket research company would end up in the bin.
Market research can be a highly scientific process, and its reputable practitioners undoubtedly provide a depth of understanding that many in-house activities cannot. The question is really whether we want people outside our business to know more about our customers than we do, or whether less precise but more direct contact would be better.
Once we have gathered the data what do we do with them? Quality function deployment calls for a straightforward list of tangible customer expectations, such as ‘easy to contact’, ‘keeps in touch’, ‘brings credibil- ity’, ‘doesn’t patronize’ and so on. These are presented in a list forming part of the ‘house of quality’. The list is arranged so that similar topics appear adjacently. This can be done in many ways. One approach that works well with groups who have brainstormed the list is to transfer the items from their flipcharts onto a fishbone diagram (sometimes known as an ‘Ishikawa’ or a ‘cause-and-effect’ diagram). There is a misconcep- tion among some people that the limbs of a fishbone diagram have to be the same regardless of its application. This is not the case and can inhibit groups rather than encouraging them; the team can choose its own themes on the basis of what the list contains (Wilson, 2000).
Another simple technique that can be very useful at this stage is the
‘how–how’ or ‘why–why’ questioning, used by many quality circles since the early 1960s. The golden rule of networking is that if you want to find out about something, you need only telephone three or four people and one of them will have all the information you need. These two tech- niques have a similar foundation: by asking ‘how?’ or ‘why?’ more than two or three times you will identify the definitive cause.
Activity: Review your own organization’s customer listening devices
Ring around your organization. Find out who collects information from customers and what they do with it.
For example, what do customers look for in a public library? Questions to ask include: ‘Why is a good catalogue important?’, ‘Why is an orderly collection of books important?’ and ‘Why are subject experts on the staff important?’After only a couple of iterations, these questions should pro- duce the common cause: easy research. The QFD diagram would show the three original factors categorized under ‘easy research’.
Such techniques are very simple to apply, but they can sound as though they are overcomplicating a simple task. However, in the more compre- hensive application of QFD they are too simple.
Where QFD is applied to a complex process, such as the design of a new drug, the development of a sophisticated computer system or the redesign of a major transport system such as a regional railway network, the num- ber of factors to be balanced can run into many thousands. At this point the human brain becomes overloaded with juggling complex details. To introduce order to such a chaotic system calls for more sophisticated tools.
One such approach is to use the family of statistical methods known as multivariate analyses (MVAs). We look at some of these in more detail in the next section, but for now we will describe only one.
Cluster analysis is one of the simpler MVAs. Like most of these tech- niques, it was originally developed to help taxonomists to determine the ancestry of plants and animals. Using a number of characteristics for each individual ‘customer requirement’, a computer arranges the require- ments into groups that share the same characteristics. The nature of the groups will depend on the type of information that is fed in. This does not normally matter when working with inanimate objects, such as products or services. In fact, it can be enlightening when the computer finds a connection that you had not thought of. Hal MacFie and John Deane of the Meat Research Institute, near Bristol, UK, used cluster analysis, among other techniques, to look at the conflicting customer requirements for a wide variety of products, including cooked meats and aviation fuel (Mottram et al., 1982). For example, they found that of all the tests required by different authorities for each batch of fuel, there were far fewer underlying parameters being assessed. For the company producing the fuel, this meant that whereas before they had to optimize over 40 dif- ferent chemical properties, now they only had to focus on half-a-dozen.
Whereas earlier applications had to use vast mainframe computing resources, which put the technique out of reach of all but the largest cor- porations, today even a notebook PC can be used.
Once established, the customer requirements and their groupings are recorded on area ‘A’ of the house of quality.