Technique helps focus the product development process

Editor’s note: Karen P. Gonçalves is president, and Michelle Gonçalves is consultant, at Delphi Market Research, Medford, Mass.

We have all heard about the Kano method, and how it can supposedly allow researchers to quickly and inexpensively break down preferences about a certain product or service and thus:

  • accurately focus the direction of development;
  • lower product and service development costs;
  • shorten time to market and;
  • increase the probability that new or refined offerings will be successful.

Clearly, these are very big claims, so it is worthwhile to explore if and how you might use this method.

The Kano method, developed by Dr. Noriaki Kano of Tokyo Riko University less than 10 years ago, is related to conjoint analysis, in that comparisons, rather than discrete answers, are an important part of the interpretation. But unlike conjoint analysis, the respondent is not asked to make trade-offs or choices among options. Rather, respondents independently rate their pleasure or displeasure with the way in which up to two dozen variables are delivered to them. The pleasure/displeasure answers allow researchers to classify each variable as attractive, one-dimensional, must-be, or indifferent. These categories are defined as follows:

  • An “attractive” attribute is simply that - attractive. There is no penalty for not including an attractive element, but including it makes the product or service better to use, differentiates it from competing offerings, and will “delight” the user. An example of this might be high-end speakers for a computer. They can add greatly to the enjoyment of the offering (one can now listen to high-quality music while working), but very few people would not buy a certain computer because it did not include high-end speakers. They are an attractive extra, marketed as an element that differentiates your computer from competing models.
  • A “one-dimensional” characteristic is one where more is always better. A lower degree of functionality in a one-dimensional characteristic will displease customers, and a higher degree of functionality will please them. An example of this might be processor speed in a computer. We all want faster computers, and we have yet to find a consumer who decides that a computer is just too fast. Furthermore, a slower computer will always be perceived as less pleasing.
  • A “must-be” characteristic is one the customer considers essential. It is much closer to a binary rather than a one-dimensional characteristic. Without it, a consumer will not buy - you are not a serious contender in an industry unless your offering provides all of the must-be characteristics. Continuing with the computer example, a CD-ROM drive (once an attractive characteristic) has become a must-be for most consumers. Without it, the computer would be vastly less functional, and with it, the computer is functional. Furthermore, there is only a small percentage of users for whom a faster CD-ROM is important.
  • An “indifferent” attribute is one that consumers do not value. It is important to note that each segment of your market may be indifferent about different characteristics. While this may not be the case for a product or service with a very narrowly defined use, a general use product or service may show a large degree of variation in the “indifferent” category. While most people use a computer, there are different types of users, all of whom have different needs. A graphic designer may consider a large, high-resolution monitor essential (a must-be); a writer might be indifferent toward monitor size and resolution.

Once each variable is categorized as attractive, one-dimensional, must-be or indifferent, researchers use the rest of the survey instrument to further break down and understand consumer needs and preferences.

Creating a Kano method survey

There are typically four components to a Kano method survey. The first is the Kano questions themselves, which consist of a series of up to two dozen paired-comparison questions, each of which is answered using an ordinal, five-point scale. It is important to note that the five-point scale is not an interval scale, which affects the analysis. There has been a fair amount of experimentation surrounding the exact wording of the five-point scale, and our company’s clients have each handled the scale differently. Typically, we work with them, either during the pre-test, or over time (when they are conducting a series of Kano method surveys), to determine the exact wording that their target markets will understand and respond to objectively. The Center for Quality of Management (CQM), Cambridge, Mass., was the first organization to publish extensively on the Kano method and their work is very helpful in deciding if Dr. Kano’s exact wording or if a variation on his words is best.

It can be difficult to get the wording of the paired-comparison questions exactly as you want them, but there are ways to help ensure that the questions are both thought-provoking and helpful to decision making. Most firms using Kano method surveys include it as a second or later part of a multi-step new product development process. This also helps to ensure that the questions are meaningful, because they are designed using newly collected data directly from a select group of customers and/or prospects. The most important goal in designing these pairs of questions is to ensure that one half of the pair is “functional” and the other half is “dysfunctional” - in other words, that respondents see one as a positive product or service characteristic, and the other as either negative or less positive than the first half.

The second part of the survey is the self-stated importance (SSI) component. In this section, consumers state how important or unimportant various product or service characteristics are to them. This more direct questioning allows researchers to glean significant insights into the nuances of consumer opinion. One can determine the degree of importance, the strength of opinion, and the unity of opinion among respondents. Once these are established, one can rank variables within each of the four categories - an important detail when as many as two dozen variables are being studied.

The next portion of a typical Kano survey is the utility section. Sometimes referred to as functionality or parametrics, these questions address how respondents use the product. While client companies may assume they know how a product is being used, there are sometimes surprises. For example, one of our clients believed that customers in its three vehicle markets used a certain component in the same way in all three types of vehicles - that component size and expected life were the only differences that might affect product design. Through a Kano method survey to all three of these markets, this client learned about several other key differences that have allowed it to refine its component design, and to gain market share in two of these three markets. Especially when an offering is complex, the utility section of the questionnaire can be invaluable.

Often, respondents can be logically grouped into different market segments, each of which uses or values the product differently. These subgroups, which are based solely on product utility, can provide important information to researchers. As mentioned earlier, each functional subgroup may have entirely different needs or usage patterns than the next. A variable which for one group is a must-be could be attractive for another. This type of specific information is crucial in focusing product development, knowing what messages to include in sales and marketing materials, and in selecting appropriate target markets.

The last part of a well-designed Kano questionnaire asks demographic questions, which are used to determine who makes up each subgroup as well as who the customers are. Unlike the utility section, which focuses on product use, the demographic section focuses on who the users are, who the decision-makers are, where they fit within the corporation (for business-to-business surveys), and how they evaluate each variable included in the survey.

When the entire questionnaire is analyzed, Kano offers a remarkably complete picture of the market’s needs, usage patterns and preferences. It also highlights which segments have different or unique needs, and how important those needs are in their purchase decisions. This analysis can help determine which characteristics of the offering need to be enhanced or completely changed, and which ones should remain the same. It can also highlight which aspects of an offering the customer simply does not care about, thus, offering opportunities to simplify or streamline the offering. Finally, Kano method surveys can be very helpful in delineating and characterizing key market segments, each of which may be targeted in a different way, to respond to their specific needs, usage patterns and preferences.

Applying the Kano Method

Kano is always administered in a visual format. The respondent must be able to see all of the paired comparisons, and to see how they relate to the self-stated importance questions. In fact, respondents often go back and forth between different sections, and it is not unusual for them to write notes in the margins (or in designated spaces for on-line surveys), to enhance their answers to different parts of the questionnaire. Therefore, telephone or in-person interviews are not appropriate. Kano works best as either an on-line survey, or a traditional mail survey. It can also be self-administered individually or in a group setting, as long as the respondent can go back and forth among the sections as often as he or she wants. This can be an important advantage for resource-constrained clients, because mail and on-line surveys tend to be much less expensive to administer than telephone surveys, they take only slightly longer in total elapsed time, and it is possible to achieve higher overall response rates with mail or on-line surveys than with telephone surveys.

Mail surveys bring with them any number of difficulties which must be addressed. Often the largest problem is acquiring an accurate, appropriate list. However, since customers are typically included among recipients of the survey, the client can generally provide, generate or purchase a reasonable list. Acquiring a good list of prospects is also crucial. Find a good list source and keep them close by. Once the prospect and customer lists have been checked for duplicates, errors and accuracy, you’re ready to send.

Attaining a high response rate with a mail survey is not as difficult as one might assume. Although telephone surveys are traditionally associated with a higher response rate than mail surveys, that has changed in recent years. Within our firm, business-to-business mail surveys typically yield response rates of about 20 percent and consumer mail surveys yield response rates of about 40 percent. Our president has conducted considerable research, both within academia and with commercial clients, on how to increase response rates to mail surveys. She has tested various combinations of outbound survey content, a range of incentives, and questionnaire length, content and format. Her dissertation included a mail survey which yielded a 75 percent response rate. Within the commercial sector, using what she and the rest of the staff have learned, our firm has attained business-to-business response rates up to 43 percent and consumer response rates up to 60 percent for traditional mail surveys.

Analysis

Occasionally, during analysis, two additional Kano categories emerge: “questionable” and “reversal.” Questionable variables are those for which the respondent has given answers which are not internally consistent. For example, they are extremely satisfied with both the functional and dysfunctional halves of a paired-comparison question. This typically means that the question was confusing to respondents, or that there was simply not enough difference in their attitude toward both halves of the question to generate two different responses. Sometimes, further analysis clarifies these responses; other times these questions need to be discarded.

A variable is classified as a reversal if it is rated most highly when dysfunctional. This implies either a failure of respondents to understand the question, or an incorrect assumption by the client regarding product utility.

Most well-designed Kano method questionnaires do not include reversals or questionable variables, but they occur occasionally. Both can be helpful in convincing clients to reconsider pre-conceived notions about a market, or in learning new things about a market that no one knew before the survey was conducted.

Conclusions

Clients tell us that Kano method surveys are extremely helpful in focusing the rest of the product development process. It can be instrumental in arguing against further expenditures on characteristics or attributes that customers simply do not care about; it can also provide an early warning system when attitudes or needs are shifting, or when what used to be “attractive” attributes are becoming “must-be” or “one dimensional.” Perhaps most important, the results of a Kano method survey are often very revealing regarding product uses or applications that the client did not know about, and in dividing the overall market into manageable, discrete segments that can be targeted separately.