Editor's note: Barbara Rugen is president of Audience Impact Research, Maineville, Ohio. She can be reached at 513-583-5704 or at brugen@audienceimpact.com. This article appeared in the June 28, 2011, edition of Quirk's e-newsletter.

Product success depends heavily on the success of designers in communicating purchase criteria and brand identity in product, package and logo design. The problem is that, in testing the design concepts, it can be very difficult for respondents to articulate and evaluate what the concepts are actually communicating to them.

Design semantics is a qualitative tool for industrial and graphic designers intended to capture what design concepts are communicating to respondents and profile their perceptions. Design semantics ensures that respondents develop a shared definition of attributes with each other and with the designers and that they discuss, measure and rate the distinguishing attributes in the design concepts.

Reporting includes a profile of each concept plotted per segment or per market, according to your needs, plus analysis of the accompanying qualitative discussion; a comparison of each profile with ratings of likeability and purchase intent; and concluding analysis of the design elements that are revealed to engage and motivate consumers in the marketplace.

Design semantics is an adaptation of the semantic differential - a quantitative tool - to qualitative research. Typically, the semantic differential serves in quantitative research to measure the strengths and weaknesses of a product or service by having respondents rate it between positive and negative descriptors; the mean of the responses is plotted to provide a profile. The data distribution is charted and percentages can be determined.

In a qualitative design study, getting such a profile enables designers to identify clearly what a design communicates to respondents. But we must rethink the standard quantitative approach for our design purposes. We replace strengths vs. weaknesses with attributes that are non-judgmental and randomly organized to capture and measure what the design communicates. Figure 1 gives an example of the different application.



Let's say the goal of the client - a manufacturer of electronic appliances - is to understand what consumers like in a television design that would stimulate interest and motivate purchase intent in the online or store retail environment. We'll use the semantic differential to get - as much as possible - a precise, measurable profile of what turns consumers on in television design and use this as directional information for the designers. Figure 2 provides an example.

The there are five separate steps necessary for success in design semantics.

Step 1: Ensure respondents share definition of terms with each other and with the designers

It's important to use the attributes the designers agree to, rather than terms the consumers may come up, which may be frustratingly off-target for what the designers need.

In focus groups, we introduce the attributes and find out what the respondents understand by these terms to make sure we all mean the same thing. We use fairly common terms but want to be certain. The best approach is to start with an exercise. There are many possibilities for an appropriate exercise but here is an example of one that we tried.

As a homework assignment, we had the respondents bring in three or four photos of furniture or décor in their homes that they really like. After they describe their photos, we suggest doing an exercise to help them pinpoint their design criteria for their homes.

Looking at your photos, circle one thing that really feels right for you and your home. Now find one thing that really represents the opposite of that.

Describe the shape or personality of the one that's right for you in one column and the shape or personality of the opposite one in a second column. Here's a list of adjectives that will help guide you:

[Attributes to be used in the semantic differential]

Respondents did this a couple of times so they became fairly comfortable using these adjectives to interpret design. Then they shared their findings with a partner. Would the partner select the same familiar adjectives to describe the circled items in the photos or not? Always they did. Then we debriefed, showing the examples of what we all meant by these terms.

Step 2: The semantic differential exercise

So now the respondents are in sync with each other in using the terms the designers want them to use. We use the semantic differential to profile and measure consumers' television design preferences.

Consumers see four television design concepts and are told all four would be equal in price and capabilities, differing only in design form. Figure 3 shows what the consumers had to do to evaluate each of the four concepts.

The results give us a profile of how the respondents see each concept compared with how much they like and want to purchase each concept. For example, if they like Model L 10 out of 10 and Model Q only 7 out of 10, we can compare their profile of what they're seeing in Model L - the one they like most - with what they're seeing in Q - the one they like less - and so on for all four concepts.

Step 3: Qualitative follow-up

The follow-up discussion covers shared likes and dislikes and focuses on the respondents' impressions of their favorites; any associations or metaphors their favorites evoke; comparing their respective favorites; and assessing what, if anything, they might change or borrow from another concept to make their favorites even more appealing.

Step 4: Competitive set - measure and discuss

Next they are shown unmarked competitor models, which they evaluate in terms of the same semantic differential attributes (Figure 4). In this way we can compare what they are seeing in the client models that they like with what they're seeing in the leading competitor models that they like.

Then we can discuss whether any of the additional selections is now their preferred choice; what elements, if any, from any of these additional models they would like to merge with their original preferred choice; what company they think produces their preferred choice; which company may make the one they like least and so on.

Step 5: Reporting

In the report, we are able to provide the composite profiles of each model, along with the qualitative comments, to analyze and guide the client on how to produce the design that best captures interest and motivates sales.

The qualitative comments will yield associations and impressions in our analysis about the design concepts and provide insights into purchase intent. The composite quantitative profiles measure in design terms what such comments mean.

As in quantitative research, the profiles are created by plotting the mean of responses for each pair of adjectives. Figure 5 shows that the aggregate of respondents in New York (total of 24 respondents) felt very strongly that Model C was a new and contemporary design concept for a television and were a little less certain that it looked compact and robust. They were neutral about whether the feel was more soft or crisp or more serious or playful.

Where the mean lands in the middle - say around 4 on a seven-point scale - that tells us that the design element either leaves most people indifferent or is polarizing. Either response is a red flag for a designer. A check of the data distribution revealed that, in this case, the New Yorkers were conflicted about those design elements.

It's important to remember that, given the small sample, the findings are directional but the measurements are precise. Besides the mean, which shows where most responses fall toward either side of the paired opposites in the overall concept profile, the data distribution is charted and percentages can be determined.

Makes such guidance possible

The product or package designer's goals are to attract consumers to the product in the retail/online environment; to motivate purchase; and to embody the brand identity. Getting consumer respondents to make this possible for designers - getting them to understand and use design language and to articulate their responses to design concepts in order to guide the designers - has always been a challenge both qualitatively and quantitatively. The benefit of the qualitative design semantics process for industrial and graphic designers is that it helps make such guidance possible in a way that is precise, insightful, reliable and measurable.