Begin at the beginning
Editor’s note: Jerry W. Thomas is president and chief executive officer of Arlington, Texas research firm Decision Analyst.
The market is changing, and the time has come to redesign the package of that old established brand. This will revitalize the brand, the marketing manager supposes. Or, a new product with great promise emerges from the laboratory and a new package must be created from scratch. In both of these instances, the process of developing a new package tends to be the same. The package design firm attempts to get the client (the brand manager, the new products manager, the marketing vice president, etc.) to lay out the vision for the new package.
In a perfect world, the client would have a clear vision and accurately communicate this vision (meaning marketing goals and packaging objectives) to the package designers. However, the manufacturer’s vision is often clouded with confusion and omission, and the result is no direction or poor direction for the new package design work.
Regardless of the clarity of the manufacturer’s vision, the package design firm goes off, ponders the vagaries of the universe and creates 20 or 30 rough designs for the new package. Typically, these designs are screened by the marketing executives and a few designs are chosen for further development. These finalists then go through another round of managerial reviews and a new design is chosen, based on the “expert” judgment of the marketing staff.
The stage is now properly set for a marketing disaster; poor package design is often a major cause of marketing failure.
Never smart enough
If the marketing professionals had flawless judgment, marketing meltdowns could be avoided. Unfortunately, marketing departments are never smart enough to see the market, or the new package, through the eyes of the average consumer. The marketing staff knows too much and is blinded by that knowledge. The marketing staff is biased by the mythologies of their profession, industry and company. The marketing staff is rarely similar to or representative of the ultimate consumers of the product (the marketing professionals tend to be much better educated and much higher in income than their customers). Also, marketing folks are often shielded by the corporate bureaucracy from the realities of the messy, helter-skelter marketplace. Lastly, two or three marketing executives are too small a sample for their decisions to be statistically reliable.
So, marketing executives and their judgments cannot be trusted. Can we turn to research for assistance, or is the research itself flawed?
Marketing research is not perfect. It has its biases and its blindness. Research tends to favor the status quo. It’s an inherent bias in virtually all marketing research. This tends to mean that an existing package, the benchmark, will almost always outscore a new package design. This bias must be understood and taken into account in interpreting the results of packaging tests. There are many other research pitfalls (sampling problems, study design issues, questionnaire problems, analytical problems, etc.) that can invalidate the research. That’s why you must often rely on researchers with some gray hair, wrinkles and humility to help oversee the research design and the interpretation of the research results. Despite its limitations, research improves the odds of success, compared to three or four executives sitting around a conference table.
Back to the beginning
What are the keys to using research to develop and evaluate package designs? Let’s go back to the very beginning of the process, the prepackage design phase. The research should ideally begin at the beginning. Good qualitative research can be invaluable in helping marketing executives set the correct goals and objectives for the new package design. This qualitative research should include an ethnographic component, the observation of shoppers in a natural retail environment, and perhaps in a usage environment. How do consumers shop the category? How much time do they spend in front of the display? How many packages do they pick up? How many shoppers read the details on the label? How many packages of what sizes are purchased? How does the consumer interact with the package in the home before, during and after usage?
After observations of actual shopping and usage behavior, in-depth interviews should be the next step. What are the primary motivations and perceptions that drive brand choice? Are respondents aware of brand names or do they buy based on the color and design of the package? Are they aware of advertising in the category, and are they aware of advertising for specific brands in the category? Can they remember and describe the existing packages in the category? What do they remember, or think they remember, about these packages? Do they buy one brand only, or shop around and buy different brands from week to week? How frequently do they buy the category and the brand? How frequently do they interact with the package once it arrives in the home? A breakfast cereal package, for example, might sit on the breakfast table every morning for a week before the package is discarded. A frozen-food package, on the other hand, may only be seen once when it is pulled from the freezer and opened. What are the advantages and disadvantages of the packaging and graphic designs for each of the major brands in the category? All of these little details, once fully understood, can help the marketing professionals set clear and relevant goals for the new package design. Helping to define and set the correct goals at the outset of the design process is perhaps the single most important contribution of research. Once the design goals are set, the package designers are ready to go to work.
As noted, the designers will typically create a large number of rough designs (anywhere from 10 to 30 or more early-stage designs). The goal of research at this point is to identify the better designs and screen out the “dogs.” Typically, this screening is done via online surveys. A sample of 200 to 300 target audience consumers are shown all of the rough designs (that is, each respondent sees every one of the designs in randomized order) and answers several questions about each design, such as:
- attention value;
- purchase interest, based solely on the package design;
- uniqueness or dissimilarity to competitive packages; and
- fit to or compatibility with the brand.
The results of these questions are combined into a scoring model so that all of the package designs can be ranked from best to worst. The output of the model is the identification of the four or five “better” designs.
Note: an alternative approach at this early stage is the creation of a number of rough package designs with the elements systematically varied according to a choice-modeling experimental design. In this approach, each respondent sees a test package in a competitive context and chooses the package she would buy. Different respondents see different sets of package designs, with different elements, illustrations and messages. Then, the statisticians crunch all of the numbers and mathematically infer the importance of the various variables and measure the relationships among the variables. Based on this enhanced understanding of the utility or importance of different variables that make up the package design, a small number of improved package designs can be created for the final stages of testing.
Fine-tuning
Regardless of method, once the “better” designs are identified, they are often put through another round of qualitative research (generally, depth interviews are better than focus groups as they can dig down into the details more extensively). The goal of this qualitative research is improvement and fine-tuning of the better package designs. Every little detail of each design is explored in these one-on-one interviews, searching for any little tweak that could improve consumer reactions to the packages.
The final step is to take each of the “better” designs, as improved by the qualitative review, and test each design against a constant set of major competitive packages. Ultimately, each package design must be evaluated in the context of this competitive environment. Each test package is placed in a simulated display, similar to a display the consumer might see in a typical store. Each package design is tested monadically among a representative sample of target-market consumers (identical or matched samples of consumers see each test package in a competitive context). At this stage, sample sizes might be 300 to 500 consumers per package design. Some of the key measurements are:
- attention value of the package design;
- purchase intent and share of market versus competitive packages;
- expected purchase frequency;
- brand fit or compatibility;
- image projected by the package design; and
- pricing expectations related to the package design.
Naturally, one of the matched sample cells is exposed to the existing package design in the same competitive context as a control or benchmark (the existing package). The test package must come very close to, or equal, the results for the existing control package, before a change to a new package is recommended (remember, the status quo bias in marketing research).
The exact nature of the testing at this final stage can take many forms. For example, displays can be simulated and the testing accomplished online. Virtual online simulations of in-store environments with simulated shopping displays can be used as the format for the final testing. Or, actual displays can be constructed and target market consumers recruited for in-person interviews. Tachistoscopes (or timed exposure) can also be employed to show the packages for fractions of seconds to help identify the attention value and recognition of various package elements (ideally used as a supplement to traditional methods), and/or eye-tracking cameras can be used to see which package elements attract the eye in what order (again, recommended as a supplement to traditional methods).
If there is any possibility that the new package design will negatively affect the product inside, in reality or in perception, an in-home usage test (or test in the product’s natural usage environment) is always recommended. The product in its new package is placed in homes of target consumers and they are asked to evaluate only the product itself. These results can be compared to results for an identical in-home usage test of the existing (or old) package. If the new package improves perceptions of the product inside, then you may choose to invest more money in introducing the new package. If the new package negatively affects perceptions of the product, then you should go back to the drawing board and start over.
Great care
In the final evaluation of new package designs for an established brand, great care must be taken to ensure that the new package design is not so different as to break continuity with existing consumers. We have seen radical new package designs trigger precipitous declines in a brand’s market share. The greater a brand’s market share, the greater the risks associated with a radical change in package design. If a brand has a tiny market share, the downside risks of a new package design are much lower. Any radical change in package design for an established brand should be supported with massive advertising, as though a totally new product were being introduced.
Finally, if a brand is not adequately supported with media advertising, the packaging must play a greater role in the brand’s marketing. Without advertising support, the face of the package must be thought of as the principal advertising medium. This makes packaging research even more important, because the package has to carry the preponderant burden of marketing the brand.