6: Finding Customer-Brand Connections
Marketers can no longer rely on the assumption that the only information the customer receives about a brand comes from them. Presently, most information comes from many sources, and customers consider the company to be less credible than most other sources of information. As such, a greater effort must be put into discovering the brand exposures from other sources.
SECONDARY DATA
A wealth of information can be gathered from secondary resources, prior to having to conduct any original resource. The author categorizes data into four area: the brand, competition, environment, and the customer.
Brand information is of critical importance: what the customers perceive the brand to mean, how it is differentiated from other brands, what value it proposes to them.
The author suggests that research can begin with the company's annual reports, marketing plans, promotional literature, and other documents (EN: this is a bad approach, as it indicates what the company wishes the brand to be, not necessarily what it really is.) The author also suggests using a clipping service to find references to the brand in print and online media, to determine how the brand is perceived externally (EN: this would need to be culled to eliminate or lessen the weight of any clipping that is a republication of a company press release). (EN: the author fails to mention any search of the blogosphere or social media sites, which is a new method and gets direct-from-consumer details better than traditional research.)
Competing brands are also important to understand, as they collectively form the customers mindset of the product category and your own brand is seen in relation to competitors - better in some ways, worse in others. Information on competitors can be gathered from largely the same sources as listed above.
The author pauses to consider the definition of "competitor." Companies often attempt to delude customers (and themselves) by choosing competitors against whom they compare favorably. In addition to being objective in identifying competing firms, it's also important to think more broadly, in terms of the ways in which customers seek to fulfill a given need, which may extend beyond conventional product categories, and be especially wary of emerging technology (e.g., Polaroid was not killed by Kodak or Fuji, whom it considered to be its competitors. Instead, it was killed by cell phones with camera functions, which it completely overlooked as a competitor.)
The author describes the marketing environment as the various "forces" outside the company's control, whose communications have an indirect impact on the brand. The examples the author provide are largely legal and regulatory factors that impact the product category (laws affecting wine sales), but it stands to reason that a cultural trend (temperance), social crusade (the AMA announces findings about the effects of alcohol on health), or even seemingly neutral events (a popular child actor killed by a drunk driver) can create a hostile environment, or even a beneficial one. As such, there are a number of "forces" that may impact consumer perceptions of a product category or specific brand, or could pose logistical problems to getting product to markets.
Environmental information is generally available in the media. Consumer publications, broadcast, and the like will provide information that affects the environment, and trade publications regularly report on environmental conditions that impact specific industries.
Finally, companies should gather information about their customer base: the current and changing conditions that affect market segments, as well as changes in their pattern of purchasing behavior for certain goods. There are a number of marketing publications and research services that can provide such information, and a firm's own customer database can be an excellent source of data on current customers.
PRIMARY RESEARCH
While secondary research can uncover general information, there are often questions that cannot be answered by conducting research to discover information, or additional research may be necessary to verify or augment the data available from other sources. This can be conducted by an in-house research team of an outside research firm - the former is more costly and resource-intensive, but the latter gives you less control over the dissemination of research findings.
(EN: The author provides a brief overview of primary research, but it's worth noting that this is a much broader subject that merits closer consideration than the scope of the author's discussion will afford.)
Once the need for information is identified in detail, a study design must be done: to define the research base and determine the best means of gathering data from them. In some instances, a series of studies is prescribed, first to discover information in general, then to delve into specific details.
A word of caution about long-term research projects: in addition to being costly and taking time to deliver results, there is the risk that the nature of the phenomenon may change over the course of the study. This is particularly problematic in industries that undergo rapid change, or products that are subject to changing tastes and fashions. On the other extreme, quick and cheap is not necessarily better - accuracy may be sacrificed for efficiency, which can be detrimental.
It's also noted that all research involves a trade-off between realism and control. Research may create an artificial situation that does not represent real behavior, and subjects who are aware of observation may change their behavior, simply because they are being observed, from its normal patterns.
In most instances, research is conducted on a sample - a small group of people whose opinions will be considered to represent those of the larger population. Generally, a "frame" is used to specify the characteristics of the test group (e.g., unmarried Asian females between the ages of 35 and 40 who have purchased a vehicle in the past 6 months), generally to match a desired market segment. In other instances, a random sample (whomever they can get to reply) can be used.
The author suggests that a larger sample, but in a vague way. (EN: the math of sampling error is easy enough: divide one by the number of participants do determine the probability of error, and subtract that from 100% to determine the level of accuracy of the study - so a sample of five is 80% accurate, a sample of 10 is 90%, etc. My market research classes suggested that a standard sample of 30 is reasonable, which would be 96.7% accurate). Even so, demographic data should be examined - as even a random sample may, by happenstance, be skewed to a certain demographic.
Finding subjects (people) for marketing studies is increasingly difficult. Response rates to surveys is very low - and generally, the more intrusive the method of study, the lower the response rate will be. (EN: it's also worth noting that sampling itself is a bias - the results of a survey reflect the opinions of people who are willing to respond to surveys, which implies character traits like extraversion and gregariousness that may not be representative of the majority of the population.)
The author discusses questionnaires as a common research device (EN: they are common, but there have been some arguments that they are overused, and improperly used - so it's important to beg the question when a "questionnaire" or "survey" is a proposed instrument.) The author provides some general remarks about the importance of wording a question to obtain the desired information without skewing the responses, and that the grouping and ordering of questions is important.
Multiple-choice questions are also common because they are easy to analyze. However, it's of key importance to include an open response (other, none of the above, don't know) and allow any question to be skipped. Forcing a response, or constraining to specific responses, detracts from the accuracy of a study.
Questionnaires that are "too long" detract from response rates - but this can be subjective. The author relates some statistics from a research group that indicates that an interruptive survey (the user didn't expect to be surveyed and is being distracted from a desired task) should be no more than five minutes, whereas an expected survey (the user has agreed to take a survey) can be between 15 and 30 minutes.
Finally, pre-testing a questionnaire is important to fine-tune and adjust. Having five or ten people "test" the questionnaire and provide feedback about the survey experience can prevent problems that would damage the results in the "actual" test. (EN: it's notable that this pre-test data should be discarded, even if the testers match the demographic profile of the desired audience.)
The next step in conducting research is to "field" the study - to deliver it to test subjects and collect their responses. It's repeated that no all subjects will respond at all, and that some will bail, and researchers are cautioned against attempting to force a response; to be especially cautious when dealing with certain kinds of respondent, such as children or the elderly; and to include demographic information to ensure that the respondents meet the desired frame.
There is a sidebar that presents a list of "rights" for test subjects, which contains some good advice: safeguard the privacy of information collected, especially contact information; disclose the name of the company doing the research; do not attempt to "sell" under the guise of doing research; be respectful of respondents' time, both in the amount of time consumed and the convenience of the time the research is done; respect their desire to refuse to answer or terminate the session without question; inform them if the interview is recorded and what will be done with the recording; and ensure that you exercise the highest standards of courtesy and professional conduct.
The author provides scant information on processing the results - literally, "answers go into some sort of spreadsheet program" (EN: this is particularly bad, though I suppose it can be said that statistical analysis is a bit boring and of little interest, my sense is the author should go a bit further, as this is a critical step in which good information can be skewed or mangled to produce inaccurate results, and marketers should pay a bit closer attention to what the math geeks do with the data.)
The author provides some ancillary information about online research, which has gained popularity because it is faster, cheaper, and easier than paper surveys or telephone interviews and is appealing to respondents because it is convenient and anonymous. Online research also enables the researcher to present simulations and use multimedia as well as asking text-based questions, so there are a wide range of possibilities. However, it's noted that the Internet cannot produce a true random sample (as the demographics of internet users are currently skewed to certain education and income levels); respondents are believed to be more prone to providing false response in this channel than in others (especially when incentives are provided); and it's more difficult to effectively segment the respondents. The author also notes that online surveys are "everywhere," implying that respondents may become inured to them if they continue to be so heavily used.
QUALITATIVE RESEARCH
While quantitative research provides statistical information that is engaging and compelling for numbers-oriented decision-makers, it is by nature limited to the kinds of information that can be quantified and it is limited to producing the outcomes that were expected in advance, even if this means the real information is prevented from being considered. Qualitative research, which involves asking more open-ended questions that identify information the researcher did not know or expect in advance.
Qualitative research is the only effective way to discover information that is presently unknown, It's recommended that, even if quantitative research is desired, it should be preceded by qualitative research (doing a focus group before a survey in order to determine what questions to ask, and what options to provide rather than going on prejudicial judgments) and it may also help after the fact, to get at some of the reasons behind answers collected.
One form of qualitative research is observation: by merely observing customers in a store, a researcher can discover which parts of the store a person visits, which products they look at, how much time they spend making decisions, etc., in a more accurate way than a survey can determine. This method (EN: called "ethnography") is common in the social sciences such as sociology and anthropology.
(EN: IMO, observation gets short shrift. The actual behavior of people trumps the way they claim to behave. Especially in the internet, behavior in a "store" is a lot easier to monitor, and in a lot greater detail, The explosion of sensor technology and social media have also made observation much easier to do, and even palatable to the subjects. This area deserves much more consideration than it is currently getting.)
Another form of qualitative research is focus groups, which is an arranged discussion with test subjects, in which information is gathered from a conversation among peers. Traditionally, a hands-off approach was suggested, but in recent years, the trend has changed to more direct involvement - getting representatives of the brand in the room to directly interact with test subjects - and sources claim that an interactive discussion produces better results (a more lively discussion, more flexibility, etc.) and are more compelling to the decision-makers.
(EN: while I'm interjecting opinion, focus groups can be valuable, but the results are often heavily skewed by social factors such as complaisance, groupthink, and posturing. I can't' dismiss them entirely, but have had very bad results come of focus groups on more than one occasion.)
A "depth interview" is another method of qualitative study, which can best be described as a one-person focus group: an interviewer interacts with an individual test subject to draw out information, letting the subject do most of the talking, and prompting for details where they seem to be needed. This enables the researcher to get beyond simple yes/no answers to probe the reasons behind them.
(EN: This technique requires more cost and time than focus groups, and produces marginally better results. The social factors of the group dynamic are removed, but there remains the social dynamic of investigator-subject, which leads to misrepresentation, prevarication, speculation, and other behaviors that produce unreliable results.)
RELATIONSHIP RESEARCH
In recognition of the growing significance of brand relationships, a number of metrics specific to relationships have been developed in order to measure the success of a brand in building relationships with customers, and comparing these measures over time and with competitors.
One source (Net Promoter) monitors the frequency with which brands are mentioned and the degree to which these mentions are positive or negative. Another (Satmetrix) examines concepts that are related to the mentions of the brand. Another (BrandAsset) positions brands against their competitors in various dimensions to determine the relative strength.
(EN: this seems a bit random and disjointed, and I expect that is due to the novelty of relationship marketing. My sense is it will be some time before a standard and reliable set of metrics is derived - and even then, I'm not certain that a single set of metrics is applicable to all brands and all customers - so my sense is that this is idea fodder, but not a clear indication of what metrics are of the greatest importance.)