Marketing Research: Market Segmentation

Probably the single most important aspect of marketing a company’s products and services is proper market segmentation then positioning your brands to laser-focus on the wants, needs, attitudes, and desires of your chosen target market. Yet, many businesses failed to update their segmentation strategy to accommodate both modern audiences and marketing practices.

In the pre-digital world, when gathering the information needed for proper segmentation was costly and time-consuming, brands used fairly anemic segments. Even some of the world’s leading companies, like Proctor and Gamble, use relatively undefined segments such as “women; 25-45, who care about their skin and are concerned with aging”.  Not only is this somewhat generic description too broad,  it doesn’t provide guidance regarding how to reach this target and what messages are most likely to drive people in the segment toward actions, like buying your products.  So how should a company define its segments?

Create market segments

Traditional segmentation

First, segments are normally defined based on some combination of demographics (age, income, education, etc), geographics (country, area (ie NW, NE), concentration (ie. urban, rural, suburban), etc), and psychographics (including attitudes, usage, lifestyle, etc). Although psychographics offers the most insight into reaching a particular target market, too often firms use demographics and simple usage to define their market. Thus, while a company spends millions of dollars developing a new product, it may doom the product to failure by not understanding how to reach the product’s intended market. Besides, knowing more about your target market helps you develop products and services consumers will find particularly valuable, which helps ensure their success and the success of the firm.

Building these segments used data generated through surveys, sometimes combined with the company’s own sales data. Gathering and analyzing data, often using tools like cluster analysis, which is somewhat theoretical, took time, money, and analytical skills. Below, see an example of cluster analysis resulting in 2 groups based on their responses. Implementation of cluster analysis requires software such as Mathworks (shown below), SPSS, SAS, or R.

Personas

Increasingly, companies build more nuanced target markets, called personas. Personas, which represent an archetypical consumer in your market, include deeper insights about the consumers that make up a group or segment, as well as critically important lifestyle issues related to the usage of the product, key decision variables used to make choices, and where you’re likely to reach the consumer.

Here’s a great example of a persona:

Building personas might involve data collected through surveys or it might happen through deep dives into the company sales data, rich data available on social platforms, website data through Google Analytics, or search data purchased from suppliers like Google. An important note about data available through third parties; personally identifiable data isn’t available, only aggregate data. For some great insights on using social media to craft market segments, check out this post.

Crafting and using personas

1. Segmenting and targeting

So, the first step is to segment the market. Qualities of a good market segment include, according to HBR:

• Large enough to warrant the effort and cost
• Stable over time
• Segments are distinct from other groups based on needs, attitudes, or usage
• Accessible through some media
• Actionable based on your product market

Once you’ve identified the segments naturally occurring in your market, you determine which segment or segments fit your products best or which fit the product you envision creating. That’s targeting. Before you go any further, you must target your most profitable segments. Ignoring an influential segment at this stage lowers your marketing returns.

2. Build personas

Once you have your target markets, the next step in building personas within the population by fleshing out your insights regarding the group with as much information as possible. Commonly, because you’re creating an image of a real person in your target market, personas are given names indicative of their key differences from other groups, such as Frugal Fred to personalize your fictional consumer. Others prefer to give personas regular names.

In building personas, consider the following types of information:

• Demographics
• Geographics
• Goals
• Frustrations, remember people buy solutions, not products
• Motivations
• Personality
• Media choices
• Technology
• Skills

Personas often include a picture of an idealized member of the group represented by the persona. A brand should develop a persona for each consumer segment in the marketplace.

Building personas requires that companies find out as much as possible about folks making up the persona. Companies conduct segmentation either qualitatively (words) or quantitatively (numbers).  I usually recommend clients use both methods of market segmentation as part of a comprehensive marketing research program. Moreover, I usually recommend that segmentation needs to be re-done periodically, as consumers constantly change.

Using personas

You must support market segmentation through anticipated improvements in your market performance otherwise building personas is a waste of time, money, and effort.

An important way brands use personas is through positioning. Positioning refers to the task of aligning your messaging with factors uncovered about the persona. For instance, Frugal Fred wants to hear about how affordable your product is rather than how fast. Similarly, showing how your product replaces several other products points to the low cost of buying your product versus needing several products.

Another way personas help your marketing efforts is by guiding new product development. Use persona, especially the part about frustrations, to identify new products your market wants.

Modifying your UX (user experience) design based on preferences reflected by your persona also provides a boost in performance.

Qualitative vs quantitative market segmentation

Quantitative market segmentation

You can survey the population for market segmentation, then use quantitative means such as cluster analysis, which I discussed earlier

I did this on a project looking at how consumers respond to needs for online privacy. I found there were 4 groups: 1) those who were concerned about privacy, but resigned to losing it online in return for the benefits they received; 2)  those who were concerned and refrained from engaging in most online activities; 3) those who were unconcerned about online privacy, and 4) those who felt that efforts to invade their privacy online would be dealt with by the police and that would deter anyone from invading their privacy (granted this is a relatively naive understanding of the internet).  Knowing how your particular set of online consumers feel about online privacy would help a firm determine elements of both online and offline communication as well as determine how they might craft the UX.

For instance, consumers in groups 1, 3, and 4 benefit from user registration as it makes their lives easier by remembering credit card information and address. Registration also allows users to review prior purchases and track shipments. Users in group 2 prefer guest access to ordering and payments through 3rd parties like PayPal, which reduces invasion of their privacy. Knowing you don’t have any group 2 users means you only need one type of UX, meaning you save some expense and aggravation.  Users in group 1 likely respond well to assurances of privacy such as BBBOnline, while the other groups don’t really care.

Qualitative market segmentation

Qualitative techniques used for building personas involve using interviews or other qualitative data, then filtering the data to develop themes that separate groups from each other.

Once you have qualitative data, interpretation is challenging. I usually start out thematically to identify patterns among responses. I can then identify participants who fit in the pattern or who fit into one of the various alternative patterns emerging from the data. Finally, I determine the demographic and geographic categories defining these participants, as well as deeper insights about attitudes, beliefs, and needs.

For instance, I was working on a logistics project for a client and found 3 groups of online shoppers: those who liked to shop online for gifts to avoid the hassle of shopping and shipping the items, those who lived in areas where they couldn’t shop in brick and mortar stores so they used online as an alternative, and those who liked to buy products online because it is an information-rich environment. Each of these groups might be targeted using different strategies including web design and off-line promotions.

Market segmentation in action

Qualitative data collection

Qualitative techniques for market segmentation include focus groups, interviews, and observations. After identifying market segments, you need to ensure you’re including representative folks from each persona in your data collection. Often, in focus groups, it’s best to only include 1 persona-type in the focus group so folks feel comfortable and that their opinions are respected. You’ll need a focus group for each segment in this situation.

All of these qualitative research methods feature a number of variations and each has its own benefits and drawbacks. Whichever technique you use, be sure to ask very general questions and allow for a great deal of introspection by participants. Projective techniques and storytelling by consumers are particularly effective ways to get information from participants. Unfortunately, clients often see these as a waste of time.

Too often, clients use their own beliefs about the product to guide them into collecting data that supports their beliefs. For instance, I was once with a client behind one-way glass as the moderator conducted a taste test of the client’s product. The focus group participants hated the product and the client almost went through the glass trying to convince them the product was good. The client failed to understand that their target audience might have different opinions than they did. Needless to say, we didn’t get anything valuable from the participants and the client lost a valuable opportunity to understand how consumers evaluated products like the one he produced — what criterion they used to determine “this is a good product”.

Another client argued against adding certain questions as irrelevant only to find out after the project was completed that they missed an opportunity to collect valuable information because they didn’t initially see why the answers mattered.

Quantitative data collection

Quantitative techniques are more straightforward and more clients feel comfortable with numeric data — almost like numbers give the information more relevance or more reality, which is definitely not the case.

Certainly, the process of using numeric data to create segments is more straightforward and many software programs, like SPSS and even Excel, accomplish this task easily with a little training. The use of web surveys, such as SurveyMonkey and Qualtrics, also offers an attractive option due to their low cost.

Unfortunately, ensuring the questionnaire asks the right questions, writing questions that aren’t leading or misleading, and that produce responses that are both mutually exclusive and collective exhaustive is very challenging. It is just this type of mistake that led to the fatal introduction of “New Coke”.  Coke, based on survey data implying cola drinkers preferred a sweeter drink, developed a new formula and immediately discontinued the manufacture of the old formula. It turned out that consumers bought Coke for a variety of reasons, including nostalgia and other emotional reasons, feelings that didn’t transfer to New Coke, causing outrage and hoarding of the remaining cases of Old Coke until Coke relented and returned to the old formula.

That’s why I recommend doing segmentation sequentially, with qualitative preceding quantitative data collection. Qualitative data would have uncovered the emotions tied to Coke that would have informed the questions asked in the quantitative study.

Next week I’ll introduce the topic of segments a firm should NOT target. In the meantime, feel free to comment on my blog, suggest new topics, ask questions, or share your experiences.  Be sure to bookmark posts you find valuable and share them with your friends and colleagues via Facebook or other social media.  For your convenience, icons appear at the bottom of this blog to help you share this information.

Need marketing help to support business growth?

We welcome the opportunity to show you how we can make your marketing SIZZLE with our data-driven, results-oriented marketing strategies.  Sign up for our FREE newsletter, get our FREE guide to creating an awesome website, or contact us for more information on hiring us.

Hausman and Associates, the publisher of MKT Maven, is a full-service marketing agency operating at the intersection of marketing and digital media. Check out our full range of services.