Jessica, a “do-it-yourselfer,” prefers to tackle her tax preparation on her own. Kim, on the other hand, is more of a “do-it-for-me” and wants the help of her tax preparation company. Both of these women fall into marketing segments created by H&R Block that help the company target more precise media and content with its digital advertising. When you set out to create audience segments, first define your brand’s goals and specify customer attributes before diving into statistical analysis of your data.
What is a segment?
A segment is a group of people or other entities, like companies or stores, which share quantifiable attributes that matter to your business. Most marketers do not use a single approach to segmentation for all their efforts, but rather apply a number of different approaches for different business reasons, according to Martin Kihn, research director, Gartner for Marketing Leaders.
For example, the retailer CVS uses five different schemas:
- Psychographic — Lifestyles and attitudes toward managing family health, for brand and other awareness campaigns
- Behavioral — Frequency of shopping by category and engagement with marketing tactics, for couponing and promotions
- Benefits — Functional or emotional benefits (e.g., whiter teeth) for product advertising
- Geography — Region and density for product mix and pricing
- Occasion — Special occasions (e.g., birthdays) or purchase intervals for reminders and recognition
Define your goal
While you can create as many segments as are statistically different and useful, good segmentation includes only those that are large enough to have an impact and that are stable, addressable and unique enough to reward different treatment. Begin by defining your goal across a framework for positioning internal/external (customer or noncustomers) and strategic/tactical goals.
Internal — Focuses on customers or others who have some engagement with your products — for example, visitors to your website. Primary uses include improving customer experience and retention, upselling and cross-selling, and gaining a greater share of wallet from competitors. Data sources include CRM, web analytics, point-of-sale systems, etc.
External — Treats people who are not existing customers, generally targeted for prospecting or growth. The outcome is often to find new prospects and leads for growth, or for new products and launches. Lacking internal (e.g., first-party) data, external segmentations rely on syndicated and other third-party data, media vendors, market research, government and other outside sources of information.
Tactical — Tactical segmentations have a goal of determining which types and groups of people are more or less likely to buy product X or service Y. For example, Kraft segmented its customer base by whether people searched for recipes and were interested in Easter for a promotion around its Easter Bunny Cake recipe.
Strategic — Strategic segmentations are more interested in exploration and discovery among groups of customers or prospects and often used by marketers who want to discover new insights for targeting, messaging or product development. This was the goal for tax preparation company H&R Block when it performed clustering analysis that identified three broad groups of customers: “Do-it-yourselfer,” “do-it-for-me,” and a previously unknown hybrid group it called “do-it-with-me,” who preferred some help but not full service.
Read related article: What’s in a Name? Creating Personas for Digital Marketing.
Develop attributes for data collection
At the heart of any segmentation exercise, think of attributes as quantifiable characteristics that can be aligned to a customer, prospect or market.
Attributes come in three basic types. They describe what people:
- Are — Persistent personal attributes such as age, gender, household income, language group and marital status. Less persistent attributes include transition states such as moving, pregnancy, wedding and bankruptcy.
- Do — Behaviors that can be observed, such as products bought, media or devices used, websites visited, content consumed. This includes where they are.
- Think — Attitudes and values that are sometimes explicit, such as when a person calls or tweets a complaint, but often must be inferred (e.g., political party based on zip code).
After defining data collection sources and integration, build segments with a needs/value or clustering exercise. If you want to determine, for example, the key factors, or needs that make a person a profitable customer, a decision tree can help spot which combinations of attributes have the greatest impact on value. Clustering methods are used to find groups of people who share similar attributes and are reasonably distinct from other groups. Both types of results give you powerful insights for product development, pricing, targeting, messaging and measurement.
Next, use personas to give depth and expression to segments that can seem cold, inhuman and incomplete. Here, a collection of attributes such as women over thirty who visit cooking sites becomes the persona, “Tammy,” a proud mother of two young children who entertains with flair.