Web Personalization Prerequisites: Identifying Audience Segments

Web Personalization Prerequisites: Identifying Audience Segments

default avatar
Thought byDave Sawyer
April 08, 2016
Audience Segments

With real-time personalization, segments are typically based on criteria that can either be automatically detected, or determined from previously gathered user data.

In the introductory post in this blog series, I described how real-time personalization can enable businesses to tailor content to the needs of specific audience groups. But in order to tailor content, you first need to define how you will identify the visitors that will receive personalized experiences. In a personalization strategy, this is accomplished through visitor segmentation.


What is Segmentation?

Segmentation refers to the process of dividing your audience into distinct groups based on specific criteria, contexts, and/or conditions. With real-time personalization, segments are typically based on criteria that can either be automatically detected, or determined from previously gathered user data. There are many different types of criteria that can be used to segment your audience, but most fall into one of two groups: implicit and explicit data.

Implicit Data

Implicit data is information that is implied or assumed. It provides an indication of a user’s intentions or needs, but is not plainly expressed by the user. Implicit data can be used to test a hypothesis, make a content recommendation, or to inform a content experiment such as an A/B test. Deriving a person’s interests based on the pages they’ve visited is an example of using implicit data.

Explicit Data

Explicit data is clear and specific, leaving no room for confusion or doubt. It can consist of automatically detected visitor attributes, or it can be data that a user has chosen to provide, such as their personal information or preferences. Tailoring an experience to the visitor’s age, gender, location, or type of device are examples of personalization based on explicit data.

Rules-based vs Predictive Personalization

Personalization based on explicit data is often referred to as rules-based personalization. Content is tailored when certain rules or conditions are met. However, many personalization products and enterprise CMS platforms provide effective ways to leverage implicit data, known as predictive personalization. Wikipedia defines predictive personalization as, “the ability to predict customer behavior, needs or wants - and tailor offers and communications very precisely”. With predictive personalization, the visitor’s behavior is analyzed and content is displayed in real-time based on that behavior. Predictive personalization “chooses” the most relevant content for the visitor based on the best performing content variation – a landing page variation that has resulted in the most conversions for the visitor’s segment, for example.


Types of Segmentation Criteria 

The most effective personalization strategies combine both implicit and explicit data to optimize the user experience. But what specific types of criteria can be used to personalize content? Let’s walk through the most common categories:

Demographic Criteria 

Demographics refer to the qualities or attributes of a specific group of people. Demographic criteria for personalization is typically explicit in that it is most often personal - provided by the visitor through an account registration, sign-up, or form fill – although it can originate from other sources. Demographic criteria may include:

  • Age or Birth Date
  • Gender
  • Income or Salary Level
  • Job Title or Occupation
  • Professional or Employment Status

Examples of Demographic Personalization:

  • A travel company who targets promotions for European tour packages to seniors.
  • An online retailer who recommendations dresses instead of slacks, based on the gender of the visitor.

Geographic Criteria

Geographic criteria is a type of demographic data that can be used to serve the needs of customers in a particular geographical region. This type of criteria may include:

  • Visitor’s specific location
  • Visitor’s general region – e.g. State/Province or Country
  • Current local time or weather

Examples of Geographic Personalization

  • A media outlet who automatically displays news coverage that is relevant to the local region or country of the visitor.
  • An online retailer who dynamically determines the local weather conditions of the visitor, and displays personalized product recommendations – such as a sale on umbrellas for visitors in locations where it is raining, and a sale on sunscreen for visitors in hot, sunny weather.

Visitor Behavior 

The behavior of the visitor can be a very effective indicator of their interests. Visitor behavior is most often composed of automatically detected criteria and implicit data about the current or past browsing sessions. Marketers and site developers can define distinct interactions that will cause a visitor to be included in a segment. Examples include:

  • The topic of content the visitor has viewed the most
  • The specific content the visitor has viewed most recently
  • The click path or order in which a user has viewed content
  • New vs. returning visitors
  • Past site downloads
  • Recent conversions or purchases

Examples of Behavior-based Personalization

  • A consumer healthcare site displays lists of recommended articles about a specific category of disease or ailment based on other articles the visitor has read recently.
  • A B2B company displays topical marketing messages on a website based on whitepapers that a visitor has downloaded previously.

Session Attributes

Personalization can also be based on explicit attributes of the browsing session and other visitor metadata. Examples include:

  • Browser or Device type
  • Source or Referral type
  • Authenticated vs Anonymous

Examples of Personalization based on Session Attributes

  • A mobile app company automatically detects that a visitor is connecting from a Samsung smartphone and delivers personalized promotions for Android apps that are compatible with the device they are using.
  • The home page of a SaaS product displays personalized messaging depending on whether the visitor has arrived from a direct link, a search engine marketing campaign, a banner ad, an organic search engine result, or a partner/affiliate site.

User Profile Criteria

User profile criteria is based on previously gathered user data. This may include account details from a CMS like Drupal, data from a customer record in a CRM like Salesforce, or personal data from a social platform like Facebook. In addition to demographic data, user profile criteria may include:

  • Interests or preferences
  • Customer type or account history
  • Subscriber information

Examples of Personalization based on User Profiles

  • A newspaper website displays personalized lists of articles based on topics that the subscriber has indicated they are interested in.
  • A customer at a "Silver" level of service is presented with a promotion to upgrade to "Gold".
  • A service displays a promotion for a customer to renew their membership based on account history data that indicates their membership is near expiration – while other visitors see promotions to become a member for the first time.

Segments vs. Personas

Now that we've described various segmentation criteria, you might be wondering: What's the difference between a segment and a persona? A persona is typically a marketing classification of an audience group defined by demographic or psychographic attributes. Sounds very much like a segment, right? These terms refer to very similar ideas and are often confused. Let's describe the difference in terms of a personalization strategy.

Consider a university’s website, which needs to serve a number of distinct audience groups, such as prospective students, current students, faculty and staff, alumni, and parents. These designations are often referred to as personas as they describe distinct types of people who share common traits or needs. A prospective student might be interested in a campus tour or dining options, while a parent might be looking at tuition information. Personas are useful tools for designing and organizing the site content by user needs and are essential to information architecture.

Segments vs Personas

By comparison, visitor segments tend to be more specific and don't only involve demographic criteria. A university might have a section of their site designed for the prospective student persona, but visitors who view content within that section can also fall into one or more other segments. For example, prospective students visiting the university site from other countries might fall into an “International prospective students” segment. Or prospective students who have visited the site previously and browsed graduate program information and might fall into a “Repeat visitor, graduate student prospect” segment.

In short, personas tend to describe the needs of distinct types of people, while segments provide a granular way to target content to subsets of visitors.

In Drupal personalization solutions, personas can be thought of as taxonomy terms or tags that content editors can use to classify which broad audience group a piece of content is intended for. Whereas segments tend to be defined based on additional attributes of the session or visitor history.



Segmentation is the core of any personalization strategy. It allows you to gain valuable insights into how distinct subsets of your audience respond to different messages, offers, or recommendations. But personalization is just one component of a digital marketing strategy and can only be effective if a business has developed a clear understanding of the needs, wants, and demands of its current and potential customers. Segmentation is made most effective when informed by data and market research.

Smart personalization strategies involve segmentation-based combinations of criteria to best organize visitors according to marketing objectives. But given the amount of available data or number of ways an audience can be segmented, it can be challenging – even overwhelming – to know where to focus. Experimentation is needed in order to identify the right level of segmentation and to determine which personalized experiences result in the most beneficial outcomes.

To learn more about personalization solutions, download my eBook, The Basics of Real-Time Personalization. It has all the content shared in the blog, and more.

FFW would love to learn more about your experiences or challenges. Feel free to leave comments or questions below, or reach out to me on Twitter: @cmsdave.

Attending DrupalCon New Orleans? Check out Dave’s session Web Personalization for Drupal Sites: Your Roadmap to Get Started.