Personalization: A Proposed Process Model

This is a proposed model for the implementation, staffing, and ongoing management of variable web content. This model is platform independent – it simply assumes content is served from a hypothetical platform capable of providing the necessary functionality.


Our primary goal is to increase the raw conversion rate of the website by displaying visitor-specific content that we believe is more likely to provoke a conversion than the default content. In the absence of a conversion, we hope to simply improve the visitor’s perception of our brand.


The following principles describe, in broad terms, our methodology and objectives.

  • We will identify specific audiences using three main categories of criteria:

    1. Session: Factors unrelated to visitor behavior, but specific to their current session (e.g. - their location, their incoming search terms, technical parameters of their web request, etc.)
    2. Global: Factors universal to all visitors and sessions (e.g.- such as time of day, available content, etc.)
    3. Behavior: The accumulated content that a visitor has consumed on the website (usually in the current session, but optionally including previous sessions as well), with the assumption that their selection of this content contributes to their audience identification
  • Visitors will “inhabit” specific audience segments due to their fulfillment of criteria derived from the above factors. This inhabitation will either happen at the instant of session initiation, or at the instant of a subsequent page request during that session, depending on defined criteria.

  • Audience inhabitation will be evaluated on every page request. Specific audience inhabitation may be used to vary content displayed in selected areas of the requested page in order to display content that:

    1. Reflects positively on the organization’s relevance and ability to fulfill the needs of an audience which the visitor inhabits
    2. Provides a conversion opportunity (e.g. – a lead generation form) using specific (1) verbiage, and (2) format that is optimized to provoke a response from an audience which the visitor inhabits
  • We will log link activation on content rendered specifically due to audience inhabitation (put another way, content that would otherwise have not been displayed, but appeared solely due to audience inhabitation)

Initial Implementation Process

  1. Identify and prioritize audiences, based on their relative value to the organization.

  2. For each audience, determine the inhabitation criteria.

    • Eliminate audiences that cannot be reliably segmented by the CMS.
  3. For all content, determine if it will provide scoring towards audience inhabitation and what weight that score will carry

  4. Determine general variable content locations within the website templates

  5. Determine what content variations will be offered to different audiences.

    • Develop necessary content that does not exist.
  6. Determine what conversion opportunities will be offered to different audiences

  7. Map this (1) content, and (2) conversion opportunities to specific template locations

On-Demand Tasks

These tasks extend existing tasks.

  • New Content Created: When new content is requested, answer the following questions:

    1. Can/should multiple versions of this content exist to address different audiences?
    2. Which audience does this content address?
    3. How will this content be scored towards audience inhabitation?
  • New Template Designed: When a new content template is designed (either in isolation, or as part of a larger re-design), it should be reviewed for potential locations for variable content to be placed

  • New Form Created: New data intake forms should be reviewed and/or configured for correct audience scoring on submission and accurate CRM profile matching

Ongoing Tasks/Meetings

  • Continuing Audience Identification: Have new audiences appeared which need to be addressed? Are existing audiences still relevant?

  • Analytics Review: To what extent are audiences interacting with variable content placement?

  • Content Review: Has all content (specifically new content) been scored consistently and correctly?