Intelligence benefits abound
Furthermore, systems like these provide more intelligence-gathering capabilities than other methods, such as shopping basket analysis, expert systems, content analysis and collaborative filtering algorithms. They do this by watching not only what customers select, but where they came from, the path they followed and the time spent looking at each item.
These systems start by simply observing activity during a training period that lasts from a few days to several weeks, depending on site traffic. During training, the system builds a history of how previous customers behaved and thus learns which items are likely to be selected under which conditions. This knowledge is then converted into scoring rules that generate multi-value profiles for each visitor and each product. Profiles and scoring rules are updated continuously as new behavior occurs, providing the most current and real-time insight into customer trends and preferences. What's more, all of this occurs anonymously.
By actually profiling the entire individual clickstream behavioral pattern, these self-learning solutions focus on the person-to-content affinities, and are not just limited to content-to-content affinity modeling that are often represented by "people who bought this item also bought these items." Some of the benefits include:
- accelerating complicated searches for which the user may not know how to define keywords (e.g., technical support site for a non-technical user)
- automatically creating document similarity map without metadata tagging or domain-specific taxonomy
- building itself from user behavior; no document parsing required
- document matching that is language-independent. Translations of the same document will have matching profiles.
Self-optimizing personalization solutions are completely data-driven and are therefore able to process large and diverse transaction histories; even across product categories, cultures and languages. Furthermore, systems like these do not rely upon any knowledge of the content other than the user's interactions with it. This greatly simplifies integration and unified content profiles across all types, from text to multi-media. These SaaS platforms can also scale to large user bases and catalogs while providing acceptable runtime performance.
Apply to all types of environments
Because of their automated profiling, content neutrality and adaptive content indexing, these types of systems can be embedded into almost any environment that can capture online behavior, including ecommerce, search, content, email, mobile and streaming media. This provides additional functionality to the single subject applications and can be used in a variety of methods such as:
- personalized product recommendations on ecommerce sites
- personalized and targeted emails
- personalized content recommendations
- community radio: for use in a group setting, a custom broadband radio station could optimize its playlist for the subjects present.
- movie guide: the affinity group would be used as a "movie night" feature, whereby the guide would suggest the best movie for two or more subjects
- gift registry: gift applications could register subjects and groups of subjects that could be used to filter gift suggestions, including restricting suggestions to objects that are compatible with both the giver and the receiver
These types of systems can also be combined with other enterprise data to produce even broader predictive models of customer behavior. Doing so extends their benefits to off-line campaigns, such as direct mail, telemarketing, media advertising and customer loyalty marketing campaigns.
Customer behaviors are notoriously complex. While traditional analytics package provide great insight into their behaviors and needs, new technologies along side existing systems can offer companies a more complete picture of what their visitors are thinking. Implementing them can mean incremental revenue, cost savings from automation and an enhanced visitor experience.
<< Previous page
Meyar Sheik is CEO and co-founder of Certona Corporation.
