In an e-commerce environment, personalisation refers to technologies that enable one-to-one communication between the vendor and the buyer (or group of buyers).
Used properly, it can add great value to the vendor's customer relationship management (CRM) processes. Effective personalisation produces measurable outcomes, enhancing the users' experience by remembering their previously registered details, thus saving them time, or by tailoring more appropriate content to the users based on their registered preferences or past behaviour.
Day to day, we often see personalisation being used in Web pages and e-mails. When we receive a newsletter from a travel company with our own name in the e-mail, that is personalisation. When you see a listing of bargain Shiraz on a wine site, while your colleague instead sees a listing of Riesling in the same area, that is also personalisation.
Personalising Web and e-mail content is vital in developing and maintaining all facets of customer experience. The knowledge gained by the vendor over the course of relationship with the customers should be fed back into the system to add more value to the customer experience.
Levels of personalisation
The depth of personalisation can be categorised into four levels:
- Simple: this form of personalisation is basic in practice. It could be as simple as merging the customer's first name in the To field in an e-mail, or including the product name of choice in the Subject field. Because of its simplicity, it is the most popular form of personalisation around.
- Preferential: this form of personalisation involves tailoring content to specific user preferences, which have been explicitly registered by the user. In the wine listing example above, you may have previously ticked the -I prefer Shiraz" checkbox on your membership form.
- Behavioural: this form of personalisation is similar to the preferential form in its tailoring of content. However, the data collection method is different. User preferences in this case have not been explicitly selected by the user. Instead, they have been implicitly captured through automated interpretation of their past navigation behaviour. These interpretations are educated guesses at best, so unless treated with care, they sometimes result in the wrong content being served to the user. In the same wine listing example above, the reason that you see the Shiraz listing is probably because you have purchased bargain Shiraz from this Web site before!
- Segmented: this form of personalisation involves delivering information to customers based on certain demographic factors, such as areas or age groups. For example, there is no point trying to e-mail your customers who are 60-years-old or above about cheap tickets to Eminem concerts.
Technical prerequisites for personalisation
Personalisation can be very useful, but it comes with a cost component that is proportionate to the complexity of the personalisation engine. Some off-the-shelf personalisation software packages come with a high price tag that needs to be justified against needs, as personalisation requirements are very different among organisations.
- Basic personalisation: all that is required to support this form of personalisation is a membership database with limited customer information, including name, e-mail and gender.
- Preferential/Behavioural/Segmented personalisation: these forms of personalisation require a more sophisticated and extensible data model that can represent users and aggregations of users in flexible fashions. A user's membership should be allowed for multiple groups at similar or different levels of abstraction, as opposed to being limited to membership in a single group with cascading inheritance of properties from other more broadly defined parent groups.
In addition, the system also requires a workflow engine that can support creating relationships of arbitrary complexity among tasks and people, and a content management architecture that allows for flexible representation of individual pieces of content as objects. These objects should be flexibly tagged with iteratively generated metadata, and associated with each other in novel and changing ways.
More complex personalisation systems of this kind often utilise common personalisation devices such as cookies and session-ids. In most cases, these devices are used to identify and track the users as they navigate around the site.
Duc Do is a director of Commercial Interactive Media. He can be contacted on 03 9419 4900 or at mailto:duc.do@cim.com.au
