What is personalisation?

Personalisation makes use of user behaviour data and machine learning to deliver content and functionality that matches specific user needs or interests.Personalisation is implicit —It is controlled by the system and does not require any conscious input and effort from the user.

In a previous post, I wrote about how people rarely read word for word online. Instead, they scan to look for keywords that point them to what they are looking for. Personalisation profiles the user and adjusts the interface according to the specific needs and interests of the user. This leads to an improved user experience and facilitates them in finding what they need and completing their tasks quickly.

There are different levels of personalisation:

• Individualised personalisation: Users are profiled individually and content is tailored to each individual user. Examples include Netflix’s homepage recommendations and Spotify’s tailored playlist.
• Role-based personalisation: Users are grouped together according to a set of pre-determined and similar characteristics. An example would be a company’s internal platform where the interface and access rights are tailored to one’s job role.
• Location-based personalisation: An example is the Coca Cola website with personalised interface and call to action for each market.

Coca Cola personalises its website UI for each market.

Customisation

Businesses are using both implicit and explicit means to understand user preferences. While personalisation is an implicit method, customisation is explicit — it is controlled by users and they are asked to make choices to tailor their experience to their own preference. It enhances user experience as the interaction is controlled by users.

Customisation works best under the assumption that users know their goals and needs best. Compared to personalisation which is done by the system, users have to take the time to configure the site in a way that is optimal for them.

Introduced in 1999, Nike was one of the first brands to offer customisation options with “NikeiD” (now Nike by You) allowing customers to design their own footwear.Pinterest enabling users to customise the type of content they want to see on their feed.

Balancing personalisation with data privacy

While traditional personalisation considers a user’s name, location, purchase and browsing history, hyper-personalisation goes one step further by using real-time data to hone in on what the user wants or needs. This is enabled by the advancement of artificial intelligence and machine learning.

Hyper-personalisation also means collecting more data from users, with or without the consent. Even when the user accepts to giving their data, there is still a loss of privacy.

Research on online behaviour has revealed inconsistencies between user attitudes and their actual behaviour. While users claim to be very concerned about their privacy, they take little effort to protect their data — this phenomenon is known as the privacy paradox.

Researchers seeking to explain this phenomenon through a systematic literature review determined that a user’s willingness to divulge privacy information is generally driven by two considerations:

• Risk-benefit evaluation: Although users are aware of the risks, the benefits outweigh the threat of privacy loss.
• Biased risk assessment: Due to the need for immediate gratification, faster and on-the-go decision making, users are biased in their risk assessment. Users may also be constrained by external factors such as low transparency, unfriendly design or hostile consumer privacy policy with all-or-nothing permission usage.

Giving users transparency over how the data is used and control over what data is shared helps to build trust. Most users are willing to share some of their data, especially when there are perceived benefits.

In an Accenture study, majority of customers are willing to share their data in exchange for perceived benefits.

There is no denying that personalisation has brought real value to both users and companies. As more companies begin to adopt hyper-personalisation which involves the collection of more data, they will need to think about whether the privacy loss on the part of user is necessary to reach their business objectives. How much and what kind of data is required? It needs to be considered and communicated in a transparent manner. This is when privacy by design comes in — privacy should be considered from the beginning of a product cycle, not as an afterthought.

Further References: