Personalization In A World Of Artificial Intelligence

V. Kumar and his colleagues examine what AI is doing to the world of marketing in their recent piece in the California Management Review. Central to their analysis is the idea of personalization which they distinguish from customization. Customization is when you give the consumer a bunch of choices. This puts a lot of work onto the consumer to configure exactly what they want. We have probably all had to do this with items of technology and had to work out how much memory etc… we need for our phones.

Personalization is different in that the consumer is less actively involved. “Personalization occurs when the firm decides, usually based on previously collected customer data, what marketing mix is suitable for the individual, whereas customization occurs when the customer proactively specifies one or more elements of his or her marketing mix” (Kumar, Rajan, Venkatesan and Lecinski, 2019, page 136).

It is clear that both have advantages. Where the customer knows what they want and doesn’t mind putting in a little effort customization is perfect. Personalization can fail if the firm doesn’t know the customer well or doesn’t have the customer’s best interests at heart. That said personalization can often be a simple solution that gets time and attention constrained consumers to their ideal choice. One downside is obvious. The firm probably has to collect a lot of information on you to be able to present you with your perfect choice. We, in effect, trade off privacy for better service. Some may be happy with that, others less so. (Obviously if the AI knows the consumer well enough to know the consumer does not want to trade privacy for personalization the consumer can be allowed to customize, rather than have their offers personalized, if they so wish. Sadly the AI will need to collect a lot of data on the consumer to know that the AI should not collect it which won’t cheer everyone up).

The authors identify three waves in the journey towards personalization. The first wave being when targeting was largely on broadcast media. If you watched certain shows you were offered certain products. Wave 2 involved more information and more sophisticated if-then type rules. This flourished with the adoption and extension of the internet. We are now in wave 3. This involves AI narrowing the vast amount of options available to the consumer. A lot of this narrowing comes early in the process, what the consumer looks at will impact what they are offered. The black box element of AI means it won’t always be apparent to the marketers why any particular consumer was shown any particular offering. In this world the curation of offerings becomes a vital skill for marketers and their AI agents. It is an interesting world that you might need more than just your natural intelligence to understand.

One of the ideas they digress to is the idea of the uncanny valley, where things are close to human but just a bit off. We can see this with movie special effects which are often great but just not quite human. Being in the uncanny valley elicits “uncanny, strangely familiar, feelings of eeriness and revulsion in observers” (Kumar, Rajan, Venkatesan and Lecinski, 2019, page 145). Our technological progress may get us into the uncanny value more. I’m not sure what to do about this but creeping out your customers probably isn’t a good idea.

Read: V. Kumar, Bharath Rajan, Rajkumar Venkatesan, and Jim Lecinski (2019) California Management Review 61 (4), pages 135-155