Email musings: Rejecting Personalization

Personalization can do a lot to help an email campaign. Split tests regularly show that personalization options hugely increase email conversion rates. However, lately I’ve been wondering about how we can use group targeting instead of personalization as a way to engage readership. Specifically, I’ve been wondering if, perhaps, we could further engage email audiences by adding them to targeted groups, and then addressing the groups directly, instead of the individuals. The reasoning has to do with how individuals interact in groups, specifically how groups provide identity, support, and guidance for their members. Would grouping people together engage them readily with the subject?

Personalization without the individual

What brought this question up was a study by Henri Tajfel titled Social categorization and intergroup behavior (1971). In this study researchers took a group of boys and divided them into groups, then asked them to distribute a resource between the members of both groups.

Tajfel hypothesized that people would form ingroups and outgroups innately, without even the need for face to face contact between members, or even any value to the formation of the group. To test this he tried a series of experiments.

In the first experiment they asked the boys to estimate a number of dots, then split into two groups. One group was told that they were split by the highest and lowest scores. The other was told that they were divided by the most and least accurate. In truth, they were randomly selected.

The boys were then told that they would be assigning monetary rewards to the others, though never to themselves. They wouldn’t know who the other children were, only their scores and what group they belonged to.

The result of the test was that, predictably, the boys favoured their own group over others, distributing more money to those in their group. The  better part was that the results held true as they made the examples ever more minimal, reducing what the boys served to gain from favouring their own group. It made no difference. no matter whether they had anything to gain from it they favoured their own.

Tajfel et al. repeated the test, further isolating the effects of the ingroup. Again, members rewarded those who were part of their own group. Tajgel concluded that:

The main findings in all three experiments is clear; in a situation devoid of the usual trappings of ingroup membership and of all the vagaries of interacting with an outgroup, the [subjects] still act in terms of their ingroup membership and of an intergroup categorization. Their actions are unambiguously directed at favouring the members of their ingroup against the members of the outgroup.

This test has been repeated with successively more minimal systems, and continued to show that we do engage in ingrouping readily. The reason for this, as Tajfel would come to argue, is that we construct our identities around the groups we belong to. 

Applying this to email marketing, appealing to a group may be more engaging than to an individual. By maintaining a structured group that you deal with as a set, with a clearly defined separation from other groups, perhaps we could create a more engaging experience by interconnecting the end user experience with individual user identity at a subjective level? I think this may happen more frequently than we think. Take, for instance,  web forums where different sub-forums create ingrouping based on interest, producing rivalries and other group to group activities, which in turn increase the self identification of a user with their community.

The difference here is, of course, that forums allow peer to peer and multi-directional communication while email tends to be pretty explicitly one way. Could we apply this same affect without the P2P communication though? I don’t know.

Stuff to try:

1. Segmenting by interest (definition of group)

2. Including contribution from “the group”. (Identification of group)

3. Giving each group information about other groups, and what they’re doing (distinction of groups)

4. Selectively including group content (moderation of group/user interaction?)