Segmentation: Beyond Demographics

Posted by: Jennifer Valenzuela
  • March 8, 2015
  • 3

SupportTeamLast week, The New York Times published an article about Fidelity Investments’ decision to customize their customer service approach based on the gender of the investor. In practice, this means a woman calling for advice may receive more small talk initially. She may field questions concerning her long-term financial goals and the important relationships in her life. A man, on the other hand, will have his questions answered immediately without much “chit chat.” The call center representative will “cut right to the chase.” The article assures readers that the end result will be the same for both investors; given the same financial situation, both callers would receive the same advice after very different customer journeys.

Although the potential impact of this approach struck me immediately, I couldn’t help but think that perhaps Fidelity got it wrong. Certainly men and women communicate differently, that’s not new. But to instruct customer service representatives to change their scripts based solely on a caller’s gender leaves out a large part of the equation.

The latter part of the article went on to suggest that perhaps differences in need are driven less by gender, and more by other factors like socioeconomic status or life stage. Experience has taught me to consider much more than the obvious when segmenting a market. Segments become more robust and usable when they include things like attitudes, needs, behaviors and life stage, for example, in addition to basic demographics.

Gender is an easy place to start. However, if Fidelity knew a caller was female but also knew, based on past behaviors, she was likely to be part of a particular segment receptive to a high risk/high yield investments, they’d change their script. Suppose they knew, based on her segment profile, her tendency to want to “cut to the chase,” they’d change their script again. Fidelity’s current, overly simplified gender strategy may actually create unwanted, unintended effects, leaving some men desiring a more personal connection, and some women rolling their eyes wondering when the real consultation will start.

Because financial planning is both important and personal, truly connecting with clients and serving their individual needs changes the game. Consumers now move money from one advisor to another with the click of a mouse—no more awkward break-up speeches! Consequently, fostering a true, personal connection is key to keeping clients: know more about them, speak to them in their own language, leave them wondering, “How do they know me so well?” These are steps in the right direction.

My advice to Fidelity? Build a segmentation tool that more accurately helps customer service representatives identify who they’re talking to, and then provide them corresponding script custom-made to each group. That way, whether a dude, a dame or some segment in-between picks up the phone, you’ll have a trustworthy system that earns their fidelity.

3 Comments
  • Wendy Jackson
    March 9, 2015
    Agree, Jen! Demos only help predict attitudes where developmental psychology is in play. The other issue here is that it sounds like they are basing their conclusion about gender differentiation on findings about past customers. That's a dangerous game - duplicate the past only when you want the result of duplicating the past.
  • Hilary DeCamp
    March 9, 2015
    I have done hundreds of segmentation in my career and can say with absolute certainty that basic demographics like gender and age rarely differentiate enough on attitudes or needs to be useful drivers of business action
  • Jonathan Weiss
    March 9, 2015
    I love this perspective. Additionally, we are increasingly getting to the point where we can make 'in-call adjustments' to customer service techniques based on the types of words callers are using, as well as their vocal tones. This way we can build upon our initial understanding about with whom we are speaking, with how they may be feeling during each particular interaction.

Add Your Comment