One hot topic at the recent Email Insider Summit in Park City, Utah, centered on the role and value of email preference centers. As with other email-marketing disputes, the debate that emerged from a panel discussion reflected general disagreement over the details and how marketers implement preference center, rather than focusing on the bigger picture. The anti-preference center camp's argument basically follows two lines of thinking: 1. Few subscribers actually use preference centers. Therefore, they have little value. 2. Consumers' behavior indicates interest and intent better than stated preferences or interests. The pro-preference center camp cites factors such as these: 1. Explicit preferences and interests can provide valuable subscriber profile data for better segmentation and targeting. 2. Preference centers provide alternatives to unsubscribing, such as changing email address, email frequency, communication preferences. 3. Profile or preference data can present a fuller picture of a consumer beyond his or her behavior alone. Looking deeper into the role and types of preference centers, we can see both camps are closer to each other than at first blush. Many folks really were talking about the efficacy of different types of preference centers. I classify preference centers into four types: 1. Opt-in (presented when someone is asked to join your email database) 2. Update (when someone clicks a link such as "Update Preferences") 3. Opt-out (similar to the update but designed specifically to offer alternatives to unsubscribing such as decreased frequency, snoozing subscriptions, etc.) 4. Purchase/Account Registration (designed mainly to be used in account registration or the purchase process; email subscriptions are a typical option in this process). Let's tackle the two core arguments against preference centers: 1: Preference centers are a waste of time because subscribers don't use them. This is often true for the update preference center, because subscribers likely don’t recall the preferences they provided at opt-in or can't easily find the preference center. Or perhaps they simply don’t see the connection between preferences and more relevant emails. To make this kind of preference center work better, you must promote the preference center beyond your welcome email. Consider sending "Update Your Preferences" emails a few times a year. Also, watch for behavioral signals. If engaged subscribers go inactive, invite them to update their preferences. Last, label the links to your preference center in your emails with the verbs that reflect what the subscriber wants to do: "Change Email Address," "Reduce Frequency" or "Modify Interests." Argument 2: Preference centers are a waste of time because people lie or don't do what they say. Both points are true. However, people will lie if you ask for information that they don't see as vital or necessary before they have learned to trust you. So, reassure customers about how you will use their data and protect their privacy to build trust. Also, don't ask for too much data too early in the relationship. That only encourages people to lie. Use progressive forms that capture data at different points during the relationship. Finally, people often buy things that don't reflect their expressed demographics or interests. They buy gifts or research products that reflect new interests. For example, someone who buys a dress could be a woman shopping either for herself or for a gift for someone else, or a man buying it for his wife, a girlfriend or daughter. Without gender data, a retailer could assume the dress purchaser is female and begin targeting a male subscriber only with offers for women's clothing and accessories. That retailer could be leaving significant money on the table. Here, capturing the right information during opt-in makes sense for the new subscriber and gives the marketer a more complete profile. Combining profile, preference and demographic information with behavioral data is the right answer. The reverse challenge is also true. Suppose a new subscriber tells you that he is male. You send him only male-oriented offers. The potential result? He goes to your competitor to shop for his wife. The problem is not with the preference center or bad data, but how you market. Knowing someone's gender does not preclude you from providing offers for the opposite gender. Rather, you might regularly include secondary gift suggestions. (It gets more complicated if your subscriber is LGBT, of course.) Finally, the concept of a preference center is morphing to include what subscribers do and do not do -- not just what they say, as I wrote in an earlier column, “Preference Centers: Rethink Rather Than Tweak.” Both sides in this debate have strong points. However, preference centers are neither dead nor valueless. Nor are they panaceas for bad marketing. As with most marketing issues, the answer and results you achieve depend upon how you actually implement a tactic or strategy. Until next time, take it up a notch.
While many marketers are aware of the role of statistical significance in A/B testing, a more fundamental understanding of the mechanics that lie beneath the phrase can improve tests' accuracy and defensibility. Read the full article on the Experian Marketing Services blog.