Boarding the B-U-S: Defining and Measuring Personalization in a Multi-Channel World

bigstock-Advertising-Truck-With-Statist-50424437-565x300

A few years ago, when marketers discussed customer personalization, they were likely referring to a customer’s first name in a subject line, customer segments based on attributes like recency and frequency, and perhaps even some “also viewed” product recommendations. These approaches proved valuable from a conversion perspective and undoubtedly still do, but, truth be told, today’s marketing pioneers understand that personalization means much more than just a name or a linear recommendation.

The new face of personalization:

  • Behavioral – Many inherent factors are important to consider to drive personalization – device tendencies, geo-location factors and time of day insights are just a few examples in a long list. If you know a customer is most likely to complete a purchase at 8pm, why would you send that customer emails at 7:15am? Relatedly, if your customer spends 5 days a month in DC on business, why are you only promoting nearby stores in her home state of Texas?
  • Usage – “RFM” (recency-frequency-monetary value) is still popular for segmenting customer bases and deploying targeted messages. This tactic, while very effective, is just one piece of a bigger puzzle. I recommend marketers continue to leverage usage information to drive personalization (e.g. search results abandonment, win-backs, etc.), but we also suggest incorporating more data on the inherent interests that underpin that usage – not just what the customer looks at, but what attributes do the products she views share (for instance, are they all red? All on sale? Does the customer only buy when free shipping is offered?).
  • Situational – Situational analytics seek to understand the “why” of someone’s browsing or buying behaviors. Consider this example: a customer first browses your website looking for pants near the end of the holiday rush, and most SKUs are sold out. This circumstance will undoubtedly impact future purchasing behaviors and needs to be accounted for and addressed head-on. Let’s take things a step further and assume this customer then goes to your store; rather than pushing her toward your shoe sale, you’d ideally want to capitalize on the reality that she is trying to find pants in stock!

Beyond just personalization, though, the tenets above are incredibly valuable for business analytics. The ability to discern variations in customer engagement, average order value, and other key performance metrics based on how/when/where a customer buys are truly invaluable.

Put in a more pragmatic way, is the customer more valuable when she shops online, offline or across both channels? Do live chat sessions yield higher average order values? Are customer who buy for the first-time on Cyber Monday more or less valuable as other cohorts? Marketers need this level of transparency into multidimensional analytics.

A unified view of the customer is table stakes for driving the type of personalization I prescribe and champion this as a former marketer myself, so even if 360-degree personalization is still a ways off on your radar, I encourage you to take the right steps toward that single view of the user today. More important though is the significance of building an analytical framework to understand the impact of every optimization decision you make.

This post from Cassie Lancellotti-Young, the VP of Analytics & Optimization at Sailthru, originally appeared on Wired Insights.