Not your grandmother’s segmentation: understanding the retail analytics behind hyper-personalization
Last January was a whirlwind for the retail industry with the excitement and grandeur of the National Retail Federation’s BIG Show 2016. After reporting our four key takeaways from NRF, we want to look behind the curtain at one of the major themes we heard almost every NRF presenter and exhibitor talking about: hyper-personalization.
Here’s my colleague Benjamin Mokotoff’s definition of hyper-personalization: Fueled by data, hyper-personalization combines curation and personalization to create tailored, curated experiences based on individual consumer preferences. To put it another way, hyper-personalization is the marketing output of an unprecedented wealth of customer data.
What happened to segmentation?
You can think of personalization as the hip, millennial grandchild of segmentation. For retailers, customer segmentation was (and still is, for some) extremely useful because it takes away the one-size-fits-all approach to marketing by categorizing customers into sub groups based on particular preferences and behaviors.
Nearly 70% of millennial shoppers or younger are willing to share personal information in return for better offers from retailers. In contrast, only 40% of baby boomers said they were willing to do the same.
That used to be enough, but it’s a different landscape today: one that centers around the whims and fancies of the individual. Thankfully, retailers, marketers, and analysts have access to the technology to analyze consumers at the individual level, and deliver the hyper-personalized experiences they’ve come to expect.
Much like how grandma and grandpa didn’t have smartphones to stay connected and communicate 24/7, segmentation was never equipped with the powerful data capture and machine-learning capabilities to analyze customers at the individual level.
Big data analytics
In the past, retailers relied on data capture methodologies like surveys and focus groups to project the preferences and behaviors of a small sample size onto their general customer base. With today’s cloud infrastructure capabilities, statistical programming languages, and data science talent, retailers can churn through millions and sometimes billions of data records to glean insights. That enables retailers to capture the actual movements and purchases of every customer—no projections needed.
One NRF 16 session that aptly demonstrated the power of data to operationalize hyper-personalization was “Value From Big Data Analytics,” presented by Williams-Sonoma VP Sameer Hassan. In this session, we learned how Williams-Sonoma listens to both explicit and implicit signals in the data. The former contains declarative data that consumers provide by filling out forms and clicking on email campaigns. The latter contains electronic behavioral data trails in their browsing history that can be mined, enabling retailers to construct customer-level treatments and offers.
Though Williams-Sonoma didn’t divulge results from these marketing tactics, others are reporting success. A study from Wharton cited that Netflix reported 60% of its sales came from machine-learning recommendations and 35% of Amazon sales came from system-generated suggestions. Regarding increasing engagement, Venture Beat reports that personalized email subject lines can increase open rates by up to 41%.
Millennials and the value exchange
A recent global survey from Aimia reports that nearly 70% of millennial shoppers or younger are willing to share personal information in return for better offers from retailers. In contrast, only 40% of baby boomers said they were willing to do the same.
This trend of younger, digitally savvy customers willingly handing over their personal data in return for rewards is known as the “Value Exchange.” The Aimia Institute says: “This exchange often happens within the context of a formal loyalty or reward program—and for the Value Exchange to work, customers must know that the program represents a ‘safe haven’ in which they can share personal information without fear that the data will be stolen, misused, or collected without their express permission.”
Kroger gave a great example of the Value Exchange during its presentation at NRF, “Getting Personal Through Customer Science.” Matthew Thompson, Kroger’s VP of Digital Business, spoke to how Kroger used customer-provided data to develop digital products for customers to gain access to digital coupons and rewards, thus driving sales by an undisclosed amount: a win-win for all.
Delight—don’t fright—your customers
Given the favorable technological and social conditions of today’s marketplace, hyper-personalization seems like a marketer’s dream. We know more than ever about our customers: their shopping habits, dress sizes, shoe sizes, home locations, work locations, and more. Retailers should have no problem offering the retail trifecta: the right product, at the right time, at the right price. But they need to proceed with caution.
One of my favorite recent examples of hyper-personalization done right comes from Marriot. Marriot used my data to collect insights about my travel profile in 2015, recommending specific cities and hotels they thought I would enjoy based on look-alike travelers with profiles similar to mine.
But delight can easily turn into fright when it comes to how marketers use the specific information they’ve collected on consumers. One recent example I received comes from a daily news aggregation newsletter called the Skimm, which I simply had not read in a while. Skimm noticed this, too, and sent an email begging me to open their emails again. Like an old ex asking for you back, it reeked of desperation.
Another example of execution gone wrong was a paid ad on a search engine. They got wind of my last name, and then offered “Rescigno” at cheap prices. Attention grabbing, yes, but it creeped me out.
Navigating a new frontier
Armed with the right tools to capture and analyze detailed data about customers to provide hyper-personalized experiences and offers, retailers are entering unknown territory. With great power comes great responsibility, which means retailers must strike a balance between offering effective marketing treatments and crossing the line into Big Brother-like territory.
There’s no playbook on how to do hyper-personalization right. It’s all unfolding in the marketplace today, making this an extremely exciting time to be in retail analytics.
Have examples of hyper-personalization done right or wrong? Share them with me on Twitter at @AnalyticsKim.