Rethinking Personalization from Segmentation to Individualization
Part 1: Why a Segment-Based Marketing Approach Is So Hard for Business to Quit
If you’ve spent any time around personalization, then you’ve probably encountered the three
great truths about it. The first is that every business needs it. The second is that every
marketing tool is keen to “help” you achieve it. And the third is that rarely, if ever, does it actually involve a real-time personalized experience for the individual shopper on the other
Instead, what you typically get is some version of a rules-based system that can distinguish
between new and returning visitors and greet each with nominally personalized content or
offers. If you’re lucky, it will also include a few “Top Sellers” – “wisdom of the crowd” type
algorithms often passed off by the provider as AI – to promote top selling or popular products.
For those with more maturity, budget and a marketing tool like Adobe, it can advance to
involve the use of probabilistic segments and marketing campaigns that test pre-defined
treatments against pre-determined audiences. While varying in depth and sophistication, what
these, and most personalization efforts, all have in common is the reliance on segments to
design the experience. This should not come as any surprise.
Segments Have Been Very Good for Business
For marketers, the customer experience has long been defined by predetermined interest,
geographic, and demographic-based segments. From a business point of view, this made
perfect sense, as the broad-based segments were the only way to reach a large group of people
with specific characteristics – such as college-educated women aged 18 to 35, or with similar
interests, like dog or cat owners.
Segmentation was and still is an incredibly effective way to reach a mass audience. So it only
follows that entire business structures — systems, logic and planning — were built up and
organized to support it. And without something to replace it, these structures remain mainly in place, which is where most companies find themselves today.
When Consumers Rise, Segments Fall
The last few years have brought wholesale change to how we consume information and
interact with the world and the brands, retailers, and media that vie for our attention and
business. Between work and home, and the phone, computer, and other connected devices,
the average person now spends 2x amount more time online than they did 10 years ago. Digital
experiences have become an extension of our physical world, and the more time we spend with
any one particular digital brand or business, the more we expect them to know us and
understand what we want and/or prefer.
While current marketing stacks can simulate personalization and provide you with the
equivalent of the coffee cup with your name on it, it can’t go deeper to suggest what you put
into it. To go further and recognize that, for example, a change in behavior, such as switching from coffee to herbal or mint tea over a 2-3 day period might indicate that you have a cold or flu and then suggest a Medicine Ball or ColdBuster tea cocktail as alternative takes a level of personalization that segmentation and the systems built up to support them alone simply can’t provide.
Lack of Technology Scale Keeps Segments in Play
One byproduct of the all the time we spend online is the terabyte-upon-terabyte of data
produced. Up until a few years ago, only a small number of very large companies had the
resources and technology to actually do something with all that information. But even for the
few companies that could afford to store the data and a smaller subset had the wherewithal to
actually process it; the limitations of the technology at the time made generating insight at the individual level virtually impossible. The processing power and algorithms that were available did allow us to do things like cluster and classify users and products, and draw basic linkages between them, but it still wasn’t powerful enough to see beyond the aggregated group and recognize the individual customer for who they are and what they really want.
New Technologies Makes Individualization Possible, Finally
It has only been recently with the advent of new technologies, like Kafka and Spark, that
allowed the real-time capture of information and the ability to process and stream large
amounts of data we could finally think about going beyond segmentation. With the technology
we have today, it’s now not only possible to actually dive deeper into the data and identify the user at the individual level but also capture events as they happen, understand behavior,
context and intent and respond in real time with an appropriate algorithmically-derived
experience or action.
But just because the technology is available doesn’t necessarily mean business can immediately
adopt and make use of it. Personalizing for the user at the individual level is difficult enough by itself, but add in the myriad complexities, challenges, and constraints—like merchandising rules, “buying” vs. “shopping”, in-session offers — of doing it within the context of commerce, and it’s no surprise that most businesses and personalization vendors have not advanced beyond segmentation.
In my next blog, we’ll take a closer look at the technologies and skills required to take your
personalization beyond segmentation, as well as the macro and organizational challenges
preventing your businesses from successfully doing so.