Online Shopping

When I began working in CRM and analytics, brands segmented audiences in a highly rigid, top-down way. To attract high-value customers, for example, retailers might simply target consumers living in high-income zip codes. Two decades into my career, I still see many brands approach customer segmentation this way.

And yet in the digital age, brands no longer need such guesswork. Instead of a top-down model based on assumptions, they can work from a bottom-up approach based on reality. With the right tools, they can quickly identify who their HVCs really are. Then they can discover the key attributes these HVCs share, e.g. product affinities, online and off-line behaviors, demographics, etc. Finally, they reach out to other consumers with the same set of key attributes in order to convert them to the brand.

I have worked with a number of top fashion retailers who have embraced modern segmentation. To help you do the same, I propose:

  • a clear definition of modern segmentation — what I call “behavioral and contextual targeting”;
  • the four dimensions of customer data that fuel modern segmentation, and
  • three critical best practices to put modern segmentation into practice.

Modern Segmentation: Behaviors and Context 

Modern segmentation does not rely on a single type of behavior or demographic attribute. Rather, it takes into account the customer’s holistic interactions with your brand over time (i.e. spending patterns, engagement signals, etc.) in order to identify profiles and preferences.

Just as important as behaviors, modern segmentation also takes into account “context.” In other words, is the customer in buying mode right now (browsing your site, clicking ads, etc.), and if so, what products and offers are attracting their attention in the moment?

It is by combining the two that you can deliver the right message to the right person at the right time, via the right channel.

The Four Dimensions of Customer Data

To make “behavioral and contextual targeting” possible, marketers must be able to creatively combine datasets across four dimensions of first-party data, including:

  1. Demographics — Who the Customer Is. More or less stable traits, like gender, age or income. For example, an older customer is, overall, more likely to be interested in certain fashion accessories than a younger customer.
  2. Historic/Behavioral data — What the Customer Did In the Past. Past behavior — purchase transactions, carts abandoned, etc. — can be a good predictor for future action. For example, a customer who has purchased a pair of shoes at the beginning of the season at full price is more likely to do so again in the future.
  3. Streaming/Contextual data — What the Customer is Doing Right Now. A customer’s current digital behaviors — web pages they are viewing, e-mails they just open, ads they are clicking on — indicate she is interested in certain products at the moment and currently in shopping mode.
  4. Predictive Analytics — How the Customer Is Likely to Respond in the Future. By finding insights locked away in data from the three dimensions above, AI and predictive analytics detect invisible patterns even the savviest marketer may miss. For example, the woman browsing that high-ticket clothing item may share key attributes with customers who purchase gourmet cookware, making for a strong cross-sell opportunity.

Best Practices for Ramping Up Modern Segmentation 

How do you jumpstart your journey to modern segmentation? Don’t try to drink the ocean — or petabytes of data — in a single gulp. To avoid overwhelm, here are three best practices that I’ve seen leading-edge brands successfully adopt:

  1. Start with a specific use case. Instead of mining all your data to identify an ideal customer, begin with a simple, practical goal you want to achieve. Then partner with data and analytics teams to discover as much as possible about customers who already engage with the brand the ways you desire. Do you want to launch a retention campaign? Start with segments based on days since last purchase and reach out to them with relevant offers. Loyalty acquisition program? Use lifetime value to understand who has the potential to increase spend velocity. Cross-sell a new category? Estimate the likelihood to purchase within it.
  2. Test, measure, iterate. Build testing into your earliest efforts at modern segmentation. This helps refine future iterations of your strategy. Just as important, it helps demonstrate the business value of this new approach to key decision-makers — critical in winning their buy-in to scale modern segmentation across all your activities. I worked with one brand that established a cross-functional team that ensures every segmentation strategy has a specific KPI goal, every test is set-up appropriately, and all results are transparent to key stakeholders. This helped turn an academic exercise into a powerful tool to drive the business.
  3. Scale with smart hub technology. In building out your test case, you may run up the limits of your brand’s IT, data and analytics resources. How could you ever scale modern segmentation given your current resources? One option is to adopt what Gartner calls smart hubs. These solutions are designed to speed and automate critical steps in modern, data-driven segmentation, including:
  • unification of data (historic/behavioral and contextual) from across disconnected marketing systems;
  • self-service access to that data for marketers;
  • predictive insights that grow out of all that data, and
  • coordinated execution across all channels.

Tamara Gruzbarg is head of industry insights at ActionIQ Inc.

load comments
blog comments powered by Disqus