Skip to main content
All case studies
Case study CPG & Retail

Driving Retention and Growth Through Customer Segmentation

Unified fragmented CRM and loyalty data across regions for a premium cosmetics brand to build a behavioral segmentation engine that enabled targeted retention and upsell campaigns across high-value customer cohorts.

877KCustomers profiled
3 years (FY17–FY19)Behavioral data analyzed
Multi-regionCRM and loyalty systems harmonized
6+Segmentation dimensions

Situation

A premium cosmetics and wellness brand was experiencing uneven growth across markets and wanted to understand what was driving customer attrition in key segments. Although each region operated its own CRM and loyalty systems, there was no unified view of customer behavior across channels and markets.

Marketing teams lacked visibility into lifecycle patterns, repeat behavior, and cross-channel engagement, making it difficult to run targeted campaigns or measure their impact. Retention and acquisition metrics were tracked in silos, and the absence of a shared customer view meant that high-value segments were being treated the same as low-value ones, with no ability to personalize at scale.

Objective

The client needed a systematic way to identify high-value customer segments and uncover where growth opportunities were being missed. Three goals drove the engagement:

  • Unify fragmented CRM and loyalty data across regions and channels into a single consolidated behavioral dataset.
  • Build a segmentation framework that captures lifecycle stage, channel affinity, purchase frequency, spend tier, and category mix, going beyond simple demographic splits.
  • Deliver an interactive dashboard that CRM, marketing, and sales teams could use to drill into segment performance by geography and prioritize campaign targeting.

Approach

We began by harmonizing customer and transaction data from all regional CRM and loyalty systems. That unification step was the prerequisite for every subsequent analysis. Without a consistent customer identity across systems, behavioral segmentation was impossible to do reliably.

Behavioral segmentationWe segmented customers across six dimensions: lifecycle stage (new, repeat, lapsed, re-activated), channel affinity (online, offline, omnichannel), purchase frequency, spend tier, product category mix, and geography. Each dimension was constructed from transaction-level data rather than self-reported attributes, making the segments behavior-driven and directly actionable for campaign targeting.

Trend analysisWe analyzed retention, acquisition, repurchase, and cross-sell behavior from FY17 to FY19 to identify which segments were growing, which were declining, and where the largest gaps between customer acquisition cost and lifetime value existed. The age-by-ethnicity growth matrices surfaced nuanced attrition patterns that would have been invisible at an aggregate level.

Interactive dashboardWe built a dashboard giving CRM, marketing, and sales teams drill-down access to segment-level performance by geography, covering total customers, retail sales, annual value per customer, items per transaction, purchase frequency, and average selling price. A repeatable annual refresh process was set up to track segment shifts and reorder campaign priorities over time.

Figure 1Customer segmentation and behavior trends (2016-2018): total customers by cohort (new, existing, re-activated), retail sales in GBP M, annual value per customer, items per transaction, purchase frequency, and average selling price, tracked across three years to identify performance shifts by segment.

Figure 2Customer growth trends by age and ethnicity segment (YoY % change): heat map comparing 2017 vs. 2016 (left) and 2018 vs. 2017 (right), surfacing which demographic cohorts are declining most steeply and where recovery patterns are beginning to emerge.

Impact

The segmentation engine uncovered key cohorts with declining retention and rising customer acquisition cost, allowing marketing teams to redirect spend toward higher-lifetime-value segments rather than spreading budgets evenly across all customers.

The channel affinity analysis gave the team improved visibility into behavior differences between online-only, offline-only, and omnichannel customers. Omnichannel customers consistently showed higher frequency, higher spend, and stronger retention, a finding that directly shaped the brand’s investment priorities for loyalty program development.

Hyper-targeted retention, reactivation, and upsell campaigns became possible for the first time, built on behavior-driven audience definitions rather than broad demographic categories. The scalable segmentation framework created a foundation for future loyalty program design, customer lifecycle marketing, and CLTV optimization that the team could maintain and refresh independently.

Where do you stand?

Where does your business stand on data and AI today?

Answer 24 yes/no questions in about four minutes. You'll see where you sit on the curve, from Intuition to Autonomy, across six dimensions.

Take the diagnostic

Rather speak to us directly?

Start a conversation