Turning thousands of customers into five clear strategies

The challenge

A global manufacturer with thousands of customers was treating its entire base as a single, undifferentiated group. Marketing, retention, and account strategy were applied uniformly, giving the same outreach to a high-value, loyal account as to a one-time, price-sensitive buyer. Leadership knew the customers weren't all the same, but they had no structured way to tell them apart or to see which relationships were quietly at risk.

After a few down quarters and major accounts churning, the client realized they needed a more nuanced approach to customer management.

The approach

We ran the engagement through our WAVES methodology, moving from raw transactional data to a strategy leadership could act on.

Wrangle. We consolidated and cleaned years of customer and transaction data into a single analysis-ready dataset. The process involved extracting CRM data, conducting thorough data exploration, and resolving data-quality issues before any analysis was done.

Analyze. We built a behavior-based segmentation model, grouping customers by how they actually buy: recency, frequency, order value, product mix, and more. On top of the segments, we modeled churn risk using Kaplan–Meier survival analysis, which estimates the probability that a customer relationship "survives" over time and exposes where retention drops off fastest.

Visualize, Enrich & Strategize. We translated five distinct segments and their per-segment churn curves into a clear picture of who each group was, what they were worth, and how urgently each needed attention, then mapped a tailored play for each segment.

The results

• Five clear, behavior-based segments replaced a one-size-fits-all view of the customer base.
• Per-segment churn risk quantified with survival analysis, so retention effort could be aimed where it mattered most.
• A segment-level strategy leadership could execute against, instead of uniform outreach. The result? Customers were better served how they wanted to be, and the business saw improved outcomes through targeted efforts, leading to their best financial year in recent history.

Why it matters

Segmentation isn't a reporting exercise. It's the difference between spending retention budget evenly and spending it where it compounds. By pairing behavior-based segments with survival-modeled churn risk, the business stopped guessing which accounts needed attention and started acting on it. This is the kind of work an embedded fractional analytics partner delivers without the cost of a full-time data science hire. Now that the system is in place, the business can continuously refine its approach based on real-time insights.