Case Study · Predictive Analytics
Customer Segmentation for Automotive Customers
Data-driven customer segments built from purchase, service and engagement behaviour, turned into named personas the marketing and sales teams can act on.
← All case studies
Client
Dealer group running brand-wide CRM campaigns
Industry
Automotive
Delivers
Behavioural segments ready to use in the CRM
The problem
Campaigns were targeted by age, location and vehicle model, and every customer received the same messages. Engagement was falling, and the team had no shared view of who their customers really were.
Approach
- Feature engineering on recency, frequency and monetary value (RFM), vehicle ownership cycle, service regularity, channel preference, finance use and responses to past campaigns.
- Clustering (k-means and Gaussian mixture models), with the number of segments chosen for stability and for how easily the business can act on them.
- Segments turned into personas, for example Loyal Service Regulars, Business & Fleet Buyers, Value-Seeking First-Time Owners, Lapsed Post-Warranty Customers and Premium Upgraders.
- A classifier assigns new customers to a segment automatically, and segment stability is monitored over time.
Activation
- Segments are synced to the CRM and marketing automation as audiences.
- Each segment gets its own messaging, offers and channel mix.
- A dashboard tracks segment size, movement between segments and revenue.
KPIs we design for
Campaign engagement
Open, click and response rates by segment
Conversion
Per segment against a generic-campaign control
Segment migration
Customers moved into higher-value segments
Shared customer language
Personas used by sales, service and marketing
Technology
PythonScikit-learnPandasdbtCRM / marketing automationBI dashboards
Have a similar challenge?
Talk to our team about what this could look like for your business.