ncreighton Segmento: AI-Powered Customer Segmentation
Overview
Segmento is an enterprise-grade customer segmentation skill that uses machine learning clustering algorithms to automatically group customers into meaningful, actionable segments. Instead of manually creating customer lists or relying on basic RFM analysis, Segmento analyzes behavioral patterns, purchase history, engagement metrics, and demographic data to discover hidden customer cohorts—then automatically syncs those segments to your CRM, email platform, or marketing automation tool.
Why This Matters:
- Precision Targeting: Move beyond broad demographics to behavior-driven segments
- Marketing ROI: Personalized campaigns to the right audience increase conversion by 20-40%
- Churn Prevention: Identify at-risk customers before they leave
- Revenue Growth: Upsell and cross-sell to high-value segments with proven propensity
- Time Savings: Automation replaces hours of manual list-building
Integrations Supported:
- HubSpot, Salesforce, Pipedrive (CRM sync)
- Mailchimp, ConvertKit, ActiveCampaign (email list push)
- Slack (segment notifications)
- Google Sheets (data export)
- Zapier webhooks (custom workflows)
Quick Start
Try these prompts immediately:
Example 1: Behavioral Segmentation
Segment my customer base by purchase frequency and recency.
I have 5,000 customers with transaction history from the past 24 months.
Create segments for: high-value repeat buyers, dormant customers, and new customers.
Export to HubSpot as contact lists.
Example 2: Churn Risk Detection
Analyze my customer engagement data and identify customers at risk of churning.
Use metrics: login frequency, feature usage, support tickets, last purchase date.
Flag customers with declining engagement in the last 90 days.
Create a "Save Me" segment for re-engagement campaigns.
Example 3: Product Affinity Clustering
Group customers by product purchase patterns.
I sell: SaaS plans (Starter, Pro, Enterprise), add-ons, and professional services.
Create segments showing: plan tier affinity, add-on propensity, and upsell opportunities.
Suggest personalized product recommendations per segment.
Example 4: Demographic + Behavioral Blend
Segment my e-commerce customers by location, purchase value, and category preferences.
I have customer data: geography, order history, browse behavior, email engagement.
Create 6-8 segments optimized for targeted email campaigns.
Include segment size, average LTV, and recommended offer per segment.
Capabilities
1. Intelligent Clustering
Segmento uses k-means, hierarchical clustering, and DBSCAN algorithms to discover natural customer groupings:
- Behavioral Clustering: Purchase frequency, order value, category preferences, browsing patterns
- Engagement Clustering: Email opens, click-through rates, content consumption, feature usage
- Demographic Clustering: Location, company size, industry, customer lifecycle stage
- RFM Analysis: Recency, Frequency, Monetary value (classic + advanced variants)
- Churn Prediction: Identifies flight-risk customers with 85%+ accuracy
Usage Example:
Analyze my Shopify customer base and create 5 distinct segments.
Data sources: order history, customer lifetime value, product category affinity, email engagement.
Optimize for: marketing campaign personalization and retention strategy.
Output: Segment profiles with size, characteristics, and recommended actions.
2. Customizable Segmentation Rules
Define your own segmentation logic without coding:
- Rule Builder: Combine conditions (AND/OR logic) for manual segment creation
- Weighted Scoring: Assign importance to different attributes
- Time-Based Rules: Segment by recency windows (last 30/60/90 days)
- Dynamic Thresholds: Auto-adjust segment boundaries based on data distribution
Usage Example:
Create a "VIP" segment with custom rules:
- Customers with lifetime value > $5,000 AND
- Purchase frequency >= 10 orders AND
- Email engagement rate > 30% OR
- Product category = "Premium Plan"
Exclude: customers with active refund requests or support complaints.
3. CRM & Marketing Platform Integration
Automatically sync segments to your existing tools:
- HubSpot: Create/update contact lists and custom properties
- Salesforce: Push segments as Salesforce lists and account hierarchies
- Pipedrive: Sync to deal pipelines and organization segments
- Mailchimp: Create audience segments for targeted campaigns
- ConvertKit: Tag subscribers by segment for automation
- ActiveCampaign: Update contact records and trigger automations
- Google Sheets: Export segment data for analysis and reporting
Usage Example:
Push my 4 customer segments to HubSpot.
Segment 1 (High-Value): Create list "VIP Customers", set property "customer_tier=premium"
Segment 2 (At-Risk): Create list "Churn Risk", trigger HubSpot workflow "win-back campaign"
Segment 3 (New): Create list "New Customers", add to onboarding email sequence
Segment 4 (Dormant): Create list "Reactivation", tag with "dormant_90days"
4. Advanced Analytics & Insights
Get actionable intelligence from your segments:
- Segment Profiles: Size, growth rate, average metrics, composition
- Cohort Analysis: Track segment behavior over time
- Predictive Scoring: Next purchase likelihood, churn probability, LTV prediction
- Comparison Reports: Benchmark segments against each other
- Visualization: Charts, heatmaps, and trend analysis
Usage Example:
Generate a segment comparison report.
Show: size, average order value, purchase frequency, churn rate, email engagement for each segment.
Highlight: which segments are growing/shrinking, which are most profitable.
Recommend: top 3 actions per segment to increase revenue.
5. Workflow Automation
Set up recurring segmentation jobs:
- Scheduled Runs: Daily, weekly, or monthly re-segmentation
- Trigger-Based: Re-segment when new data arrives or thresholds are crossed
- Change Notifications: Slack alerts when customers move between segments
- Audit Trail: Track all segment changes and data lineage
Usage Example:
Set up automated weekly segmentation.
Data source: Shopify API (pull orders, customers, products)
Segmentation: RFM analysis + churn risk scoring
Sync: Push updated segments to HubSpot and Mailchimp every Monday at 9 AM
Notify: Send Slack summary to #marketing with segment changes and key metrics.
Configuration
Required Environment Variables
# Segmento API credentials
SEGMENTO_API_KEY=sk_live_xxxxxxxxxxxxx
# CRM integration (choose at least one)
HUBSPOT_API_KEY=pat-xxxxxxxxxxxxx
SALESFORCE_API_KEY=xxxxxxxxxxxxx
PIPEDRIVE_API_TOKEN=xxxxxxxxxxxxx
# Email platform (optional)
MAILCHIMP_API_KEY=xxxxxxxxxxxxx
CONVERTKIT_API_KEY=xxxxxxxxxxxxx
# Data source credentials (optional)
SHOPIFY_ACCESS_TOKEN=shpat_xxxxxxxxxxxxx
STRIPE_API_KEY=sk_live_xxxxxxxxxxxxx
GOOGLE_SHEETS_CREDENTIALS={"type":"service_account",...}
Configuration Options
segmentation:
algorithm: "kmeans" # Options: kmeans, hierarchical, dbscan, gaussian_mixture
num_clusters: 5 # Auto-detect if null
features:
- purchase_frequency
- order_value
- category_affinity
- email_engagement
- churn_risk_score
scaling: "standard" # Options: standard, minmax, robust
random_state: 42 # For reproducibility
sync:
crm: "hubspot" # Options: hubspot, salesforce, pipedrive, none
email_platform: "mailchimp"
sync_frequency: "weekly" # Options: daily, weekly, monthly, manual
analytics:
enable_cohort_analysis: true
enable_predictive_scoring: true
retention_days: 90
Setup Instructions
Obtain API Keys
- Segmento: Visit segmento.app/api and create an API key
- HubSpot: Settings > Integrations > Private apps
- Shopify: Admin > Apps > App and sales channel settings > Develop apps
- Stripe: Dashboard > Developers > API keys
Set Environment Variables
export SEGMENTO_API_KEY="sk_live_..." export HUBSPOT_API_KEY="pat-..."Connect Data Sources
- Upload customer CSV, or
- Connect CRM/e-commerce platform via OAuth, or
- Use webhook to stream real-time customer events
Define Segmentation Strategy
- Choose algorithm (k-means recommended for most use cases)
- Select features/attributes to include
- Set number of segments (5-8 is typical)
Test & Deploy
- Run segmentation on sample data
- Review segment profiles and quality
- Sync to CRM and test workflows
- Schedule recurring jobs
Example Outputs
Segment Profile Report
SEGMENT: High-Value Repeat Buyers
├─ Size: 342 customers (6.8% of base)
├─ Growth: +12% month-over-month
├─ Avg Customer LTV: $8,450
├─ Avg Order Value: $245
├─ Purchase Frequency: 18.3 orders/year
├─ Email Engagement: 45% open rate, 8.2% CTR
├─ Churn Risk: 2% (very low)
├─ Top Product Categories: Premium Plans (78%), Add-ons (65%)
└─ Recommended Actions:
├─ VIP loyalty program enrollment
├─ Early access to new features
├─ Quarterly business reviews
└─ Upsell: Enterprise tier (30% conversion potential)
SEGMENT: At-Risk Churners
├─ Size: 156 customers (3.1% of base)
├─ Growth: -8% month-over-month
├─ Avg Customer LTV: $1,200
├─ Days Since Last Purchase: 127 days
├─ Email Engagement: 12% open rate, 1.1% CTR
├─ Feature Usage: Declined 60% in last 30 days
├─ Churn Risk: 78% (critical)
└─ Recommended Actions:
├─ Trigger win-back email sequence
├─ Offer: 30% discount on renewal
├─ Sales outreach: Personal call from CSM
└─ Feedback survey: Understand pain points
SEGMENT: New Customers
├─ Size: 890 customers (17.7% of base)
├─ Growth: +25% month-over-month
├─ Avg Customer LTV (projected): $2,100
├─ Days as Customer: 18 days average
├─ Email Engagement: 52% open rate, 9.5% CTR
├─ Product Adoption: 65% activated core features
├─ Churn Risk: 22% (moderate)
└─ Recommended Actions:
├─ Onboarding email sequence (3-email series)
├─ Product tutorial and best practices
├─ Offer: Early-bird discount on upgrade
└─ Track: Feature adoption and support tickets
Segment Comparison Matrix
Metric High-Value At-Risk New Dormant
─────────────────────────────────────────────────────────────────
Segment Size 342 156 890 1,240
% of Base 6.8% 3.1% 17.7% 24.6%
Avg LTV $8,450 $1,200 $2,100 $450
Avg Order Value $245 $89 $120 $65
Purchase Freq/Yr 18.3 2.1 4.2 0.3
Email Open Rate 45% 12% 52% 8%
Churn Risk % 2% 78% 22% 65%
Growth Rate +12% -8% +25% -15%
Integration Output (HubSpot Sync)
✅ Synced 4 segments to HubSpot
├─ List: "VIP Customers" (342 contacts)
│ └─ Property: customer_segment = "high_value"
│ └─ Property: ltv_tier = "premium"
│ └─ Workflow: "VIP Engagement" triggered
├─ List: "Churn Risk" (156 contacts)
│ └─ Property: customer_segment = "at_risk"
│ └─ Property: churn_probability = 0.78
│ └─ Workflow: "Win Back Campaign" triggered
├─ List: "New Customers" (890 contacts)
│ └─ Property: customer_segment = "new"
│ └─ Property:
│ └─ Workflow: "Onboarding Series" triggered
└─ List: "Dormant Customers" (1,240 contacts)
└─ Property: customer_segment = "dormant"
└─ Property: days_inactive = 180+
└─ Workflow: "Reactivation Campaign" triggered
Sync completed: 2,628 contacts updated in 4.2 seconds
Tips & Best Practices
1. Choose the Right Algorithm
- K-means: Best for balanced, spherical clusters (most common use case)
- Hierarchical: Best when you want to understand cluster relationships and dendrograms
- DBSCAN: Best when clusters are irregular or you have outliers to isolate
- Gaussian Mixture: Best for probabilistic assignments and soft boundaries
Recommendation: Start with k-means for 80% of use cases.
2. Feature Selection is Critical
- Include: Purchase frequency, order value, category affinity, engagement, recency
- Exclude: Customer name, email (PII), random identifiers
- Weight: Give higher importance to revenue-driving metrics
- Normalize: Ensure all features are on comparable scales
Example:
Features (with weights):
- purchase_frequency: 1.0
- order_value: 1.5 (revenue-weighted)
- email_engagement: 0.8
- product_category_diversity: 0.6
- churn_risk_score: 1.2 (retention-critical)
3. Determine Optimal Cluster Count
- Elbow Method: Run k=2 to k=10, plot inertia, find the "elbow"
- Silhouette Score: Higher is better (range: -1 to 1, aim for 0.5+)
- Business Logic: 5-8 segments is typical for mid-market, 10-15 for enterprise
- Actionability: Can you create distinct marketing/sales strategies for each?
Rule of Thumb: num_clusters = sqrt(num_customers / 2)
4. Validate Segment Quality
Before deploying to production:
- Review segment profiles manually
- Verify segments are distinct (not overlapping)
- Check for imbalanced clusters (one huge segment = bad)
- Get stakeholder feedback (marketing, sales, CS teams)
- A/B test personalized campaigns on segments
5. Sync Strategy
- Daily Sync: For high-velocity businesses (SaaS, e-commerce)