Audience Segmentation AI Prompts

Build a churn prediction model to identify at-risk customers for proactive marketing intervention. Customer data: {{customer_data}} Churn definition: {{churn_definition}} Marketing interventions available: {{interventions}} 1. Feature engineering: Build predictive features for each customer (measured over the last 30/60/90 days): - Recency: days since last purchase or login - Frequency: purchase count in the last 90 days vs prior 90 days (trend) - Monetary: spend in last 90 days vs prior 90 days (trend) - Product usage: number of distinct products/features used - Engagement: email open rate, app sessions - Support signals: number of complaints or returns - Payment signals: failed payments, subscription downgrades 2. Churn probability model: - Logistic regression for interpretability (preferred for marketing teams) - Or gradient boosted trees for accuracy - Training: last 6 months of data; validation: most recent 30-60 days - Output: probability of churn within the next {{horizon}} days per customer 3. Risk tier definition: - High risk: churn probability > 60% - Medium risk: churn probability 30-60% - Low risk: churn probability < 30% - Size each tier: count and revenue at risk 4. Expected value of intervention: - For high-risk tier: Expected savings = (customers x churn probability x avg LTV x save rate) - Save rate: historical % of at-risk customers who respond to an intervention - Compare to intervention cost: is the program economically justified? 5. Intervention strategy by tier: - High risk: highest-value offer, personal outreach from CSM or account manager - Medium risk: automated personalized email with value reminder + soft incentive - Low risk: monitor; include in standard engagement program 6. Measurement plan: - Control group: randomly hold out 10% of at-risk customers from interventions - Measure: 60-day churn rate in treated vs control group - Calculate: incremental save rate and revenue impact Return: feature engineering spec, model approach, risk tier definitions, intervention strategy, expected value calculation, and measurement plan.

Build and operationalize customer segments for targeted marketing. Customer data: {{customer_data}} (demographics, behavioral, transactional, engagement) Marketing goals: {{goals}} Channels available: {{channels}} 1. Segmentation approach selection: Demographic segmentation: - Age, gender, location, income, job title - Pros: easy to understand and action - Cons: weak predictor of behavior for most products Behavioral segmentation: - Purchase history, product usage, channel preferences, engagement frequency - Pros: directly tied to marketing-relevant actions - Best for: personalization, cross-sell, win-back RFM (Recency, Frequency, Monetary): - Recency: how recently did they purchase? - Frequency: how often do they purchase? - Monetary: how much do they spend? - Quintile score (1-5) on each dimension; combine into segment labels Psychographic / attitudinal: - Values, motivations, lifestyle - Pros: powerful for brand messaging - Cons: requires survey data, harder to operationalize 2. RFM segmentation execution: For each customer, compute R, F, M scores (1-5): - Champions: RFM = 5,5,5 (buy often, recently, high value) - Loyal customers: 4+,4+,3+ - At-risk: previously high RFM but R has dropped - Potential loyalists: recent but low frequency - Win-back: low R, previously decent F and M - Lost: low on all three dimensions 3. Segment sizing and value: - Size (count and % of customers) - Average order value, purchase frequency, LTV by segment - Total revenue contribution by segment 4. Segment-to-channel mapping: For each segment: which channels and messages are most appropriate? - Champions: VIP program, referral program, early access - At-risk: re-engagement email, win-back offer - Potential loyalists: loyalty nudge, second purchase incentive 5. Personalization rules: - What content, offer, and message should each segment receive? - Build a segment x message matrix Return: RFM segment definitions and scoring logic, segment sizing table, revenue contribution, channel mapping, and personalization rules.

Build a lookalike audience strategy based on best-customer characteristics. Seed audience: {{seed_audience}} (your best customers by LTV or conversion) Available platforms: {{platforms}} (Meta, Google, LinkedIn, programmatic DSP) Campaign goal: {{goal}} 1. Seed audience definition and quality: - Define 'best customers': top 10% by LTV, or converted within 30 days, or completed {{action}} - Minimum seed size: 1,000 users for reliable lookalike modeling (500 minimum, 5,000+ recommended) - Seed quality check: are seed customers actually your most profitable? (Not just most active) 2. First-party data preparation: - Match rate optimization: use email, phone, MAIDS for highest match rates - Hashed PII: never pass unhashed emails to platforms - Audience freshness: use customers acquired in the last 6 months for best results - Exclude: existing customers from prospecting lookalike campaigns 3. Platform-specific lookalike construction: Meta Lookalike Audiences: - Similarity range: 1% (most similar) to 10% (broader reach, less similar) - Recommendation: 1-2% for highest intent, 3-5% for broader prospecting - Layer with interest targeting for higher precision Google Similar Audiences / Customer Match: - Smart Bidding automatically adjusts bids for similar audiences - Customer Match can be used for similar segments via automatically created lists LinkedIn Lookalikes: - Most valuable for B2B: match on company, industry, job title characteristics - Seed with MQL or customer list from CRM 4. Testing framework: - A/B test: lookalike 1% vs lookalike 3% vs interest targeting vs no audience filter - Measure: CPA and conversion rate per audience type - Duration: minimum 2 weeks, 50+ conversions per variant for statistical reliability 5. Performance benchmarks: - Lookalike audiences should outperform broad targeting by 20-40% on CPA - If lookalike is not outperforming: seed audience may not be differentiated enough Return: seed audience definition, data preparation checklist, platform construction guide, testing framework, and performance benchmarks.

Develop data-driven marketing personas for {{product}}. Data sources: {{data_sources}} (CRM, survey, behavioral analytics, interviews) Existing customer base: {{customer_count}} customers 1. Quantitative customer profiling: From CRM and analytics data, compute for each customer: - Company size, industry, geography (for B2B) - Demographics: age, gender, job title (for B2C or B2B) - Behavioral: acquisition channel, feature usage, purchase frequency - Value: LTV, plan/product tier, churn risk score 2. Cluster analysis: - Apply k-means clustering (k=3 to 5) on the behavioral and firmographic features - For each cluster: profile using the mean values of each feature - Name each cluster based on its defining characteristics 3. Qualitative enrichment: - For each cluster, pull the top 10% by LTV and review their CRM notes, support tickets, and survey responses - Identify: primary use case, main problem solved, key decision criteria, objections to purchase - Add the 'voice of the customer' to each persona: direct quotes from reviews or interviews 4. Persona template per cluster: - Name and title: a memorable label (e.g. 'The Efficiency-Focused Ops Manager') - Demographics/firmographics: size, role, industry - Primary goal: the main outcome they are trying to achieve - Pain points: the problems your product solves for them - Decision criteria: how they evaluate solutions - Information sources: where they get industry information - Objections: their most common reasons not to buy 5. Persona sizing and value: - What % of current customers does each persona represent? - What % of revenue does each persona account for? - Which persona is under-represented in the customer base vs addressable market? 6. Marketing application: - Recommended messaging for each persona - Recommended channels to reach each persona - Recommended content type for each persona Return: cluster analysis, persona template for each cluster, sizing table, and marketing application guide.