Customer Retention Engine

Customer Retention Engine

Overview

Acquiring a new customer costs 5–7x more than retaining an existing one. A retention engine identifies customers who show declining engagement — reduced purchase frequency, decreasing order values, browsing without buying — and intervenes with personalized campaigns before they lapse. Klaviyo and AutomateWoo can build these workflows with no custom code using predictive analytics built into the platform.

Note: For reactivating already-lapsed customers, see @win-back-reactivation. This skill focuses on proactive churn prevention before customers lapse.

When to Use This Skill

• When repeat purchase rate is declining month-over-month
• When a significant percentage of customers only ever purchase once
• When you want to proactively contact customers before they go fully dormant
• When building a post-purchase nurture program beyond the first 30 days
• When needing to identify which customers are worth offering a discount vs. which will repurchase anyway

Core Instructions

Step 1: Choose the right tool for your platform

Step 2: Define your churn threshold

Churn timing depends on your product's natural repurchase cycle:

In Klaviyo, go to Analytics → Predictive Analytics to see Klaviyo's automatically calculated churn risk for your customer base. Klaviyo uses your actual order history to set these thresholds — no manual calculation needed.

Step 3: Build retention flows

---

Shopify with Klaviyo

Flow 1: Early warning — customers approaching their expected repurchase date

1. Go to Klaviyo → Flows → Create Flow → Start from Scratch 2. Set trigger: Metric → Expected Date of Next Order is in 7 days

• Klaviyo calculates this automatically using predictive analytics

3. Add an email action with subject line: "Your favorites are waiting for you"

• Include: personalized product recommendations based on past purchases
• Use the {{ person.predicted_next_purchase_date }} variable to acknowledge timing

4. Wait 3 days → Add a conditional: "Has placed an order since flow start?" → If No: send a follow-up email

Flow 2: High-value at-risk customers — personalized offer

1. Create a new flow triggered by: Segment → "High-Value At-Risk" segment (configure segment below) 2. Email: personalized exclusive offer (free shipping or 15% off for top-tier; no discount for mid-tier) 3. Wait 5 days → If no purchase: SMS follow-up (for SMS-consented customers)

Create the "High-Value At-Risk" segment in Klaviyo: 1. Go to Klaviyo → Segments → Create Segment 2. Conditions:

• Predicted Churn Risk equals High OR Mid
• AND Total Customer Value (Historic CLV) greater than $150
• AND Has not placed order in last 60 days

3. Save as "High-Value At-Risk"

Flow 3: One-time buyer reminder

1. Create a flow triggered by: Metric → Placed Order 2. Wait 45 days 3. Check: "Total number of orders equals 1" → If yes: 4. Email: "Based on your [product name] purchase, you might love these..." with 3–4 complementary product recommendations 5. Wait 14 days → If still only 1 order: send final email with 10% welcome-back discount

---

WooCommerce with AutomateWoo

1. Go to AutomateWoo → Workflows → Add Workflow 2. Set trigger: Customer → Win Back

• AutomateWoo has a built-in "Customer Win Back" trigger that fires at configurable intervals after last purchase

3. Set timing rules:

• Workflow A: 45 days since last purchase — send personalized recommendations email
• Workflow B: 75 days since last purchase — send exclusive offer email with discount

4. Add a rule to each workflow: "Customer has not purchased since workflow was created" — auto-cancels on conversion 5. For VIP customers: add a rule "Customer total spend > $500" to route to a separate workflow with better offers

AutomateWoo also includes built-in RFM segmentation — go to AutomateWoo → Reports → RFM Analysis to see your customer segments visually.

---

BigCommerce

1. Install Klaviyo from the BigCommerce App Marketplace 2. Klaviyo automatically syncs all BigCommerce order history 3. Follow the same Klaviyo flow setup described in the Shopify section above 4. BigCommerce order events trigger Klaviyo flows in real-time after the integration is connected

---

Custom / Headless

Send order events to Klaviyo's API to trigger retention flows:

// Send order event to Klaviyo when an order is completed
async function trackKlaviyoOrder(order: Order) {
  await fetch('https://a.klaviyo.com/api/events/', {
    method: 'POST',
    headers: {
      'Authorization': `Klaviyo-API-Key ${process.env.KLAVIYO_PRIVATE_KEY}`,
      'Content-Type': 'application/json',
      'revision': '2024-10-15',
    },
    body: JSON.stringify({
      data: {
        type: 'event',
        attributes: {
          metric: { data: { type: 'metric', attributes: { name: 'Placed Order' } } },
          profile: { data: { type: 'profile', attributes: { email: order.customerEmail } } },
          properties: {
            $value: order.subtotal,
            OrderId: order.id,
            Items: order.lineItems.map(i => ({ ProductName: i.name, ItemPrice: i.price })),
          },
        },
      },
    }),
  });
}

Klaviyo's predictive analytics then automatically calculates churn risk and expected next purchase date based on these events.

Step 4: Configure intervention timing and incentives

Match interventions to customer value tier:

Step 5: Measure retention lift

Track these in Klaviyo or AutomateWoo dashboards:

Best Practices

• Never discount VIP customers reflexively — high-LTV customers who are slightly overdue may just be busy; a soft nudge without discount often works and protects margin
• Personalize based on actual purchase history — Klaviyo's dynamic product blocks automatically show recommendations based on what the customer has bought
• Set a contact frequency cap — in Klaviyo, use Smart Sending and set a daily message limit; at-risk customers should not receive more than one touchpoint per week
• Use Klaviyo's predictive analytics — do not manually calculate churn scores; Klaviyo computes expected next purchase date and churn risk automatically from your order data
• Close the loop with CS for high-value customers — for accounts over $1,000 LTV showing high churn risk, route to a customer success rep via Gorgias or Zendesk for personal outreach

Common Pitfalls

Related Skills

• @lifecycle-marketing-automation
• @win-back-reactivation
• @loyalty-program-optimization
• @email-marketing-automation
• @email-list-segmentation

Behavioral Triggers and Retention Campaign Analytics

Problem/Feature Description

Orchard Home is an online home goods retailer that has started sending retention emails but isn't sure whether they actually work. The head of growth wants two things: (1) a set of real-time behavioral triggers so that emails fire when customers show specific disengagement signals rather than on a fixed schedule, and (2) a measurement framework that produces reliable evidence that the campaigns are generating incremental repurchases — not just capturing customers who would have come back anyway.

The team also wants to make sure they're not burning goodwill by emailing enthusiastic VIP customers with aggressive discount offers every time they take a brief pause, and they want to understand whether certain campaigns are causing customers to unsubscribe at elevated rates.

Output Specification

Produce a TypeScript file behavioral-triggers.ts implementing two event-driven trigger functions that fire based on customer activity signals.

Produce a second TypeScript file campaign-analytics.ts implementing a lift measurement function that compares campaign recipients against a control group and returns repurchase rates, percentage lift, and revenue recovered.

Also produce an ANALYTICS_NOTES.md that covers: 1. What signals trigger each behavioral email and under what conditions 2. How the control group is selected and what fraction of at-risk customers it represents 3. How the measurement handles the comparison to calculate lift 4. How to avoid over-discounting VIP customers 5. What unsubscribe tracking should look like

Customer Churn Scoring System

Problem/Feature Description

Bloom & Co is a mid-sized online florist with a growing repeat customer base. Their marketing team has noticed that many customers who bought once or twice simply disappear, and the company has no way of knowing who is truly "gone" versus who is just between purchase cycles. The data team has access to the full order history per customer but currently has no scoring system to classify who is at risk of churning.

Your task is to design and implement a TypeScript churn scoring module that the team can run daily as part of their data pipeline. The module should assign risk levels to customers based on how overdue they are relative to their own individual purchase patterns. Customers with only a single order should receive a sensible fallback estimate. The team also has a mix of subscription-box customers and one-off purchasers in the same database, and they want to make sure the scoring logic applies correctly only to the right population.

Output Specification

Produce a single TypeScript file named churn-scoring.ts that implements the scoring logic as described. The file should export:

• A RetentionScore type or interface
• A calculateRetentionScores function (can use stub/mock data access for database calls)
• A getAtRiskSegments function that groups scored customers into named segments

Also produce a short DESIGN_NOTES.md explaining: 1. How risk thresholds are determined 2. How single-purchase customers are handled 3. Which customers are excluded and why 4. What segments are produced and their definitions

The functions should be complete enough that a reviewer can understand the full scoring and segmentation approach by reading the code and notes.

Automated Retention Intervention Workflows

Problem/Feature Description

UrbanKit is an online streetwear retailer whose marketing team runs several email campaigns manually today. Each week a team member exports a list of lapsed customers and sends them a bulk discount code. The problem: customers who just bought yesterday sometimes receive the email anyway, high-value customers are getting the same 20%-off code as occasional buyers (which has started to erode gross margin), and some customers are getting retention emails on top of a newsletter plus a flash sale notification in the same day.

The engineering team has been asked to replace this manual process with an automated workflow system. The new system should automatically classify customers into risk tiers and dispatch the appropriate intervention — with guardrails to prevent over-contacting customers or sending unnecessary discounts to people who are likely to repurchase on their own.

The team wants the system to run once per day at a specific time and handle at least three distinct intervention types: early-warning customers who are just beginning to slip, high-value customers who are truly at risk, and first-time buyers approaching their expected repurchase window.

Output Specification

Produce a TypeScript file retention-workflows.ts that implements the automated workflow runner described above. Use stub/mock implementations for database calls and email sending — the logic and guard conditions are what matter.

Also produce a WORKFLOW_NOTES.md that documents: 1. When each workflow fires and for whom 2. What each intervention sends (content/parameters) 3. How duplicate sends are prevented for each workflow type 4. How frequently a customer can be contacted (across all workflows) 5. When a pending retention job should be cancelled