Playbook Overview
| # | Workflow | Revenue Impact | Setup Time | Difficulty | Priority |
|---|---|---|---|---|---|
| 1 | Abandoned Cart Recovery | 10 to 15% cart recovery rate | 3 to 5 hours | Easy | Start Here |
| 2 | Post-Purchase Upsell | 8 to 15% AOV increase | 3 to 4 hours | Easy | Start Here |
| 3 | AI Review Generation | 3 to 5x review volume | 2 to 3 hours | Easy | Start Here |
| 4 | Dynamic Retargeting | 3 to 5x ROAS on retargeting | 4 to 6 hours | Medium | Week 2 |
| 5 | Inventory-Based Ad Pausing | Eliminate wasted ad spend | 4 to 8 hours | Medium | Week 3 |
| 6 | Customer Winback | 5 to 12% reactivation rate | 3 to 5 hours | Medium | Week 3 |
| 7 | Predictive Segmentation | 20 to 40% higher email revenue | 5 to 8 hours | Advanced | Month 2 |
Workflow 1: Abandoned Cart Recovery
The AI-Optimized Cart Recovery Sequence
| Step | Timing | Channel | Message | AI Logic |
|---|---|---|---|---|
| 1 | 30 min after abandon | SMS | Reminder with product image and direct cart link | Only send SMS if customer has opted in and cart value > $50 |
| 2 | 1 hour after abandon | Personalized subject line, product images, social proof (reviews) | AI selects subject line variant based on past open behavior | |
| 3 | 24 hours | Urgency message: low stock, price drop alert, or trending product signal | AI chooses urgency type based on real inventory and price data | |
| 4 | 48 hours | Email + SMS | Incentive offer: free shipping, 10% off, or gift with purchase | AI determines minimum incentive needed based on customer LTV tier |
| 5 | 72 hours | Final reminder with stronger incentive or alternative product suggestions | AI recommends similar products if original items are out of stock |
Smart Incentive Tiering
Do not offer the same discount to every customer. Segment by LTV tier: high-LTV customers often convert without any discount (just a reminder), mid-LTV customers respond to free shipping, and low-LTV or first-time visitors may need 10 to 15% off. This approach protects margins while maximizing recovery.
Tools and Setup
| Component | Recommended Tool | Alternative | Cost |
|---|---|---|---|
| Email/SMS automation | Klaviyo | Omnisend, Drip | $45 to $150/mo |
| Cart tracking | Shopify / WooCommerce (native) | BigCommerce, Magento | Included |
| AI subject lines | Klaviyo AI / Phrasee | ChatGPT for manual generation | $0 to $500/mo |
| Personalization | Klaviyo Predictive Analytics | Nosto, Dynamic Yield | $0 to $500/mo |
Workflow 2: Post-Purchase Upsell Sequence
| Step | Timing | Channel | Content | AI Logic |
|---|---|---|---|---|
| 1 | Immediately (checkout) | On-site popup | One-click upsell: complementary product at 10 to 20% off | AI selects product based on purchase + collaborative filtering |
| 2 | 30 min post-purchase | SMS | "Your order is confirmed! Add [product] before we ship?" | AI picks highest-converting cross-sell for this product category |
| 3 | 3 days post-purchase | Product education + "complete the set" cross-sell | AI segments by new vs. returning customer, adjusts messaging | |
| 4 | 7 days (post-delivery) | How-to content + replenishment reminder or accessory upsell | AI predicts replenishment timing based on product consumption rate | |
| 5 | 14 days post-purchase | VIP offer or loyalty program invitation (if high LTV signal) | AI scores customer LTV probability, only triggers for high-value |
Protect the Post-Purchase Experience
The post-purchase window is fragile. Never upsell before the order ships if you have fulfillment delays. If shipping takes more than 3 days, delay step 3 until after delivery confirmation. A customer waiting for a late package does not want to buy more.
Workflow 3: AI Review Generation
| Step | Timing | Action | AI Logic |
|---|---|---|---|
| 1 | Delivery confirmed + 3 days | Send review request email with product photo and one-click star rating | AI determines optimal send time based on customer timezone and past email engagement |
| 2 | Delivery + 5 days (if no review) | SMS follow-up with direct review link and small incentive (loyalty points, 5% next order) | Only triggers if customer opened the email but did not click |
| 3 | After review submitted | If 4 to 5 stars: auto-publish, ask for photo/video. If 1 to 3 stars: route to support team | AI sentiment analysis gates reviews before publication, flags negative experiences |
| 4 | Ongoing | AI generates review response drafts for approval | Personalized responses based on review content, product, and customer history |
| 5 | Monthly | AI identifies top-reviewed products for UGC campaigns and ad creative | Aggregates review data to surface highest-performing products and customer quotes |
| Review Platform | Best For | AI Features |
|---|---|---|
| Yotpo | Mid to large brands | AI review request optimization, sentiment analysis, UGC curation |
| Judge.me | Budget-friendly option | Automated requests, photo reviews, AliExpress import |
| Stamped.io | Growing brands | AI-powered smart reminders, NPS integration, loyalty program |
| Okendo | Premium brands | AI attributes, customer profiles, zero-party data collection |
| Loox | Visual/photo reviews | AI photo review collection, video reviews, discount incentives |
Workflow 4: Dynamic Retargeting
| Audience Segment | Trigger | Ad Content | Bid Strategy | Expected ROAS |
|---|---|---|---|---|
| Product viewers (high intent) | Viewed product 2+ times or added to wishlist | Exact product images + reviews + urgency (low stock) | Aggressive: highest CPM willing to pay | 5 to 8x |
| Cart abandoners | Added to cart but did not purchase | Cart products + incentive ladder (free shipping > 10% off) | High: second priority | 4 to 7x |
| Category browsers | Browsed category but no product view | Best sellers from that category + social proof | Medium: moderate CPM | 3 to 5x |
| Past purchasers | 30 to 60 days since last purchase | Complementary products + "customers also bought" | Medium: focus on repeat purchase | 3 to 6x |
| Lapsed customers | 90+ days since last purchase | New arrivals + winback offer | Low: testing re-engagement | 2 to 4x |
Feed Quality Is Everything
Dynamic retargeting is only as good as your product feed. Ensure every product has high-quality images, accurate pricing, real-time stock status, and complete attributes (color, size, category). Use DataFeedWatch or Feedonomics to optimize your feed before running dynamic ads.
Workflow 5: Inventory-Based Ad Pausing
| Stock Level | Action | Automation Logic | Tool |
|---|---|---|---|
| In stock (50+ units) | Normal bidding and promotion | No changes needed | Standard ad platform |
| Low stock (10 to 50 units) | Add urgency messaging: "Only X left" | AI updates ad copy/overlay with real stock count | Shopify Flow + Google Merchant Center |
| Very low (under 10 units) | Reduce ad spend 50%, shift to waitlist capture | Auto-reduce CPC bids, add "notify me" landing page | Custom script or Feedonomics rules |
| Out of stock | Pause all paid promotion immediately | Auto-pause Shopping ads, pause retargeting for this product | Shopify Flow + Google Ads API / Meta API |
| Back in stock | Resume ads + send restock notification to waitlist | Auto-enable ads, trigger "back in stock" email/SMS blast | Klaviyo back-in-stock flow + ad platform API |
Workflow 6: Customer Winback
| Stage | Timing | Channel | Message | AI Logic |
|---|---|---|---|---|
| At-risk detection | 60 days no purchase (or 2x avg. purchase cycle) | Internal alert | Flag customer in CRM as at-risk | AI compares individual purchase frequency vs. cohort average |
| Soft re-engagement | 60 to 75 days | "We miss you" + new arrivals in their preferred category | AI selects products based on past purchase category and browse history | |
| Incentive offer | 75 to 90 days | Email + SMS | Personalized discount: 15 to 20% off or BOGO on favorites | AI calculates minimum incentive needed based on customer price sensitivity |
| Last chance | 90 to 120 days | Stronger offer + "your preferences have been saved" messaging | AI determines whether to increase incentive or suggest alternative products | |
| Sunset | 120+ days with no engagement | System | Suppress from regular marketing, move to quarterly win-back cadence | AI monitors for any re-engagement signal to reactivate immediately |
Workflow 7: Predictive Segmentation
| Segment | Definition | Marketing Strategy | Expected Lift |
|---|---|---|---|
| High predicted LTV | AI scores top 20% by future revenue potential | VIP treatment: early access, exclusive offers, loyalty program fast-track | 25 to 40% higher retention |
| Likely to purchase (7 days) | AI identifies customers with 70%+ purchase probability this week | Light-touch nudge: new arrivals email, no discount needed | 15 to 25% higher conversion |
| Price sensitive | AI detects customers who only purchase on sale | Target with flash sales and clearance, exclude from full-price campaigns | 20 to 30% more efficient promotions |
| At-risk of churn | AI predicts customers likely to lapse based on engagement patterns | Proactive winback before they fully lapse (see Workflow 6) | 10 to 20% churn reduction |
| Category enthusiasts | AI clusters customers by predicted next-purchase category | Category-specific campaigns with personalized product recommendations | 20 to 35% higher email click rates |
| Segmentation Tool | Best For | AI Capabilities |
|---|---|---|
| Klaviyo Predictive Analytics | Shopify and e-commerce brands | Predicted LTV, predicted next order date, churn risk, gender prediction |
| Nosto | On-site personalization | AI product recommendations, segment discovery, A/B testing |
| Dynamic Yield (Mastercard) | Enterprise personalization | Predictive targeting, affinity engine, deep learning segments |
| Bloomreach | Large catalogs (1,000+ SKUs) | AI search, merchandising, content personalization, segment prediction |
| Retention.com | Identity resolution + re-targeting | Identify anonymous visitors, predict purchase intent, enrich profiles |
Implementation Roadmap
6-Week Implementation Plan
- Week 1: Abandoned Cart Recovery. Build and launch the 5-step cart recovery sequence. This is the highest-impact, lowest-effort workflow. Most brands recover 10 to 15% of abandoned carts within the first month.
- Week 2: Post-Purchase Upsells + Review Generation. Set up the post-purchase upsell sequence and review request automation simultaneously. These two workflows are complementary and share timing logic.
- Week 3: Dynamic Retargeting. Optimize your product feed, set up audience segments, and launch dynamic retargeting campaigns. Focus on cart abandoners and product viewers first.
- Week 4: Inventory-Based Ad Pausing. Connect your inventory management system to your ad platforms. Set up stock-level thresholds and automated pause/resume rules.
- Week 5: Customer Winback. Build the winback sequence with at-risk detection, re-engagement emails, and escalating incentives. This workflow needs 60+ days of data to trigger, so launch it early.
- Week 6: Predictive Segmentation. Enable Klaviyo Predictive Analytics (or your segmentation tool of choice). Build segments based on predicted LTV, purchase probability, and churn risk. Start using these segments for all campaign targeting.
Measure Before Scaling
Run each workflow for at least 14 days before optimizing. AI models need data to learn, and premature changes prevent the algorithms from converging. Set a baseline, wait, measure, then iterate. The exception is obvious errors (wrong products, broken links) that should be fixed immediately.
Methodology & Sources
Workflows developed from e-commerce client implementations, Klaviyo and Shopify Flow documentation, industry case studies, and performance data from DTC brands doing $1M to $50M in annual revenue.
- Klaviyo, "E-commerce Benchmarks Report," 2025
- Shopify, "State of Commerce Report," 2025
- Baymard Institute, "Cart Abandonment Rate Statistics," 2025
- McKinsey, "The Value of Personalization in E-commerce," 2024
- Yotpo, "State of E-commerce Reviews," 2024
- Dynamic Yield, "Personalization Maturity Study," 2024
Frequently Asked Questions
Most e-commerce brands implementing the full playbook see a 15 to 30% lift in total revenue within 90 days. The biggest gains come from abandoned cart recovery (recovering 10 to 15% of abandoned carts) and post-purchase upsells (adding 8 to 15% to AOV). Stores doing $1M+ annually typically see $150K to $300K in additional revenue from these automations.
The core stack is Klaviyo (email/SMS), your e-commerce platform (Shopify, WooCommerce, or BigCommerce), and a reviews platform (Yotpo, Judge.me, or Stamped). For advanced workflows, add a personalization engine (Nosto or Dynamic Yield) and a feed management tool (DataFeedWatch or Feedonomics).
Plan for 4 to 6 weeks total. Start with abandoned cart recovery (week 1), then post-purchase upsells (week 2), review generation (week 3), and layer in the remaining workflows over weeks 4 to 6. Each workflow takes 2 to 5 hours to build and test.
Yes, but prioritize differently. Stores under $500K should focus on abandoned cart, review generation, and post-purchase upsells first. These three workflows alone can add 10 to 20% in revenue. Save inventory-based ad pausing and predictive segmentation until you have enough data (usually 1,000+ orders).
Not if you set up proper frequency caps and suppression rules. Each workflow in this playbook includes recommended timing and suppression logic. The key is relevance: a well-timed abandoned cart email with the right product feels helpful, not pushy. Monitor unsubscribe rates, and if any flow exceeds 0.5% unsubscribe rate, adjust timing or content.