Every brand I audit has the same pattern. They've invested in the flows their agency recommended. Welcome series, abandoned cart, post-purchase. Then they plateau. Email revenue is flat. Retention is stagnant. And nobody can figure out why.
The answer is almost always sitting right there in their Klaviyo account. Predictive analytics. The feature that tells you which customers are about to churn, when each customer is most likely to reorder, and which customers will generate the most lifetime value. Ignored. Unused. Generating nothing.
This is how to unlock it, what to build with it, and what it means specifically for CPG, wellness, and food and beverage brands.
What Klaviyo's Predictive Analytics Actually Are
Klaviyo's predictive analytics is a machine learning system trained on your store's actual order data. It is not a generic industry model. It learns from the specific purchasing behaviour of your specific customers and generates individual predictions per profile.
These predictions update dynamically as customer behaviour changes. If someone who was on a regular purchase cadence suddenly goes quiet, their churn risk score rises. If a first-time buyer places a second order earlier than expected, their predicted LTV increases. The model adjusts in real time.
To unlock predictive analytics, your account needs:
- →
At least 500 customers who have placed an order
- →
An active ecommerce integration (Shopify, BigCommerce, or Magento)
- →
At least 180 days of order history
- →
Orders placed within the last 30 days
- →
Some customers who have made 3 or more purchases
Once unlocked, the predictions appear directly in each customer profile and become available as conditions in your segment builder and as filters in your flows. There is no additional setup required. The data is already there. Most brands just never touch it.
The 4 Predictions Klaviyo Makes Per Customer
Once active, Klaviyo generates four predictions for every qualifying customer profile. Each one has a distinct use case.
Predicted LTV
The total revenue Klaviyo expects a customer to generate over their lifetime, combining what they have already spent with a forward-looking estimate based on their purchase pattern. This is the number that should be driving your retention spend decisions. A customer with predicted LTV of 400 GBP deserves different treatment to one predicted at 60 GBP. Most brands treat them identically.
Expected Date of Next Order (EDONO)
Klaviyo's prediction of when a specific customer will place their next purchase, based on their individual purchase cadence. For consumable categories, this is highly accurate. A customer who consistently buys a 30-day supplement supply will have an EDONO that clusters around day 25 to 27. Building a reorder reminder flow triggered around the EDONO is one of the highest-leverage retention moves a consumable brand can make.
Churn Risk
The predicted probability that a customer will not place another order in the next 90 days, expressed as Low (under 33%), Medium (33 to 66%), or High (above 66%). This is the prediction most brands sleep on. By the time a customer is classified as lapsed, it is too late. Churn risk lets you intervene while they are still active, before the relationship breaks down.
Predicted Number of Future Purchases
An estimate of how many more times Klaviyo expects this customer to buy. Useful for identifying customers who look like one-time buyers early, so you can target them with conversion campaigns before the window closes. Also useful for validating whether your product mix or subscription options are actually driving multi-purchase behaviour across cohorts.
The 3 Segments Every Brand Needs to Build
Predictive analytics only generates value when you build segments around it. Here are the three foundational segments to create the moment your account qualifies.
Segment 1
High-LTV, High Churn Risk
Predicted LTV above your 75th percentile AND Churn Risk = High
This is your most urgent segment. These are your most valuable customers showing early signs they are about to leave. They have the highest revenue concentration and the shortest window to save them. A typical intervention for this segment is a direct, personalised email from the founder, a time-sensitive offer, or early access to something new. Not a generic win-back. Something that makes them feel seen. This segment should be reviewed weekly without exception.
Segment 2
High-LTV, Reorder Window Approaching
Predicted LTV above your 75th percentile AND Expected Next Order Date within 7 days
These are your VIPs who are about to repurchase. Your job here is not to interrupt them with a discount. It is to meet them at the moment of highest intent and make the experience frictionless. A gentle reminder email, a personalised product recommendation based on their purchase history, or a bundle offer that increases AOV without competing against their existing buying decision. Hit them at the right moment with the right message and you will see revenue without cannibalising margin.
Segment 3
Mid-LTV, Low Churn Risk
Predicted LTV between your 40th and 75th percentile AND Churn Risk = Low
This segment is often overlooked because nothing is going wrong. But these are your upgrade candidates. They buy regularly, they are unlikely to churn, and they are the cohort most likely to respond to subscription offers, loyalty programmes, or bundle upgrades. A small increase in purchase frequency or AOV across this segment compounds significantly over 12 months. Ignoring them because they are not in crisis is a missed growth lever.
5 Flows to Build With Predictive Data
Segments tell you who. Flows tell you when and what to say. These are the five flows that generate the most return from Klaviyo's predictive data, in priority order.
Churn Prevention Flow
Trigger: Customer's churn risk score moves to High. Split by LTV tier on entry. High-LTV customers get 3 emails over 10 days: a founder-voice personalised check-in on day one, a soft product re-engagement with social proof on day four, and a direct offer with urgency on day ten. Mid and lower LTV customers get a shorter two-email sequence with a single offer. The goal is re-engagement before the customer mentally disengages, not recovery after they have already left.
Predicted Reorder Reminder
Trigger: 5 days before Expected Date of Next Order. This is the highest-impact flow for consumable DTC brands. Send a single email that acknowledges what they previously bought, confirms it should be running low, and makes reordering frictionless with a direct add-to-cart link or one-click reorder. For wellness and supplement brands, add a 7-day early bird offer for customers who reorder before the predicted date. This compresses the repurchase window and increases purchase frequency.
LTV Tier Upgrade Flow
Trigger: Customer's predicted LTV crosses from your mid tier into your high tier. This is a behavioural signal that a customer is deepening their relationship with your brand. Acknowledge it. This flow introduces subscription options, loyalty rewards, or a VIP access offer. It turns a growing customer into a committed one. Most brands only think about subscriptions at the acquisition stage. Upselling to subscription post-purchase, when predicted LTV rises, is dramatically more effective.
VIP Retention Flow
Trigger: Quarterly, for all customers in the High-LTV, Low Churn Risk segment. Your safest customers still need to feel valued. This is not a promotional email. It is a relationship touchpoint. Exclusive first look at new products, a thank you from the founder, a community invitation, or early access to a sale. Keeping your best customers feeling recognised is what maintains their low churn risk over time. The brands that lose their VIPs are usually the ones who only email them when they have something to sell.
One-Time Buyer Conversion Flow
Trigger: Predicted number of future purchases is zero or one AND it is 30 days post-first-purchase. Klaviyo will flag customers who look like they are unlikely to return. Act early. This flow gives you a concentrated window to convert a first-time buyer into a repeat one. Use social proof from customers with similar first purchases, address the most common objection for non-repurchasers in your category, and present a clear second-purchase incentive. Customers who make a second purchase within 60 days of their first are 3x more likely to become long-term buyers.
Why This Matters More for CPG and Wellness Brands
Klaviyo's predictive models are effective across ecommerce, but they are especially powerful for consumable categories: supplements, food and beverage, skincare, cleaning products, and wellness. Here is why.
Consumable products have predictable repurchase windows driven by usage rate. A customer buying a 60-capsule supplement bottle is almost certainly going to run out around day 30. A customer who buys a bag of premium coffee will need to reorder within 2 to 3 weeks. This behavioural regularity makes the Expected Date of Next Order prediction extremely accurate for CPG brands, far more than for, say, a fashion or home goods store where repeat purchases are driven by inspiration or occasion, not necessity.
This means the predicted reorder reminder flow is not a nice addition for consumable brands. It is a core revenue engine. Brands that get this right and send the reminder email 5 to 7 days before the predicted reorder date see materially higher open and conversion rates than standard campaign sends, because they are landing in the inbox at the exact moment of highest intent.
Every Man Jack, a personal care brand, adjusted their replenishment flow to trigger on each customer's predicted next order date instead of a fixed schedule. The result was a 25% year-over-year increase in flow revenue. The product did not change. The audience did not change. The timing changed.
For wellness and F&B brands specifically, the churn risk prediction is also more meaningful than in other categories because the health and wellness purchase cycle has a natural attrition point. Customers buy at peak motivation and fade as the habit breaks. Churn risk rises before any revenue signal appears. Catching it early, at the churn risk threshold rather than at the last-order date, is the difference between recovering a customer and losing them permanently.
The 3 Mistakes Most Brands Make
Building win-back instead of churn prevention
Win-back flows fire after a customer has lapsed. Churn prevention fires while they are still active. These are fundamentally different interventions. A customer with High churn risk who receives a well-timed email is 4 to 6x more likely to respond than a customer who has already been inactive for 90 days. The difference is acting on the prediction, not waiting for the outcome. Most brands build win-back because it is the default recommendation. Churn prevention is the higher-leverage play.
Treating all predicted churners the same
A customer with predicted LTV of 600 GBP and High churn risk should receive a very different intervention to a customer with predicted LTV of 40 GBP and High churn risk. The first one deserves a direct, personal, high-effort outreach. The second may not be worth the margin cost of an aggressive offer. Build conditional splits in your churn prevention flow so that the level of intervention scales with the value of what you stand to lose.
Ignoring the data after the initial build
Predictive segments are not static. Klaviyo updates them continuously as customer behaviour changes. A customer who drops from Low to High churn risk in a single week needs an immediate intervention, not one that triggers on the next scheduled campaign. Set up real-time segment monitoring and ensure your flows are configured to trigger within 24 hours of a churn risk change. The speed of response matters as much as the message.
What This Looks Like in Practice
I worked with a wellness supplement brand doing around 85,000 GBP a month. Their Klaviyo was active. Welcome series, abandoned cart, basic post-purchase. Email was generating about 18% of total revenue, which looked fine on paper.
When we audited the account, predictive analytics was fully unlocked. The data had been sitting there for eight months. Nobody had built a single segment from it.
We built three segments: High-LTV High Churn Risk, Reorder Window, and Mid-LTV Low Churn Risk. We added five flows. Within 45 days, email revenue was up to 29% of total revenue without a single new customer acquired. The churn prevention flow alone recovered 14 customers in its first month, representing over 6,200 GBP in predicted LTV.
The platform did not change. The audience did not change. What changed was that we stopped sending emails on a schedule and started sending them when the data said to.
Klaviyo Audit
Find Out What Your Klaviyo Account Is Leaving on the Table
I will review your predictive analytics setup, your segment structure, and your active flows, then show you exactly where the revenue gaps are and what to fix first. No agency pitch. No fluff. Just a clear roadmap for what to build next.
Book Your Klaviyo AuditFrequently asked questions
What is Klaviyo's predicted LTV?
Klaviyo's predicted LTV is a machine learning model that estimates the total revenue a customer will generate over their lifetime, combining their historical spend with a forward-looking prediction based on their purchase behaviour. It updates dynamically as behaviour changes and is available as a segment condition to let you target customers by lifetime value tier.
How do I access predictive analytics in Klaviyo?
Klaviyo's predictive analytics unlock automatically once your account meets the data requirements: at least 500 customers who have placed an order, an active ecommerce integration, at least 180 days of order history, orders within the last 30 days, and some customers with 3 or more purchases. Once unlocked, predictions appear in each customer profile and are available as segment conditions and flow filters.
What is Klaviyo's churn risk prediction?
Klaviyo's churn risk is the predicted probability that a customer will not place another order in the next 90 days, expressed as Low (under 33%), Medium (33 to 66%), or High (above 66%). It updates continuously based on purchase frequency, email engagement, and time since last order. You can segment by churn risk to intervene before customers lapse.
What is the expected date of next order in Klaviyo?
The expected date of next order (EDONO) is Klaviyo's prediction of when a customer is most likely to place their next purchase, based on their individual purchase cadence. For consumable DTC brands like supplements, wellness products, and food and beverage, this is especially accurate because purchasing is driven by consumption rate. Brands that build reorder reminder flows around the EDONO see average flow revenue increases of 25% or more year over year.
How is Klaviyo's predicted LTV different from historical LTV?
Historical LTV is simply the total revenue a customer has already spent. Predicted LTV combines that with a forward-looking estimate of future spend. Historical LTV tells you what happened. Predicted LTV tells you what is likely to happen. Acting on predicted LTV lets you allocate retention spend to customers with the highest future value, not just those who have already spent the most.
Can small DTC brands use Klaviyo's predictive analytics?
Yes, but you need to meet the minimum data threshold: at least 500 customers who have placed an order, 180 days of order history, and some customers with 3 or more purchases. Brands below this threshold will not see predictive data. The most practical early-stage approach is to focus on clean Shopify and Klaviyo integration, consistent event tracking, and strong first and second purchase flows so that when you hit the threshold your data is accurate enough for reliable predictions.
About the author
Caner Veli built Liquiproof from zero to 3,000+ global retailers in under 6 years. He now helps DTC and CPG brands fix broken growth engines and scale 2x-15x in 90 days.