This is the single most common revenue leak I find when I audit a DTC email account. Not bad creative, not poor subject lines, not the wrong sending frequency. It is the absence of any meaningful distinction between the people on the list. One audience, one message, sent to everyone, every time.
AI-powered customer segmentation changes that. Klaviyo already has the capability built in. Shopify is feeding it the data. Most brands have it switched off, or built segments so complex they became unmaintainable and got abandoned. This is the guide to doing it in a way that actually generates revenue.
What AI Segmentation Actually Does (and How It Differs from Rule-Based Logic)
Traditional email segmentation works on rules. You build a list of conditions: bought in the last 90 days, opened the last three emails, average order value above a certain threshold. You apply those rules, create a segment, and send to it. The problem is that rules reflect the past. A customer who bought four months ago and is about to buy again looks identical to a customer who bought four months ago and has gone permanently cold, until you wait long enough to confirm which is which.
AI segmentation in Klaviyo uses machine learning trained on your order history to predict what is likely to happen next. It calculates each customer's expected date of next order, their predicted lifetime value, and their current churn risk score. These are not static flags that you set once. They update continuously as your customers' behaviour changes.
The practical difference is timing. Rule-based segmentation tells you what happened. Predictive segmentation tells you what is about to happen. A customer whose churn risk just moved from low to medium is not lapsed yet, but they are heading there. That is the window to intervene, not the window that opens after they have already gone.
The Three Tiers That Move Revenue
You do not need fifty segments. You need three tiers that are clean, maintained, and each receiving a message that actually matches where that customer is in their relationship with the brand.
Champions
5-10% of your list. 30-40% of your email revenue.
Champions are your best customers. They bought recently, they buy repeatedly, and they spend the most. In Klaviyo's RFM framework, these are customers scoring in the top tier across Recency, Frequency, and Monetary value. The platform calculates this automatically.
The mistake most brands make with Champions is treating them like everyone else. They send them the same broadcast campaign with the same 15% off code that goes to the entire list. Champions do not need a discount. They are already bought in. What they respond to is recognition: early access to new launches, a personal note from the founder, an invite to leave a review or refer a friend, access to something that feels exclusive.
A single well-crafted Champions campaign, with no discount, typically outperforms a sitewide promotional email on click rate and revenue per recipient. The segment is small. The upside is disproportionate.
Active Buyers
Your current base. Focus here is on frequency and AOV.
Active buyers are customers who have purchased at least once in the last 90 to 120 days and have engagement scores above your account average. They are not Champions yet, but they are not at risk either. They are the segment with the clearest path to increasing revenue without increasing the customer count.
The lever here is Klaviyo's predicted next order date. If a customer typically repurchases every 45 days and they are now at day 40, the platform knows. A replenishment nudge timed to that window, referencing their specific previous purchase, converts at a significantly higher rate than a generic promotion. You are not guessing at timing. You are using their own purchase pattern to tell you when to send.
Cross-sells and bundle offers belong here too. Customers who have purchased one product line and have a moderate to high predicted CLV are the right audience for an introduction to a complementary product, not your coldest subscribers who are still deciding whether the brand is worth a second look.
At-Risk and Lapsed
The segment most brands reach too late.
This is where most DTC brands run their win-back campaigns, and most run them reactively, ninety or a hundred and twenty days after someone stops engaging. By that point, reactivation rates are low. The window has passed.
Klaviyo's churn risk scoring changes the timing. When a customer's churn risk moves from low to medium, that is the moment to send a proactive win-back, not a final attempt. Brands that trigger win-back sequences at the medium churn risk stage recover 12 to 18% more at-risk customers than those using fixed time-based triggers. The message changes too. At medium risk, you do not need a panic discount. A direct email acknowledging they have not been back, reminding them what they bought, and offering something relevant is usually enough. Save the heavy offer for confirmed lapsed customers.
RFM in Klaviyo: Setup and What to Do with Each Tier
Klaviyo calculates RFM scores automatically across your connected Shopify data. You do not need to configure the scoring model. You need to configure what happens to each tier.
Inside Klaviyo, go to Segments and create a new segment using the "Predictive Analytics" property group. You will see RFM tier as a selectable condition. Build a segment for each tier, then assign each to a different campaign cadence. Champions receive fewer emails, higher exclusivity. Loyal customers and potential loyalists receive your standard campaign frequency. At-risk customers move into an automated win-back flow triggered by the churn risk condition rather than a fixed calendar date.
The configuration takes about thirty minutes. The part that takes longer is being honest about what each segment should hear. Champions should not receive the promotional campaign you built for cold subscribers. At-risk customers should not receive a new product launch before they have been won back. Each tier needs its own message. That clarity is what produces the revenue lift.
The most common mistake brands make with Klaviyo segmentation is building too much at once. A complex structure with twenty conditions that takes six weeks to configure loses to a simple three-tier setup that is live this week. Complexity without activation is just planning.
The Micro-Segment Trap: When Simple Beats Complex
There is a seductive logic to hyper-precise segmentation. If three tiers are good, surely thirty are better. Customers who bought on a Friday afternoon and engage with your weekend content deserve a Friday-specific segment. Customers in a specific geography with a particular product preference warrant their own slice.
In practice, micro-segmentation delivers diminishing returns quickly, and it creates a maintenance burden that most DTC teams cannot sustain. When a segment is too small to reach statistical significance, you cannot measure whether it worked. When a segment structure is too complex to maintain, it gets abandoned. Neither outcome generates revenue.
The practical rule: start with three predictive tiers. Measure the revenue per send for each tier over sixty days. Add complexity only when you can measure the incremental lift from each additional layer. A brand generating forty percent more revenue per send from three clean segments does not need a fourth segment until those three are optimised. Most brands are not there yet.
What This Looks Like in Practice
A wellness supplement brand I worked with was running a single monthly broadcast to 18,000 subscribers. Open rate was 22%. Click rate was 0.8%. Email revenue was flat despite a growing list.
We built three segments in Klaviyo using predictive analytics. Champions (roughly 1,800 subscribers) received a dedicated new product preview sequence with a personal message from the founder and a referral ask. Active buyers received a replenishment flow triggered by Klaviyo's predicted next order date for their primary product. At-risk customers received a two-email proactive win-back, triggered at medium churn risk rather than a fixed ninety-day lapse.
Within sixty days, email revenue per send had increased by 34%. No change to creative quality. No increase in campaign frequency. No new tools or platforms. The improvement came entirely from sending the right message to the right group at the right time using data that was already in Klaviyo.
Inside the system
How we build this for brands
When we take a brand on at Purposeful Profits, the email segmentation work begins by pulling Klaviyo and Shopify data into a VOC engine that mines customer reviews and purchase behaviour to identify the language patterns, buying triggers, and objections most common in each segment. Champions have different concerns than at-risk customers. The messaging for each tier should reflect that gap. We then build and deploy lifecycle flows in Klaviyo across welcome, replenishment, win-back, and post-purchase journeys, each tied to the relevant predictive segment rather than fixed time delays. The flows are live from day one. Measurement runs weekly via a profit and cash-flow dashboard that surfaces email revenue per send, segment performance, and churn recovery rates so the operator can see exactly what each tier is generating.
Part of this runs live for portfolio brands today. The full system is what we design and deploy when we take a brand on, covering segmentation architecture, flow builds, copy, and ongoing performance reporting through a single weekly dashboard rather than a stack of disconnected metrics.
Klaviyo Audit
Find out which segments are carrying your revenue
Caner reviews your Klaviyo account, identifies your highest-value segments, maps the gaps between what you are sending and what each tier actually needs, and shows you the exact flows and campaigns to build first.
Book Your AuditFAQ
Common questions
What is AI customer segmentation for DTC brands?
AI customer segmentation uses machine learning to group your customers by behaviour, purchase patterns, and predicted future actions rather than static rules. In Klaviyo, this includes Predictive Analytics features like churn risk scores, predicted next order dates, and lifetime value predictions. Instead of segmenting by 'bought in the last 90 days', AI segmentation identifies customers who are likely to buy again within a specific window, or flags customers whose purchase probability is dropping before they lapse.
How do I set up AI segmentation in Klaviyo?
Klaviyo's Predictive Analytics is already inside your account if you're on a paid plan with sufficient data (at least 500 orders). Navigate to Segments, create a new segment, and you will see predictive properties available as filters: predicted CLV, churn risk level, expected date of next order, and predicted number of purchases. Start by building three segments: Champions (high CLV, low churn risk, recent purchase), Active Buyers (moderate CLV, purchase in last 120 days), and At-risk (rising churn risk or no purchase past predicted window). These three tiers alone will change the revenue your email channel generates.
Can I build AI segmentation myself without an agency?
Yes. Klaviyo's predictive segmentation features are built in and do not require a developer or a data scientist. The setup is done entirely inside the Klaviyo interface. What takes time is not the technical configuration but the strategy: deciding which segments to build first, what to send each tier, and how to measure the revenue lift. Most operators who try to do this alone build too many segments too quickly, which creates complexity without proportional lift. The better approach is three clean tiers, activated this week, iterated monthly.
What is RFM segmentation and does Klaviyo do it automatically?
RFM stands for Recency, Frequency, and Monetary value. It scores each customer on when they last bought, how often they buy, and how much they spend in total. Klaviyo automatically calculates RFM tiers and labels them: Champions, Loyal Customers, Potential Loyalists, At-risk Customers, and others. You do not need to configure the scoring. You need to configure what to send each tier. Champions respond to VIP treatment, early access, and referral asks. At-risk customers respond to time-limited offers and direct messaging. Sending the same email to both groups is the single most common revenue leak in DTC email channels.
How much revenue lift can I expect from proper email segmentation?
Segmented email campaigns generate approximately 30% more opens and 50% more click-throughs than non-segmented sends, according to HubSpot's 2025 State of Marketing Report. AI segmentation improves revenue per user by 10 to 20% on average. The lift varies significantly by how segmented your current sending is. Brands moving from one broadcast list to three active tiers (Champions, Active, At-risk) consistently see email revenue per send increase by 25 to 40% within 60 days without changing creative quality. The improvement comes entirely from sending the right message to the right person at the right time.
What is churn risk in Klaviyo and should I use it?
Churn risk in Klaviyo is a predictive score that indicates the probability of a customer not purchasing again, based on their historical behaviour patterns. Klaviyo categorises customers as low, medium, or high churn risk. Brands that activate churn risk as a segment trigger and send proactive win-back sequences to customers when they reach medium churn risk (rather than waiting until they fully lapse) recover 12 to 18% more at-risk customers compared to time-based win-back triggers. It is already in your account. If you are not using it, you are running win-back campaigns six weeks late.
About the author
Caner Veli built Liquiproof to global distribution across 3,000+ retailers, then exited. He now runs Purposeful Profits using a combination of operator strategy and AI-powered systems he has built and uses daily, having 10x'd monthly revenue in his own business in the last 90 days. He works with DTC and CPG brands in supplements, skincare, wellness, and food and drink. Work with Caner.