Predicted CLV in Klaviyo: how it works and how to build a segment

How do I predict customer lifetime value for ecommerce customers?

If you use Klaviyo, you can use predictive analytics to estimate how much a customer may spend over the next year. The result, called Predicted CLV, can help you find customers who may become high-value buyers, decide where retention effort is worth the investment, and build dynamic segments for follow-up.

You don't need to calculate the prediction yourself. Klaviyo builds the model from the order data in your account and retrains it at least once a week. The useful part is what you do with the signal: look at patterns across groups of customers, then choose an experience that fits the group.

What is Predicted CLV?

Predicted CLV is Klaviyo's estimate of how much a customer may spend in the next year.

It sits alongside two related values:

• Historic CLV: the value of a customer's past orders, taking refunds and returns into account.

• Predicted CLV: the amount Klaviyo estimates the customer may spend in the next year.

• Total CLV: Historic CLV plus Predicted CLV.

For example, a customer with a Historic CLV of $400 and a Predicted CLV of $100 has a Total CLV of $500.

The distinction matters. Historic CLV describes what already happened. Predicted CLV helps you plan for what may happen next. They answer different questions, so don't use one as a substitute for the other.

How Klaviyo calculates Predicted CLV

Klaviyo doesn't use a simple ecommerce formula such as average order value multiplied by purchase count and customer lifespan. It trains a model on order behavior in your account, then uses those patterns to estimate future spend.

That means the quality of the prediction depends on the quality of the purchase data. If you send Placed Order events through a custom integration or the Klaviyo API, include the actual order value in the $value field. Missing or incorrect values can affect the result.

What data does predictive analytics require?

Predictive CLV is available when your account meets the requirements for predictive analytics. In general, you need:

• At least 500 customers who have placed an order. This means purchasing customers, not total or active profiles.

• An ecommerce integration, such as Shopify, BigCommerce, Magento, or WooCommerce, or Placed Order events sent through the API.

• At least 180 days of order history.

• At least one order within the last 30 days.

• At least some customers who have placed three or more orders.

These requirements help the model identify repeat-purchase patterns. If predictive analytics appears on a profile but the fields are blank, Klaviyo may not have enough information about that individual to make a prediction.

How should you interpret Predicted CLV?

Treat Predicted CLV as a planning signal, not a promise. Some customers will spend more than predicted, and some will spend less. The number becomes more useful when you evaluate a segment or cohort instead of making a major decision from one customer's value.

For example, adding the Predicted CLV values for the current members of a segment gives you an estimate of the revenue that group may generate over the next year. It doesn't guarantee that revenue, and it doesn't include customers you may acquire later.

Use the signal to answer practical questions:

• Which existing customers may become future VIPs?

• Which customers may justify a higher-touch retention experience?

• How much potential future revenue is associated with a current audience?

• Which customers may be better suited to a lower-cost re-engagement approach?

Predicted CLV and Churn Risk Prediction also answer different questions. Predicted CLV estimates future spend. Churn Risk Prediction estimates how likely a customer is to stop purchasing. Used together, they can help you prioritize where to act first.

How to build a Predicted CLV segment

1. Confirm that predictive analytics is available

Check that your account meets the requirements above. If you don't see Predicted CLV in the segment builder, your account may not qualify yet, or the model may not have enough data.

2. Open the segment builder

In Klaviyo:

1. Go to Audience > Lists & segments.

2. Select Create New.

3. Select Create segment.

Segments are dynamic. They grow as profiles meet the conditions and shrink when profiles no longer meet them.

3. Add the predictive analytics condition

Add this condition:

Predictive analytics about someone > Predicted CLV

Choose an operator and enter a value. For example, to create an audience of customers predicted to spend at least $500 in the next year, use:

Predictive analytics about someone > Predicted CLV > is at least > 500

The right threshold depends on your business. Look at your customer-value distribution, average order value, margins, and campaign goal before choosing a number. A threshold that works for one brand may be meaningless for another.

4. Add audience and engagement conditions when needed

A CLV segment can include profiles who aren't eligible to receive a message. A segment isn't consent. If you plan to send email or SMS, add the relevant subscription or consent conditions with an AND connector.

You can also narrow the audience by purchase or engagement behavior. For example:

• Predicted CLV is at least $500.

• AND someone has placed an order at least once over all time.

• AND the person can receive email marketing.

• AND the person has opened an email in the last 90 days.

This separates the audience you want to analyze from the audience you can contact right now.

5. Name and create the segment

Use a name that explains both the signal and the time frame, such as High predicted CLV - next 12 months. Create the segment, then review the profiles before you send a campaign. Large segments may take time to populate.

Segment ideas you can use

Future VIPs

Use a higher Predicted CLV threshold to find customers who may become high-value buyers. Depending on your business, you could test:

• Early access to a product launch

• A loyalty or rewards invitation

• Premium product recommendations

• Cross-sell or upsell messages based on purchase history

Don't automatically lead with a discount. Customers who already show strong predicted value may respond better to relevance, access, or recognition.

Lower predicted-value re-engagement

Use a lower threshold to find customers who may need a more efficient retention approach. You might test:

• A win-back campaign

• Product education

• Lower-priced or entry-level recommendations

• A controlled incentive for engaged customers who haven't purchased recently

If you use this audience for email, add engagement and deliverability conditions so you aren't sending to an unengaged or suppressed audience.

Value and risk together

Combine Predicted CLV with Churn Risk Prediction to create a simple prioritization framework:

• High Predicted CLV + high churn risk: prioritize personalized retention outreach.

• High Predicted CLV + low churn risk: protect the relationship with VIP treatment, cross-sell, or early access.

• Low Predicted CLV + high churn risk: use scalable, lower-cost re-engagement.

• Low Predicted CLV + low churn risk: maintain the relationship with regular post-purchase or replenishment messaging.

This is where behavioral segmentation gets useful: the value forecast gives you one signal, while purchase and engagement behavior add context.

How to estimate a segment's potential revenue

To estimate the next-year revenue represented by a current segment:

1. Create a segment of the customers you want to evaluate.

2. Export the segment with Predicted CLV included.

3. Add the Predicted CLV values together.

The sum is an estimate of the future spend associated with the current members of that segment. For a related example of using customer-value fields in an export, see this Community discussion about exporting CLV data.

Common mistakes to avoid

Treating one prediction as a guarantee

Use individual values to prioritize and group-level results to plan. Review how the segment performs over time before changing your strategy.

Trusting incomplete order data

For custom integrations, check that each Placed Order event contains the actual order value. Your prediction can only be as useful as the data behind it.

Ignoring your product's replenishment cycle

General purchasing predictions don't know the exact replenishment window for the product a customer last bought. If your products have predictable replenishment timing, use product-specific replenishment flows alongside the broader CLV signal.

Treating a segment as permission to send

Add the channel-specific consent or subscription conditions before sending. A profile can meet a value threshold and still be suppressed or unsubscribed.

Setting a threshold once and forgetting it

Customer behavior, prices, product mix, and business goals change. Review the segment size and performance regularly, and adjust the threshold when it no longer separates audiences meaningfully.

FAQ

Is Predicted CLV the same as lifetime value?

No. Historic CLV describes what a customer has already spent. Predicted CLV estimates what they may spend over the next year. Total CLV combines the two.

Is Predicted CLV the same as average order value?

No. Average order value describes an individual order. Predicted CLV estimates a customer's future spend over the prediction period.

Can I create a Predicted CLV segment for any account?

Only when the account meets the predictive analytics requirements, including enough purchasing customers, order history, recent order activity, and repeat-purchase data. A Community example about CLV for WooCommerce covers the same qualification issue.

Why is my Predicted CLV segment empty?

Check that your account qualifies, that the segment uses the Predictive analytics about someone condition, and that the threshold fits your customer-value distribution. Also confirm that your order events contain accurate value data.

Can I change the prediction period?

The standard predictive analytics experience describes Predicted CLV as expected spend over the next year. If your account includes customizable CLV functionality, you may be able to adjust the prediction window to match your business or purchase cycle. Changing the window changes the meaning of the value and any segments built from it.