How we can work together →

Shopify Consultant

Plan and scale growth for your Shopify store

Fractional Ecommerce Director

Ongoing leadership for next-stage ecommerce growth

Ecommerce Consultant

Hired to solve specific growth challenges

Ecommerce Coaching

1-2-1 mentoring for founders and ecommerce leaders

Conversion Strategy

How to convert more visitors into buyers

Retention Strategy

Grow your ecommerce business on repeat

SEO & GEO Consultant

Become the answer in Google and AI search

1-Day Ecommerce Growth Blueprint

One focused day to map your next stage of growth

Replatform to Shopify

Build your growth machine right from day one

Testimonials

What founders say about working together

Not sure which fits? Let’s talk it through.

“Working with Ian is probably the best thing that has happened to us in the last 6 years.”

Rivki Stoll · Private White VC

Data →

Ecomonomics

The numbers behind your ecommerce growth

Attribution and measurement

Know what is really driving your sales

Technology →

Tools and martech

The tools that power modern ecommerce growth

SEO and AI visibility

Get found in search and in AI answers

Affiliates and partnerships

Collaborations, partnerships and affiliate programmes

Human touch →

Brand and positioning

The insight and stories that make your brand matter

Retention and subscriptions

Growing profitable repeat business

Content and community

Copy, stories and community that build connection

New lessons every weekday · 100s of articles, written by me, not a robot

Explore all growth lessons →

,

How To Optimise Post-Purchase Emails Using Klaviyo’s Predicted Next Order Date

The second order is where the money is

Most post-purchase emails are timed by habit. Thank you on day one, a review request on day seven, a “we miss you” with 10% off on day thirty. Every customer gets the same schedule, whether they’ll be ready to buy again in two weeks or four months.

That costs you twice. Send the nudge too early and it’s noise, or worse, a discount handed to someone who would have come back at full price anyway. Send it too late and they’ve already gone somewhere else. You paid to acquire that customer. The second order is where you start to earn it back, and a schedule that ignores when people actually reorder leaves that second order to chance.

The fix isn’t more emails. It’s sending the one that matters at the point each customer is most likely to buy again, and saving your discounts for the people who need a reason.

Klaviyo already has a prediction for that sitting on your customer profiles. Most brands I look at have never used it.

What Klaviyo’s predicted next order date actually is

Klaviyo is the email platform I work with and the one I recommend. If your account qualifies, its predictive analytics add a set of predictions to each customer profile: predicted and historic customer lifetime value, churn risk, average time between orders, and an expected date of next order.

A few things about that date, straight from Klaviyo’s own help docs, because they decide how you use it:

  • It’s built from the customer’s own order pattern. Where there isn’t a pattern, or not enough data, Klaviyo predicts from how your other customers behave.
  • One-time buyers still get a date. It’s based on data across all your customers. That makes it most useful for exactly the people you most need a second order from.
  • People who’ve never ordered don’t get one. This is a post-purchase tool, not a welcome-flow tool.
  • It ignores what they bought. The prediction doesn’t take the products in the order into account, so it won’t know that a 30-day supply runs out in 30 days.

Check you qualify before you build anything

Predictions only switch on once your account has enough data. Klaviyo’s requirements are: at least 500 customers who have placed an order, an ecommerce integration such as Shopify (or the API sending placed orders), at least 180 days of order history, orders in the last 30 days, and some customers who have placed three or more orders.

The quickest check is to open a customer profile and look for the predictive analytics section. If it says the profile doesn’t qualify, you’re not there yet. Don’t try to force it. Build a simpler time-based flow for now and come back when the data’s there.

Where I’d start

This is the order I’d work through it with any Shopify brand that wants more second orders without throwing discounts at everyone…

1. Decide who the flow is for. Klaviyo itself doesn’t recommend counting down to the expected date for repeat customers, and if you have a lot of repeat buyers it suggests you may want to use it only for people who have bought once. I agree. Your repeat customers already have a rhythm. The one-time buyer is the one at risk, so start there.

2. Leave replenishment where it is. If you sell something with a known cycle, like a 30-day supply, a replenishment flow based on that cycle will beat a prediction that doesn’t know what was in the basket. Klaviyo says the same: this flow shouldn’t replace replenishment flows. Use the predicted date for products where the gap between orders genuinely varies.

3. Build a date-triggered flow on the expected date of next order. In Klaviyo’s own repeat-purchase guide, the flow is date-based, triggered by the expected date of next order, with the repeat set to never. Then add the flow filters that stop it misfiring: exit anyone who has placed an order since starting the flow, and if you run subscription or replenishment flows, exclude people who’ve had one recently so nobody gets two reminders for the same thing.

4. Split on predicted lifetime value before you reach for a discount. Klaviyo’s guide has a neat way to do this. Export your “placed one or more orders” segment with predicted CLV, work out the average, and add a conditional split on predicted CLV being at least that figure. The customers above the line don’t need money off. Give them a reason to come back that costs you nothing: what to try next, a reason to join your loyalty programme, a request for feedback. The customers below the line are where an incentive can earn its keep, and even then I’d try a bundle or a free gift with purchase before a straight percentage off.

5. Make the content about the next purchase, not the last one. The prediction doesn’t know what they bought, so your emails have to. Use what you know about their first order to suggest the natural next product. If you’ve got the data on what customers buy on their second order, use it. My bundles guide covers how to find that.

6. Treat the date as a window, not an appointment. It’s a prediction. Don’t write “time to reorder!” as if you know they’ve run out. Write as if you’re a helpful reminder that turns up around the right time, and give the sequence a couple of messages either side of the date rather than betting everything on one send.

7. Run it for a few months, then judge it on second orders. Look at the second-order rate of one-time buyers who went through the flow, and what it cost you in discounts, compared with how one-time buyers behaved before. If the high-CLV path converts without an offer, you’ve just stopped paying for orders you were going to get anyway.

What to measure

  • Second-order rate for one-time buyers. The main number. If this doesn’t move, nothing else matters.
  • Placed order rate and revenue per recipient by path. Compare the high and low predicted-CLV branches separately, not blended.
  • Discount cost per second order. What you gave away to get each repeat purchase. This should fall on the high-CLV path.
  • Unsubscribes and spam complaints. If these jump, your timing or tone is off, even if revenue looks fine.

The mistakes I see most

Using it on everyone. Repeat customers already have a pattern. A countdown to a predicted date for someone who buys every six weeks like clockwork adds noise, not orders.

Letting it replace replenishment. If you know the cycle, use the cycle. A prediction that doesn’t look at the product can’t beat a known run-out date.

Discounting by default. Sending the same 10% to every customer near their predicted date trains your best customers to wait for it. Split on predicted CLV first.

Forgetting the exit filter. Without the “has placed an order since starting this flow” filter, people who’ve just reordered still get the reminder. It’s the fastest way to look like you’re not paying attention.

Building it before the data is ready. If your account doesn’t meet the thresholds, there’s nothing to trigger on. Get the basics right first.

My verdict

The brands I’m working with right now are fighting the same battles: ad spend that’s crept up faster than the profit, no clear picture of how to scale, and not enough modern ecommerce know-how in the team to get more from the tools they’re already paying for. Post-purchase email is a good example. Most brands have the prediction sitting in Klaviyo and are still sending everyone the same day-30 discount.

If you qualify, a date-triggered flow on the expected date of next order, aimed at one-time buyers and split by predicted lifetime value, is one of the cheapest retention wins going. It uses data you already have and an email platform you already pay for.

It isn’t the right tool for everyone. If you’re under the 500-customer threshold, it won’t switch on. If you sell a consumable with a fixed cycle, a replenishment flow will do the job better. And if almost nobody buys from you twice because of what you sell, no prediction will change that: the work is in the product and the offer, not the timing.

For a Shopify brand with enough order history and a second-order rate that should be higher, it’s where I’d start. You can find out more on the Klaviyo site, and my Klaviyo tool page covers how I use it with Shopify brands.

Related reading: How To Optimise Post-Purchase Klaviyo Flows Using Post-Purchase Surveys and How To Optimise For Business on Repeat: Engineering the Second Purchase. More guides on this problem: A post-purchase experience that doesn’t bring customers back.

More on this lever: The Retention Lever: How To Optimise Repeat Purchase Using Your Own Data.

Ian


Manage The Machine

Get my weekly round up of Ecommerce Growth insights and tutorials every Monday at 9am.

Straight from my client work. No theory. Written by me, not a robot.

Free. Every Monday at 9am. Unsubscribe any time.

Ian Rhodes

Written by Ian Rhodes

Independent Ecommerce Consultant & Coach with 28 years of ecommerce leadership and 750+ brands worked with. Hired to prioritise and lead the work that delivers next stage growth.

More growth lessons