To increase repeat purchases, fix your shipping and returns terms first, then build the 3 automated flows that ask for the next order. Personalization comes after both. It is the last 10% of the job, and most guides sell it as the first.
To increase repeat purchases, fix your shipping and returns terms first, then build the 3 automated flows that ask for the next order. Personalization comes after both. It is the last 10% of the job, and most guides sell it as the first.
What the data says, against what the guides say
Search this question and every result gives you the same ending: personalize the experience, launch a loyalty program, use AI recommendations.
One page on that results list asked actual shoppers. Mollie surveyed 10,000 European consumers on what makes them buy from a store again. The top factor was free shipping, with 89% calling it crucial to the decision. Free returns came fourth.
Not personalization. Not points. Shipping.
That should change what you do on Monday. A shipping threshold is a decision you can make this afternoon. An AI recommendation engine is a project.
Know your number before you change anything
Published averages for ecommerce repeat purchase rate land around 28%, and 20% or above is widely treated as normal. Alex Schultz, formerly of Facebook, put the healthy range at 20 to 30% of customers returning monthly.
Treat all of those as rough. They mix subscription boxes with furniture brands, and they rarely say what window they measured. Your own number from last year is a better comparison than any of them.
Measure it the same way every time:
Take everyone who placed a first order in one month.
Count how many placed a second order within 90 days.
Divide.
Pick 90 days and never change it. Changing the window moves the number more than anything you build.
1. Fix the terms before the tactics
Free shipping over a threshold. Free or cheap returns. A reorder that takes 2 clicks, not a fresh checkout.
None of this is clever. All of it outranks personalization in what shoppers say drives them back. If your shipping is $8 on a $40 product, no email sequence fixes that. Every guide telling you to segment your list is skipping the reason people are not returning.
Work out the threshold that keeps your margin intact, then say it plainly on the product page rather than at checkout.
2. Build the post-purchase flow
This is the single highest-value automation for repeat purchases, and most brands do not have one.
The window between order placed and order delivered is the only stretch where a customer thinks about you daily. They are tracking the package. They are looking forward to it. Most stores fill that window with a shipping confirmation and nothing else.
That window is where the second order gets set up. Not sold, set up. Show them how to use the thing properly. Tell them what people usually buy next and why. Make the reorder path obvious before they need it.
3. Time the replenishment flow to the product, not the calendar
If your product runs out in 6 weeks, the email lands at week 5.
This is the flow where timing beats copy by a wide margin. A perfectly written reorder email that arrives 3 weeks late loses to a plain one that arrives the day someone notices the jar is nearly empty.
Work out the real consumption cycle for your top 5 SKUs. Not the one on the label, the one your reorder data shows. Then set the delay to just before it.
4. Cross-sell the obvious next product
Pick the product people buy second, and offer that one.
You do not need a recommendation engine for this. Export your order data, find the customers with 2 or more orders, and look at what the second order was. On most stores 3 or 4 products account for the majority of them. That list is your cross-sell logic, and it took an afternoon.
5. Win back the ones who stopped
Someone who bought 3 times and then went quiet is worth more attention than someone who never bought at all.
Define quiet in terms of your own cycle. For a coffee brand that might be 60 days. For a jacket brand it might be a year. Then send 3 emails, and do not lead with a discount. If your winback opens with 20% off, your best customers learn that going quiet is profitable.
6. Now personalize
Everything above works without knowing anything about an individual customer. It works on timing, terms, and product logic.
Personalization is what you layer on once those exist. Product recommendations based on what someone bought. Replenishment timing set per SKU rather than per store. Different winback copy for a 1-time buyer and a 5-time buyer.
It is real and it works. It is also the refinement, not the foundation, and building it first means personalizing a system that is not sending anything useful yet.
You do not need a CDP
Most articles on this topic will tell you to centralize your data in a Customer Data Platform.
At $1M to $10M in revenue, you do not need one. Your Shopify order history and your Klaviyo profiles already hold everything the 5 steps above require: what someone bought, when, how often, and whether they opened the last email.
A CDP solves a problem that appears when you have several disconnected systems and a team arguing about which one is right. Buying it earlier means paying to unify data you could query in an afternoon, while your post-purchase flow still does not exist.
Where this sits
Repeat purchases are what the whole retention system is for. The flows named above are 5 of the 12, and they only work in order. Building the cross-sell flow before the post-purchase flow means recommending a second product to someone who has not finished being impressed by the first one.
Get the whole system
12 flows, 37 emails, one job: repeat purchases. Every trigger, every send delay, every exit condition, written out so you can build them this month.