Find what to give your new customers in the first 30 days
find something specific for them
70 to 77% of customers never buy again after their first order.
Of the customers who do come back, the timing is specific. About half of all second orders happen within 30 days of the first one.
Wait past 45 days with nothing sent, and the odds of a second order fall from somewhere around 15 to 20% down to 3 to 5%.
This is based on different ecom studies.
Many merchants have the same welcome flow for every first-time buyer.
Product-specific content in this window is what actually cuts returns and buyer's remorse, since confusion about a new product is a common trigger for both.
A generic thank you doesn't address it - whoever bought a skincare product and whoever bought a health supplement get the same email, if anything.
With Claude, you can find specific things to send different types of buyers.
And here’s a prompt that I used to do the same.
The exact prompt:
Pull my first-time buyers who ordered in the last 60 days. Group them by what they bought, by product or category, whichever gives cleaner groups. For each group, name the one thing about that product most likely to confuse a first-time owner or lead to a return, say what a touchpoint should say about it, and recommend when to send it - a few days after delivery for most products, or timed to when a consumable or subscription item would run out. Flag it if fulfillment data isn't available so the timing counts from the order date instead.
Run it directly from here → Open in Claude
What comes back:
Every first-time buyer from the last 60 days, grouped by what they bought
The one likely point of confusion or return risk for each group, specific to what they bought
What the touchpoint should say about it
When it should go out - a few days after delivery for most products, or tied to when a consumable or subscription item runs out, counted from delivery instead of the order date
Anyone already past their window with nothing sent yet, flagged separately
Why a loyalty program is the wrong first move
A loyalty program is the wrong place to start with someone who just bought once. It takes 6 to 12 months to show any real movement in lifetime value or repeat purchase rate - time a brand-new customer isn't going to sit around for.
Timing does more work in the first 30 days than a reward will do.
Most post-purchase flows count day 4 to 5 from the order date, off a calendar. That breaks the moment shipments run late: the product-education email shows up before the box does. So base the timing on the fulfillment event instead.
Customize for your specific product type:
For something durable or a bigger purchase, make the content around how to get the most out of it rather than when to reorder.
If you track return reasons, ask Claude to pull those instead of inferring the likely confusion point from the product listing alone - a real reason is better.
If you already run a welcome series for new customers, ask Claude to build around it instead of duplicating what's already going out. You can also ask to turn any group into a Klaviyo flow entry point. It can even write back to your account, if needed.
One Claude Skill for all Shopify operations
This runs on The Agentic Operator, a Claude skill I built for operators covering catalog, PDP content, segmentation, retention, ads, analytics, and more.
Create a project in Claude for your Shopify operations
Add Shopify connector (Settings → Connectors) and authenticate
Klaviyo, same way
Download the skill file - Get it here
Add it under Customize → Skills → Add → Upload a skill
Type /the-agentic-operator, then ask anything
- Ankit
If this was useful, the next one will be too.
Weekly. Free.
Unsubscribe with one click.
