Advertise to 38,864 E-Commerce Businesses
Promote your tool, service or upcoming event - Reach Founders
Good morning!
In today’s newsletter,
Our very own DTC Tool Stack
Start 2026 with a stronger retention strategy
Why your win-back flow should vary by cancellation reason
How to create ads using your best customer reviews
AI Launch Codes: Why AI Struggles With Creative Idea Generation
This issue takes 3 minutes to read.
Let’s dive into it👇
DTC Tool Stack
We’re introducing our very own curated ecom tool stack, one every DTC team should have.
👉 Includes tools for almost anything you’re struggling with.
Retention
Start 2026 with a stronger retention strategy
Want to improve your retention strategy in 2026?
We've partnered with Flowium to offer a free 30-minute lifecycle consultation.
Get an expert review of your email/SMS flows and actionable tips to increase conversion and customer lifetime value.
Retention Recipe
Why your win-back flow should vary by cancellation reason
Subscribers don't all cancel for the same reason. And yet most brands send everyone the same "We miss you! 20% off" email.
Someone who cancelled because of price needs a different message than someone who has too much product stacked up.
Your subscription app already captures why people cancel. Route that data into Klaviyo as an event.
Then build ONE win-back flow with dynamic blocks that change based on the cancellation reason.
Examples:
"Too expensive" → Offer smaller pack sizes or show ROI
"Not seeing results" → Share customer wins and usage tips
"Have too much product" → Let them pause or adjust delivery frequency
"Switching to a competitor" → Remind them what they'll lose
Start with your 2-3 most common cancellation reasons and keep one generic message for everyone else.
How to set it up:
Route cancellation reasons from your subscription app (Recharge, Skio, etc.) into Klaviyo as custom events
Build one flow
Add dynamic blocks that show different content based on the cancellation reason
Post Credit: Jeremy Horowitz
Copy That Converts
How to create ads using your best customer reviews
Your best ad copy already exists.
It's sitting in your reviews and comments section.
The key is to make it look native, and not like an ad.
It should look like content someone would naturally scroll past.
Why it works:
People trust other customers more than they trust brands. When someone sees a real review instead of a polished marketing copy, they actually believe it.
Plus, you don't have to guess what resonates. Your customers already told you what they love in their own words.
How to do it:
Find your best reviews (ones that mention specific results or emotions)
Screenshot them
Put on a simple, clean background
Make sure it looks native to the platform (not overly designed)
Run as an ad
Why AI Struggles With Creative Idea Generation
Large language models like ChatGPT and Claude are remarkably good at summarizing, rewriting, and executing well-defined tasks.
But they consistently struggle with creative idea generation.
When asked to brainstorm, they tend to produce outputs that are safe, repetitive, and structurally similar.
Even with:
Different prompts. Different brands. Different products.
You get the same ideas.
A new research paper titled
“Divergent–Convergent Thinking in Large Language Models for Creative Problem Generation” finally put a name to this behavior.
The researchers show that modern LLMs tend to collapse creativity because they try to satisfy all requirements immediately.
They call the result something many of us already feel in practice:
The Artificial Hivemind.
How Most of Us Brainstorm Ideas With AI
Here’s how most e-commerce teams brainstorm ad angles for Meta ads using AI.
We’ve been taught that the “right” way to prompt follows a structure like:
Role – who the model should pretend to be
Context – background, examples, or brand details
Task – what we want it to do
Outcome – how the output should look
So I used this structure to prompt AI
Consider yourself to be a senior performance writer with over 20 years of experience in creating high performing meta ads for your clients.
Create 20 high performing meta ad angles for this product Obvi Collagenic Fat Burner.
https://myobvi.com/products/obvi-collagenic-fat-burner?srsltid=AfmBOooGDyb3hm4KZ72edy8BxTNsmbXivGVOr0ho94rGK5ebM1aJUvH_&selling_plan=405504261&variant=32376605278257At first glance, this feels logical.
Clear role. Clear task. Clear performance goal.
So what happens when you run it?
You get outputs that are:
familiar
safe
eerily similar to what you’ve seen before
Here is what I got.
What’s Actually Going Wrong
The paper explains that creativity in humans happens in two separate phases.
First, we use divergent thinking — generating many possible ideas without judging them.
Then we use convergent thinking — narrowing those ideas into a single, workable answer.
That separation is critical for creativity.
Large language models don’t do this by default.
When prompts include roles, goals, and constraints upfront such as high-performing, improves ROAS, lowers CAC, improves conversion rates — the model immediately enters convergent thinking, trying to satisfy every requirement at once.
The paper shows that this early convergence is what collapses creativity.
The model isn’t failing to be creative. It’s being asked to decide before it’s allowed to explore.
Using the Paper’s Insights, Here’s What Changed
Here’s what the next 20 ad angles looked like.
The first set followed familiar patterns.
These don’t.
Let’s Break Down the Process
This post is for paid subscribers
We built AI Launch Codes to help you scale smarter. We test AI tools, build workflows, and write prompts—so you don’t have to. But testing takes time, money, and effort. This paid upgrade helps us keep experimenting, while you get the winning playbooks delivered straight to your inbox.
UpgradeWhy You’ll Want It:
- Use AI to scale faster, cut costs, and boost efficiency
- Ready to use workflows, automations and prompts
- Real world use cases, no theory