I’m a solo developer. I built a Shopify app that sends a personalized
email after each order — it writes a unique message per order, flags VIP
and at-risk customers, times the send, and tracks what each email brings
in. It’s approved and listed in the App Store.
What I don’t have yet is real merchants using it. So I’m looking for 5
stores willing to install it and tell me honestly what works and what
doesn’t. Free while we do this.
Repeat-purchase categories fit best — supplements, coffee, pet, skincare,
anything consumable.
To be upfront: it does not do upsells or discount codes, and I’m early
enough that you’d be shaping it rather than using a finished product.
If that sounds useful, reply here and I’ll follow up.
Clear scope and honest disclosure — that’s refreshing. For the five-store test, I’d avoid making opens the success metric. Track support contacts, repeat purchase, unsubscribe/complaints, and a small holdout.
If the app flags at-risk customers, also ask merchants whether that flag changed an action. Otherwise the email may feel personalized without proving incremental value.
Thanks — that’s a sound list, and it lines up with how this is built. Revenue attribution per message is already in the app, which is why opens were never going to be the number I report.
On the comparison group: the assignment is deterministic rather than a stored random draw — a hash of shop plus customer, so someone lands in the same group every time and can’t drift between them. Holdout customers also aren’t counted as sends, so they don’t distort the volume figures or cost the merchant anything.
It’s off by default and switched on per store, deliberately. Withholding email from a slice of someone’s customers is their call to make, not mine, and it has to be said out loud before anything runs.
At this scale I’ll report absolute counts rather than percentages — 6 of 40 against 2 of 40, not “3x”. Percentages on small samples are theatre.
On at-risk: a flag that changes nothing downstream is worthless, agreed. Here it feeds the segmentation and the send timing rather than sitting in a dashboard as a label.
I’ll come back to this thread with real numbers once the first stores have run a cycle.