App Marketplace Review Outreach
We find merchants who said in a Shopify App Store review that a tool isn't working for them, build them the fix for their store and send it with a short how-to video.
119 reviewers found, 45 custom demos built, 25 videos recorded, 16 sent.
TL;DR
- We scrape reviews on the Shopify App Store to find merchants who publicly said a tool isn't working for them.
- Each reviewer is matched to their store, gets a custom demo built for their store and a short how-to video, and only then a personal message.
- 119 reviewers found, 45 custom demos built, 25 videos recorded, 16 messages sent so far.
- Honest status: building and sending work. Automatic reply tracking never ran reliably, so that is the next fix.
The idea
Cold outreach usually asks for something: a call, a reply, a budget. We flipped it. Merchants tell the whole world what's broken in their app reviews: "the flows are confusing", "I never got my welcome emails working". So we show up with the fix already built, and the message is a delivery note rather than a pitch.
How it works
- 1Scrape the marketplace. We collect reviews of an app on the Shopify App Store (we started with Klaviyo): stars, text, date and how long the merchant has used it.
- 2Segment the reviewers. Claude sorts every reviewer into Detractor, Mixed, New user, Established fan or Power user. New users and detractors are the people we can help most.
- 3Match review → store → contact. Each review is matched to the actual store and a contact, with a confidence score (high, medium, low) and a note on how the match was made.
- 4Build the asset first. A Klaviyo demo for that exact store, plus a short how-to video recorded with our pitch video engine.
- 5Draft, approve, send. A personal email naming the one thing we changed, with the demo and the video. A person approves every batch before anything goes out.
A second source: Shopify stores running Gorgias
The same "build first" approach also runs on a list of 23,500 Shopify stores that use Gorgias. A scoring script keeps stores of the right size with a real mobile number in nearby or English-speaking markets, which leaves about 1,000 qualified leads. We rebuild their welcome email as an interactive version, record a 90-second walkthrough and send it by WhatsApp or SMS, always with an opt-out line.
Under the hood
| Part | Detail |
|---|---|
| Lead store | Notion database: store, app, stars, segment, review text, match confidence, demo link, video, email, sent at, follow-ups, outcome |
| Status flow | New → Enriched → Needs Review → Demo Built → Draft Ready → Approved → Sent → Replied (or Skip) |
| Message rules | Name the actual change and one number, never two. No "I'd love to". Opt-out in every message. Nothing is sent before the asset exists. |
| Compliance | Human approval per batch, opt-out, low volume |
Results so far
| Metric | Value |
|---|---|
| Reviewers scraped (Klaviyo app, Aug 2026) | 119 |
| Reviewers matched to a store | 113 (49 with high confidence) |
| Custom Klaviyo demos built | 45 |
| How-to videos recorded | 25 |
| Emails sent (with follow-ups) | 16 |
| Gorgias store list scraped / qualified | 23,500 → ~1,000 |
What we learned
- Building works; closing the loop didn't yet. The daily reply checker that should move replies into the CRM never ran reliably, so replies were found by hand. Lesson: build reply detection before you send.
- Most reviewers are happy. 78 of 119 gave five stars. Filtering to new users and detractors first saves a lot of demo work.
- Contact data is the bottleneck. 45 reviewers had no findable email.
Next steps
- A daily reply check (Gmail and WhatsApp) that creates a card in Artikuno OS.
- More apps beyond Klaviyo: Omnisend, Privy, Smile.io, Mailchimp and Seguno are already in the schema.
Connected projects
- Demo and video engine from → Upwork Demo Factory and Personalised pitch videos
- Replies land in → Artikuno OS
Images.
Ready to build your AI project?
Free 30-minute discovery call. No pitch. Just a conversation about what AI could do for you.
- Free · 30 min
- A concrete first step you can act on
- In German or English, online