Test run: 26 women's clothing boutiques, 7 drafts worth sending

Before asking anyone to pay for MAALI, we used it ourselves. This is a real campaign from our own testing, with real numbers. Store names are left out because those businesses didn't sign up to be in a case study.

One honest caveat: we haven't sent these emails yet, so this case study is about research quality and cost, not reply rates. We'll update it when we have reply data.

The offer

The campaign was for AbrimaAI, our sister product: an AI content tool for apparel stores that puts garment photos on AI models in different poses and scenes, without a photoshoot.

The ideal customer, as written in the campaign:

Small and mid-size Shopify apparel brands that add products often and rely on flat-lay, hanger or plain photos, or have few on-model images. Not a fit for stores that already have rich professional on-model photography for every product.

Filters: US only, at least 5 products, a product added in the last 180 days, minimum fit score 40.

Finding stores

We used Find stores with the focus "women's clothing boutiques". MAALI searched the web, found 23 candidate stores and checked each one: is it a live Shopify store, does it have a public catalog, does it publish a business email? 20 passed and were added to the campaign. Together with a handful of stores we had pasted in by hand earlier, the campaign had 26 stores.

What the free checks skipped

7 of the 26 stores were skipped before any AI research, at no cost:

That's more than a quarter of the list ruled out for free. Without the email check alone, four emails would have had nowhere to go.

What the AI research found

The remaining 19 stores were researched: MAALI read their catalogs, photos, apps and pages, and scored each against the ideal customer.

• 12 scored below 40 (average 25) and were set aside. Every one of them already had professional on-model photography, often five to nine images per product, which is exactly the "not a fit" we described. A generic lead list would have sent them all the same email.
• 7 scored 40 or higher and got drafts. The six waiting for review averaged a fit score of 52; the one we approved scored 80.

The drafts fell into two patterns. The strongest fits had the photo gap the offer is built for: the top store (fit 80) showed only flat-lays and hanger shots, and another had a single image on 49% of its 402 products. The rest were very active catalogs, with hundreds of products and new ones added daily, where producing images is a constant workload. Each draft opened with what MAALI saw on that store, naming real products, then offered a free sample for one product.

What it cost

At today's pricing, this campaign would use:

• 3 credits for Find stores (one search, which added 20 stores)
• 19 credits for research (1 per researched store; the 7 skipped stores were free)

That's 22 credits for 7 researched, qualified drafts, or about $5.80 on the Growth plan. The time spent was mostly reading the drafts.

What we learned

1. The "not for" half of the ideal customer matters most. Writing who the offer isn't for is what let MAALI score 12 strong-looking boutiques as poor fits.
2. Most of the list shouldn't be emailed. Only 7 of 26 stores were worth a message. That's normal for a specific offer, and it's the point: fewer, better emails.
3. The email check pays for itself. Stores without a published address are skipped for free, before research.

Try it on your own offer

The Free plan includes 10 store researches with your first 3 drafts unlocked, so you can see what MAALI finds for your service before paying anything.