I'm building AdTestLab — a tool that takes Shopify founders' daily Meta + Shopify numbers and gives them a verdict: kill, continue, or which

Here's the strange part: it requires manual data entry.

In a market where everyone's chasing automation (Triple Whale, Lifetimely, every dashboard that auto-syncs Meta + Shopify), I deliberately went the other way. Not because I couldn't build the integrations — but because the integrations don't actually solve the problem.

Here's what I noticed during my own product testing:

1. Meta and Shopify never agree on order count. Meta uses 7-day click attribution and double-counts. Shopify is ground truth. Triple Whale doesn't fix this — they just show both messy numbers.

2. The decision "should I kill this test?" rarely fails because data was wrong. It fails because the founder didn't actually look at the data.

3. Course-trained dropshippers already track daily — usually in a Google Sheet. The friction isn't manual entry; it's not having a clear framework around the numbers.

So I built a tool that:

- Forces you to enter the numbers (5 min/day)

- Shows you what's actually happening in the funnel

- Gives a rule-based verdict (CPA target, click-through health, funnel stage breakdown)

- Layers an AI diagnosis from Claude on top for nuanced reads

Free during beta. I want to validate whether the "manual is intentional" positioning resonates with real users, or if I'm being stubborn about a real problem.

Stack: Next.js + Firebase + Claude API. Built solo over ~2 weeks with Claude Code as my dev assistant.

Live: https://ad-test-lab.vercel.app/

Would love feedback from anyone who's run Meta tests on Shopify. Especially curious to hear if "manual entry as a feature" lands or feels like a cop-out.