Before Diwali I ran an automated shopper (CartBeam) through 75 independent Indian Shopify stores: open a product, pick a variant, add to cart, check the cart, and open checkout on a phone-sized screen. It stops before address and payment and never places an order. This is a small, one-time sample, not a market survey, and it only covers one product path per store on mobile.
1. What we could test
- 54 of 75 (72%) passed all five steps.
- 15 (20%) could not be tested: bot protection, sold-out products, or a layout the checker does not recognise. We report these as "could not check", not as broken.
- 6 (8%) were flagged by the automated run. I checked them by hand: three simply require a customer login before checkout, two had a removed product, one I could not verify. None was a real checkout failure that I could prove. So the honest count of broken checkouts I can confirm is zero.
2. Which checkout is behind the store
Detected from the page, n = 75.
- GoKwik: 32 (43%)
- Unknown or custom checkout: 21 (28%)
- Native Shopify checkout: 7 (9%)
- Shiprocket Checkout (Fastrr): 6 (8%)
- Shopflo: 6 (8%)
- Razorpay Magic Checkout: 3 (4%)
3. What I got wrong the first time
My first run flagged 17 of 40 stores as broken. I re-ran each and got the same failure twice, then checked eight by hand: all eight were fine. The checker was calling bot challenges, sold-out products and unrecognised layouts "failed". After I gave those their own result, the first 40 became 36 passed, 4 could not be checked, 0 flagged. Reproducing a failure twice is not verification; a person looking once is.
4. What this does not tell you
It does not cover theme-only script breaks, payment authorisation, address and shipping steps, desktop, or any time other than the moment of the check. A pass is one path at one time, not a guarantee.
If you run a Shopify store and want one product path checked on mobile for free, paste a public product link at https://cartbeam.infinityore.com.
Top comments (0)