Why US shoppers weren't converting
Fewer than 3 in 1,000 US visitors bought. I found out why.
of visitors bought
lower order rate after forced login
ranked fixes with a metric each
Fewer than 3 in 1,000 US visitors bought over three months.
Analytics showed where shoppers dropped, but not why, or which fix would pay off first.
Sole researcher, starting from the stakeholder brief: goals, customer, channels and known pain points.
Traced most of the loss to product pages with no way forward and a checkout that punished logging in. Handed over six ranked fixes.
A fine-jewellery brand, a lot of ad traffic, almost no orders
The brand sells fine jewellery to the Indian diaspora in the US and wanted to grow significantly. Nearly 9 in 10 visitors were new, most arriving on mobile from social ads.
Ads brought traffic but almost nobody bought. The team needed to know where people were dropping and which fixes would pay off first.
Numbers for where, walkthroughs for why
To start from the goal, the customer and the numbers the team already watched.
To land as a shopper would, from Facebook, Instagram, Google and the India site, and see what ad visitors saw.
Funnels show where people drop and where tracking is blind, so I cut them by source, landing page, device and login state.
To tie every screen from landing to payment to its tracking event, and mark the friction where it happened.
To test the menu, search and filters against Baymard guidance, not just my opinion.
Where the funnel leaked
Step conversion over 3 months, rounded. Payment itself worked; the loss sat on the product page and between starting checkout and entering an address.
Logging in at checkout cut orders by 20×+
- Cart
- Guest formtax and coupons hidden
- Address
- Payment
- Cart
- Login formno exit; cart cleared
- Back to browsingre-add the item
- Cart
- Sign up
- Profile pagedead end
Logging in at the cart cut the order rate by more than 20× compared with shoppers already signed in. Part of the gap is self-selection: signed-in shoppers were mostly returning customers.
Every path, walked and annotated
What I found, and what I recommended
- Cart
- Login formno exit; cart cleared
- Back to browsingre-add the item
- Cart
- Login, cart keptwith an exit
- Checkoutaccount auto-created for new buyers
- Keep shoppers in checkout after login, keep the cart, add an exit; auto-create accounts for new buyersThe 20×+ order-rate gap
- Show cart summary, state tax, coupons and points upfront for guestsDrop between checkout start and address
- Breadcrumb and "view similar" on product pages, A/B tested2 in 3 product-page exits
- "Save this design", price-drop alerts and a recently viewed stripAnonymous returning visitors
- Regroup navigation to 6–9 primary items13 mixed categories
- Fix tracking: missing events, test traffic, personal data in URLs, UTM namingBlind spots in the data
I handed the team a ranked fix list, an annotated journey map and a log of tracking gaps, plus the metrics to judge each fix by: checkout completion for signed-in versus guest shoppers, product-page exits, and sign-ins on return visits.
Interview recent US buyers and drop-offs to test the trust gaps the data hinted at: delivery dates, sizing, and reviews from Indian rather than diaspora buyers.
Have a funnel that leaks?
Open to remote and hybrid roles, and contract projects.