The most expensive bug is the one nobody reports
A broken page gets reported. A checkout that quietly loses orders does not: the customer leaves, returns to Google, and buys from someone else. Analytics records a visit, not a lost sale, so it never shows up as an error. In our archive the e-commerce work is mostly this: stores with traffic and poor conversion, where the fix was small and the revenue effect was not.
Cause 1: a surprise at the last step
The classic. Delivery, tax or a handling fee appears only at the final step, after the customer has entered everything. This is the single most common cause of abandoned checkouts. Show shipping and tax earlier — in the cart, not at the last click — and the abandonment falls away. It is a small change with a disproportionate effect.
Cause 2: required fields nobody should be required to fill
Account creation, a phone number, a company field, a postcode format the customer does not recognise. Every field is friction and every field costs completions. Guest checkout should exist; optional should genuinely mean optional. Validation should be helpful rather than a wall of red text at the end.
Cause 3: mobile friction
A checkout that works on a laptop and fights on a phone is losing the majority of traffic, because most traffic is on a phone. Test it on a real phone on mobile data — not a desktop browser resized. The specific offenders are usually: date pickers that fight the keyboard, autoplaying video, and tap targets that are slightly too small. See our Core Web Vitals post for the speed side of the same problem.
How to find out which one is yours
- Place a real order on a real phone, end to end, paying with a real card.
- Watch where you hesitated. That hesitation is your top cause.
- Check the cart-to-checkout step count. Six screens to buy a $20 item is too many.
- Ask the five customers who bought recently what nearly stopped them.
The honest version: this is usually a few hours of work, not a rebuild. Our reviewed project history shows what that work looks like when it is done, including the reviews where it went less smoothly than planned.
Next step
Tell us roughly how many orders a week you should be getting and roughly how many you are getting. That gap is the thing worth fixing, and we will tell you if the fix is small: When did a customer last tell you checkout was difficult — and did anyone believe them?