When the traffic arrives and the money doesn’t
The version of this that gets sold is testing — headlines, button colours, a six-week programme. The version that’s usually true is that something on the path to checkout is mechanically broken and nobody has walked it.
The examples are unglamorous and they repeat: a size guide that opens over the buy button on a mid-size phone. A cart that empties on back-navigation. A discount field that rejects the code in the email you just sent. A search that returns nothing for the word your customers use for your own product.
None of these show up as a bug report, because the customer doesn’t know it’s broken — they assume it’s them, or they assume you’re out of stock, and they leave. In analytics it reads as a drop-off, which reads as a traffic quality problem, which is where the budget goes.
Walk your own paths
On your phone, as though you’d never seen the store. Not on desktop, and not as the person who knows where everything is.
Search for one of your products using the word a customer would use, not the name in your catalogue. If it returns nothing, that’s revenue.
Add something to the cart from a collection page, not a product page. Then hit back. Then go to checkout, and hit back again. Broken back-navigation on a cart is one of the most common untested paths there is.
Apply a discount code from your most recent email. All the way through, on the phone.
Then open your analytics and find the biggest drop between two steps — and walk that step slowly. What you find is usually mechanical, not persuasive.
What the work involves
Walking the paths, on real devices, including the browsers inside Instagram and TikTok if that’s where your traffic comes from. These behave differently and almost nobody tests them.
Reading the analytics for where people actually stop — then going and standing at that step rather than theorising about it.
Fixing what’s mechanically broken first. This is most of it, and it’s the part with the clearest line to revenue.
Then the judgment calls — the account wall before checkout, the shipping cost that appears three steps too late, the form asking for things you don’t need. Stated as recommendations with reasoning, because some of them are business decisions rather than mine.
A before and after on the paths that changed, so the work has an answer rather than an impression.
What this isn’t
It isn’t a testing programme. I’m not going to run six weeks of A/B tests on a store without the traffic to resolve them — most stores this size get a bad number and a false conclusion. If you have the traffic for real testing, that’s a different engagement and worth being explicit about.
It isn’t copy or creative. I’ll tell you when a page is asking for too much too early. I won’t write the offer.
It isn’t a promise of a percentage. Anyone who quotes you an expected uplift before seeing your store is quoting you an average from someone else’s store.
And it isn’t a substitute for a product problem. If the thing isn’t selling because of price, positioning or demand, a better checkout gets you a smaller version of the same result.
What the standard says about this
- The conversion paths, walked — Whether the paths that carry money were walked or assumed.
- The browsers your customers actually use — Whether the tested browsers match the ones in your analytics.
- Checkout, end to end — Whether the checkout completes, not merely renders.
Two ways to find out
The checks above are published and free. Run them yourself, or send me the store and I'll run them for you.