Most conversion projects start with a redesign. This one started with a single bar in the mini cart.
On the Thermos store, we added one element to the mini cart: a free delivery progress bar showing how close the shopper is to the threshold. No redesign. No replatform. We ran it as a controlled test for 12 days, with 25,850 visitors split between the old cart and the new one.
Thermos Mini Cart Test
Conversion rate: +35.74%
Revenue per visitor: +66.41%
Products per order: +23.47%
Confidence: 99.75% across 12 days and 25,850 visitors
Paul Ryazanov · MageCloud · One element, one controlled test, one live store
Why I Publish the Confidence Number
I give the confidence number because most case studies do not. A percentage on its own tells you nothing. Without the sample size and the duration, you are being shown a coincidence and asked to treat it as a strategy. Anyone can find a good week, screenshot it and call it a result.
That is also why I do not recommend testing to every store that asks for it. A small store simply does not have the traffic to reach a number like 99.75% in a sensible amount of time. I wrote about exactly that in why I told a client not to start A/B testing. Before any of this, it is worth knowing where you actually stand, which is why I always ask founders to benchmark their conversion rate first.
Why Being Close to Something Changes Behaviour
People respond to being close to something. Show them they are £4 from free delivery and a good share of them will add the fourth item. The threshold was always there. The bar simply made it visible at the exact moment the shopper was deciding whether to check out.
The numbers show where the extra revenue came from. Revenue per visitor rose by almost twice as much as conversion. Products per order went up by nearly a quarter. The bar did not only persuade more people to buy. It persuaded the people who were already buying to buy a little more.
What This Test Does Not Prove
This is one store, one free delivery threshold and one 12-day window. It does not mean every store should copy the same bar and expect the same lift. A free delivery threshold only works if your margin can carry it, and the right threshold depends on your average basket.
It also does not mean conversion rate is the number to chase. I have argued that conversion rate is becoming a weak metric on its own. What makes this result worth sharing is that revenue per visitor and products per order moved with it. If only conversion had moved, I would be far less interested.
How to Find Your Own Version of This
This result came from the cart, not from more traffic. I have argued before that most stores do not have a traffic problem, they have a conversion problem. Buying more traffic into a cart that does not convert only makes the leak more expensive.
Testing is how you find out which small change matters, rather than guessing. I have been making the case for A/B testing as the cheapest growth lever most stores never pull for more than a decade, and this test is a good example of why.
We audit Magento, Shopify and WooCommerce stores every week. If you want to know what your own version of this is, send me your URL and I will tell you what I would test first.