Cart Drawer vs Add-to-Cart Banner: +16% Conversion and Two Opposite Device Results
[ SUMMARIZE WITH AI ]
[ FREE CRO TEARDOWN ]
Find the 3 biggest revenue leaks on your store.
Every day a conversion leak goes unfixed, you're paying for traffic that doesn't buy. Get a 5-minute Loom through your PDP, cart, and checkout, with mockups of the fixes. No pitch.
Get My TeardownEveryone tests what goes inside the cart drawer. Almost nobody tests whether the drawer should be there at all.
We did. A high-ticket equipment brand was using their theme's default add-to-cart notification: a small banner that appears, confirms the item, and disappears. We replaced it with a full cart drawer. Line items you can edit, a free shipping progress bar, trust badges, checkout button in the footer.
Six weeks. 84,005 visitors. 347 orders.
Conversion rate rose 16%. Revenue per visitor rose 18%.
Then we split the data by device and found the average was hiding two completely different stories.
Why We Questioned the Cart Drawer Itself
Look at cart drawer advice online and it is all interior decorating. Add a free gift meter. Add trust badges. Add upsells. Move the price onto the button.
All of that assumes the drawer exists.
Plenty of Shopify themes still ship a notification banner instead. Item added, here is a confirmation, back to browsing. It is not broken. It is just quiet. The shopper never sees their cart, never sees the total, never sees how close they are to free shipping.
Our take going in: on a store where people buy one expensive thing, the banner is fine. Nobody is building a basket. There is nothing to show them.
We were wrong about that, and the way we were wrong is the useful part.
The Test: Banner vs Drawer on a High-Ticket Store
Component: Add-to-cart interaction, sitewide
Platform: Intelligems
Test type: Control plus one variation
Duration: 6 weeks
Traffic: 84,005 visitors, roughly 50/50 split
This store is not typical DTC. Average order value sits around $700 and conversion rate runs near 0.4%. That is normal for the category. High consideration, long research cycle, few orders relative to traffic. It matters for reading the numbers below, because 347 orders across six weeks is a thin sample no matter how much traffic you send at it.
Control: the notification banner
Theme default. Add to cart, a banner confirms it, the shopper carries on. The header cart icon goes to the cart page. Checkout happens from there.
Variation: the cart drawer
Be clear about what this actually was. Not one change. Four, shipped together:
- The drawer itself, full-screen on mobile so the background page is gone entirely. Editable line items with quantity controls, and a checkout button straight to checkout that skips the cart page. The header cart icon opens the drawer instead of navigating away.
- A free shipping progress bar with a $100 threshold and a meter that fills as the cart grows.
- Visible savings per line item: discount percentage, amount saved, and the original price struck through, plus a subtotal row showing the discounted total against the original.
- Three trust signals in the footer covering customer service, warranty, and after-sales support.
Product recommendations were deliberately left out. That was the one thing we refused to bundle, because an upsell block would have made the revenue result meaningless.
Bundling four changes is a tradeoff we made on purpose. The brand needed a decision about the drawer, not a six-month research programme. The cost is that this test tells you the bundle works and cannot tell you which part did the work. Keep that in mind when you read the device split below, because we think one of these four explains most of it.
The Headline Result, and Why It Is Not the Whole Story
| Metric | Banner | Drawer | Change |
|---|---|---|---|
| Conversion rate | 0.382% | 0.444% | +16% |
| Revenue per visitor | baseline | +18% | +18% |
| Average order value | baseline | +1% | flat |
Now the part most case studies leave out.
This test reached 86% to 92% probability of beating control, depending on the metric. That is below the 95% threshold most teams use to call a winner. The confidence interval on conversion rate ran from -6% to +44%. On revenue per visitor, -12% to +59%. Both ranges still include zero.
So this is a directional result, not a proven one. The brand shipped it anyway, and we agreed with the call. When the effect is consistently positive across two independent metrics, the mechanism is obvious, and the change costs nothing to maintain, waiting another six weeks for a decimal point is a worse decision than shipping.
We are telling you the confidence number because the alternative is pretending 16% is settled fact. It is not.
Two Devices, Two Different Mechanisms
Here is where the average lies.
| Metric | Mobile | Desktop |
|---|---|---|
| Conversion rate | +26% | +6% |
| Average order value | -21% | +33% |
| Revenue per visitor | -1% | +41% |
Same drawer. Opposite behaviour.
On mobile, the drawer removed friction. A lot more people completed a purchase, 26% more. But they bought less per order. Our read: the banner forced a trip to the cart page, and that trip was doing accidental work. It gave people a full screen to reconsider, review, and add. The full-screen drawer replaced that trip with something faster and more focused, which is exactly what it was designed to do. More buyers, moving quicker, each carrying a smaller basket. Revenue per visitor came out flat.
On desktop, the drawer grew the basket. Conversion barely moved, only 6%, but order value rose 33% and revenue per visitor rose 41%. Our read: the free shipping meter and the savings display, which are the two of the four changes that give a shopper a reason to add rather than a reason to relax. Desktop has the screen space to show line items, a progress meter, and per item savings all at once. Mobile does not, even full-screen.
Be careful with these two rows. Mobile ran on 109 orders and desktop on 78. That is small. We treat the direction as a signal worth acting on and the exact percentages as noise.
The point that survives the small sample: a cart drawer is not one change. Ours was four, and they do not pull in the same direction. Removing friction and merchandising the cart are different jobs, and how much screen the shopper has decides which one wins.
What We Would Tell You Before Copying This
Check your device mix first. This store runs about 72% mobile traffic. Because mobile revenue per visitor came out flat, most of the revenue gain was carried by a quarter of the audience. If your split is different, your result will be different.
A drawer is not automatically an upgrade. If your cart page is doing real work, converting well, showing recommendations that get clicked, then skipping it has a cost. Measure order value, not just conversion, or you will ship a mobile conversion win that quietly shrinks baskets.
Run the free shipping meter as its own test. It is our prime suspect for the desktop result and we cannot prove it, because we shipped it inside a bundle of four. If we ran this again it gets its own variation. Learn from our tradeoff instead of repeating it.
Do not add recommendations on day one. Leaving them out is why we can say anything useful about the drawer at all. When the winner shipped to the brand's rebuilt theme, we removed the recommendations block that theme already had, because it was never part of what won.
High-ticket is not exempt. That was our going-in assumption and the desktop numbers argue against it. Even when people buy one thing, showing them the cart and a shipping threshold changes what they put in it.
Drawer vs Banner Questions
Should every Shopify store replace its add-to-cart notification with a cart drawer?
No. This test says a drawer is worth testing, not that it always wins. The result depends on your device mix and on how well your existing cart page performs. On a store with a strong cart page and mostly desktop traffic, skipping that page can cost you.
Why did mobile conversion go up while average order value went down?
The banner forced shoppers to visit the cart page to check out. That detour added friction, which cost conversions, but it also gave shoppers a full screen to review and add items. Removing it converted more people at a smaller basket size. Net revenue per visitor on mobile came out flat.
Is 86% probability to beat control good enough to ship?
It depends on the cost of being wrong. Here the change was permanent, cost nothing to maintain, and moved two independent metrics in the same direction. Shipping was the better business decision than running another six weeks. On a change that is expensive to reverse, wait for 95%.
Do cart drawers work on high-ticket stores?
This one did, mainly on desktop. Around $700 average order value and a 0.4% conversion rate, the drawer still raised order value 33% on desktop. The free shipping progress bar appears to be the reason, because a shipping threshold is easy to clear when items are expensive.
What should we test inside the cart drawer after switching to one?
Isolate the free shipping meter first, since it is the most likely driver. Then trust badges, then recommendations. Test them one at a time. Shipping a drawer with four new elements at once means you learn that the bundle worked, not which part did.
How long should a cart drawer test run?
Long enough to collect a real number of orders, not a real number of visitors. This test had 84,005 visitors but only 347 orders, and orders are what the maths runs on. On a low conversion rate store, plan for six weeks minimum and expect wide confidence intervals anyway.