The upsell and cross-sell examples that raised order value most reliably in our archive share one trait: the offer is obviously relevant to what the buyer already chose.
A leather strap for a watch, install supplies for a golf grip, the same jeans in a new color after checkout and an annual plan beside a monthly one all won; a mystery box, an unrelated strap row and a paid engraving step at checkout all lost.
How we picked: every example is a test from the LiftArchive database, with the published number, who ran it and the evidence type.
We have not run these tests ourselves, and we say where a result is a forecast, a before-and-after or short of significance.
New to the terms? Our guide to upsells vs cross-sells covers the difference, and upselling covers the strategy.
All 19 upsell and cross-sell examples at a glance
| # | Example (brand or page) | What they did | Result | Type |
|---|---|---|---|---|
| 1 | PerTronix | Changed cart recs heading to 'Rev up your cart' | +40.37% RPV | Ecommerce |
| 2 | Jackson's | Rewrote the cart upsell heading | +83% recommendation add-ons | Ecommerce |
| 3 | Furniture store | One-click leather kit add in the cart | +$55 average order value | Ecommerce |
| 4 | Watch brand | Swapped which product the cart upsell offered | +15.26% revenue per visitor | Ecommerce |
| 5 | Gift wrap offered where shoppers looked for it | Added a 'Buying for someone else?' gift wrap block | +$58,860 per month | Ecommerce |
| 6 | Food brand | Free shipping bar with 'Upgrade to 10-Pack' buttons | +24% RPV | Ecommerce |
| 7 | WedgeGuys | One-click 'Frequently Bought Together' block near add to cart | +10.87% revenue per visitor | Ecommerce |
| 8 | Stone Creek Coffee | Moved add-ons into the purchase panel | +4% revenue per visitor | Ecommerce |
| 9 | Ulta Beauty | Overlay of relevant products after add to bag | +9% revenue | Ecommerce |
| 10 | Bear Mattress | Redesigned mobile add-on section with thumbnails and question-led copy | +24% purchases, +16% revenue | Ecommerce |
| 11 | Shipping insurance as a checkbox in checkout | Checkbox add-on above the payment step | 50% of orders took the upsell | Ecommerce |
| 12 | Yensa | Post-purchase: second foundation at 20% off vs a concealer | Converted at over 15% | Ecommerce |
| 13 | The Perfect Jean | Post-purchase: same product, different color, 40% off | 1.8% to 18.6% | Ecommerce |
| 14 | BOOM! | Post-purchase discount shown in dollars, not percent | 13.59% vs 10.3% take rate | Ecommerce |
| 15 | Hydroviv | Post-purchase replacement-filter subscription | 41% take rate | Ecommerce |
| 16 | Paltalk | Pop-up offering a larger pack when one is selected | +USD 3,900 revenue | App |
| 17 | Nitro | Added software assurance, backup CD and multi-license options | +21% average order value | SaaS |
| 18 | eteamz | Rebuilt the page shown when the premium trial ends | +43% upgrades | SaaS |
| 19 | Preu AI | Added a discounted annual plan to the paywall | +80% proceeds per user | App |
The 19 upsell and cross-sell examples
Examples 1-6 are in the cart, 7-10 on the product page, 11 in checkout, 12-15 after purchase, and 16-19 are pack and plan upgrades in apps and SaaS.
For the full set, browse cart upsell tests, product page upsell tests and checkout upsell tests.
1. PerTronix: a cart heading written for its buyers
Blend CommerceWhat they did: Automotive parts brand PerTronix kept the same cart drawer recommendations but swapped the default 'You May Also Like' heading for 'Rev up your cart', written for car and engine enthusiasts.
Why it works: A generic heading reads like filler.
A heading in the buyer's own language makes the add-ons feel picked for them rather than pushed at them.
Result: +40.37% revenue per visitor, with average order value up 21.18% and conversion rate up 12.91%.
Blend Commerce ran it in Rebuy Smart Cart.
It came close to, but did not reach, statistical significance, so treat the size with care. See the test.
Copy this: Replace 'You may also like' with one short line that sounds like your brand and names the job the add-on does.
2. Jackson's: 'Add more crunch to your cart!'
What they did: Snack brand Jackson's ran the same cart drawer recommendations under 'Add more crunch to your cart!' instead of 'You may also like'.
Why it works: Same idea as PerTronix, tested on a different product.
The copy gives the shopper a reason to add something, not just a label for a carousel.
Result: +83% recommendation add-ons, with conversion rate up 5% and checkout clicks up 6%, so the upsell did not pull people away from buying.
Blend Commerce ran it, and it reached statistical significance. See the test.
Copy this: Test your cart heading before you test which products sit under it.
It is the cheapest change on this list.
3. Furniture store: a one-click leather care kit in the cart
GrowthRockWhat they did: An online furniture store added a low-key strip to the cart offering a $40-80 leather conditioning kit.
A plus icon added it in one click.
Why it works: The kit protects a large purchase the shopper has just committed to.
It was also hard to find elsewhere on the site, so the cart was the first time many buyers saw it.
Result: +$55 average order value over 41 days and more than 4,000 transactions.
GrowthRock ran it. Significance was 92%, a little under the usual 95% bar. See the test.
Copy this: Find the accessory that protects or maintains your main product and offer it in the cart with a single click, no product page detour.
4. Watch brand: a leather strap beat a mystery box
What they did: A watch brand tested three cart drawer upsells: the existing mystery box, a leather strap and a watch case.
Why it works: A strap is an obvious fit for a watch.
A mystery box asks the shopper to gamble on something unrelated at the moment they are deciding whether to pay.
Result: +15.26% revenue per visitor ($1.71 to $1.97) and +18.19% conversion rate.
Average order value dipped 2.48%, but more people bought.
Milos ran it. The watch case also beat the mystery box, by less. See the test.
Copy this: Before redesigning the cart, test which single product you offer there.
The pick matters as much as the layout.
5. Gift wrap offered where shoppers looked for it
What they did: Survey answers showed customers wanted gift wrap but couldn't find it.
Oliver Kenyon added a 'Buying for someone else?' block to the cart with a gift wrap image, a £6 price and its own add button.
Why it works: It answers a need customers already had.
It is cheap and fits almost any order.
Result: +$58,860 per month, which is a forecast from Oliver Kenyon rather than measured revenue.
Average order value and gift wrap attach rate both went up; the sizes were not shared. See the test.
Copy this: Read your post-purchase survey and support tickets for add-ons people ask about.
Put the most requested one in the cart.
6. Food brand: pack upgrade buttons on the free shipping bar
What they did: A DTC food brand sold in packs added a card to the top of the cart drawer: a free shipping progress bar with markers for the 10-pack and 15-pack discounts, plus a one-click 'Upgrade to 10-Pack' or 'Upgrade to 15-Pack' button.
Why it works: This is a quantity upsell, not a new product.
The shopper already likes the item, and the button makes the next price break one tap away.
Result: +24% revenue per visitor among shoppers who opened the cart, at 88% probability to beat control on 119 vs 94 orders.
Convertibles ran it. The sample is small, so read it as directional. See the test.
Copy this: If you have quantity discounts, show the next tier in the cart with a button that upgrades the order in place.
7. WedgeGuys: Frequently Bought Together for a product that needs supplies
What they did: A golf grip page added a Frequently Bought Together block below the add-to-cart area, offering two install supplies with one 'Add all 2 to cart' button.
Why it works: Installing a grip needs extra supplies, so the add-ons are close to required.
Result: +10.87% revenue per visitor, with conversion rate up 8.09%, over 16 days on 12,421 visitors. Alex Stoykov ran it. See the test.
Copy this: If your product needs something else to work, offer that item on the product page with one button that adds both.
8. Stone Creek Coffee: 'Complete Your Routine' inside the buy box
What they did: Stone Creek Coffee moved a recommendation module from low on the product page, where only 2.41% of mobile visitors reached it, into the purchase panel.
It now suggests equipment that goes with the coffee.
Why it works: An upsell no one scrolls to cannot sell.
Moving it next to the buy button puts it where attention already is.
Result: +4% revenue per visitor across devices. Desktop gained strongly while mobile lost (revenue per visitor -9%), and none of it reached significance.
Blend Commerce ran it in Intelligems. See the test.
Copy this: Check a scroll map before testing upsell content.
If almost no one sees the module, move it first, and split results by device.
9. Ulta Beauty: recommendations right after add to bag
What they did: When a guest added an item to the bag, Ulta Beauty showed an overlay with the products most relevant to that guest.
Why it works: The shopper has just said yes to one item.
That is the moment they are most open to a second, and the overlay asks without sending them anywhere.
Result: +9% revenue, with add-to-bag clicks up 15.1%, conversion rate up 2.3% and items per order up 4.7%.
Ulta Beauty ran it, published as a VWO case study. See the test.
Copy this: Show two to four relevant add-ons in a confirmation overlay after add to cart, with a clear path to checkout.
10. Bear Mattress: add-ons that answer 'Why do I need this?'
What they did: Bear Mattress gave each mobile add-on a thumbnail, rewrote the copy around questions like 'Why do I need a protector?', only showed the frame after a foundation was added, and added a 'Mattress only' option.
Why it works: Shoppers skip add-ons they do not understand.
Explaining the reason for each one turns an upsell into advice.
Result: +24% purchases, +16% revenue. Add-to-cart for the protector rose 51.75% and the foundation 73%.
The test is credited to VWO and ran 19 days. Several changes shipped together. See the test.
Copy this: Write one line for each add-on that answers why the buyer needs it, in their words.
11. Shipping insurance as a checkbox in checkout
What they did: Fashion stores added a checkbox above the payment section of the Shopify Plus checkout offering shipping insurance against damage, loss and theft for 3.95 euros.
Why it works: A small, low-risk add-on in a single checkbox costs the buyer almost no effort to accept.
Result: 50% of orders took the upsell, shared by Kasper Dolk.
This is an observed take rate, not an A/B test, and the box in the example is pre-ticked. See the test.
Copy this: Offer one cheap, relevant protection or service add-on as an unticked checkbox, and check local rules before you pre-tick anything.
12. Yensa: a second unit beat a complementary product
What they did: After buying foundation, Yensa customers were offered either a complementary concealer or a second unit of the same foundation at 20% off.
Why it works: The buyer has just chosen this exact product and shade.
A second unit needs no new decision about fit or color.
Result: The second-foundation offer converted at over 15% and beat the concealer, whose rate was not shared.
Zipify's OCU Plus team ran it. The discount differed between offers, so product choice is not the only difference. See the test.
Copy this: In your post-purchase offer, test 'the same thing again, cheaper' against your best cross-sell.
13. The Perfect Jean: the same item in a new color
VisuallyWhat they did: The Perfect Jean replaced a post-purchase offer of a hoodie at 33% off with 40% off the item the shopper just bought, in a different color.
Why it works: It is the same logic as Yensa: the shopper already knows the fit. A new color is an easy yes.
Result: Offer conversion went from 1.8% to 18.6%.
Visually ran it. This is a before-and-after comparison after many rounds on Visually, and the discount also rose, so both played a part. See the test.
Copy this: For apparel and anything with variants, make your post-purchase offer the same product in another color or flavor.
14. BOOM!: '$6 off' beat '15% off'
What they did: After buying a Boomstick Color, BOOM! by Cindy Joseph buyers were offered another one at either '15% off' ($23.80) or '$6 off' ($22).
Why it works: On a low-priced item, a dollar figure is easier to picture than a percentage.
The dollar version was also slightly cheaper, so price and framing are mixed together.
Result: 13.59% vs 10.3% take rate, and +$0.60 per visitor for the dollar version. Ezra Firestone ran it. See the test.
Copy this: On items under about $100, test your upsell discount written as a dollar amount.
15. Hydroviv: a filter subscription right after checkout
What they did: Right after buying a water filter system, Hydroviv buyers saw a one-click offer for a recurring subscription to the matching replacement filter.
Why it works: Every buyer will need filters later.
Offering the plan now saves them a task and turns one order into recurring revenue.
Result: 41% take rate among under-sink system buyers.
Hydroviv set it up itself. It was measured after launch, not split tested. See the test.
Copy this: If your product uses consumables, offer the refill subscription on the thank-you page, matched to what they just bought.
16. Paltalk: a pop-up suggesting the bigger credit pack
What they did: Each time a Paltalk user selected a virtual credit pack, a pop-up asked if they wanted a higher-value pack to save more.
Why it works: The user has already decided to buy.
A size-up offer framed as savings asks for a small step, not a new decision.
Result: +USD 3,900 revenue over a 40-day, 50/50 test.
The test is credited to VWO. Only the dollar amount was shared, so the relative lift is unknown. See the test.
Copy this: If you sell credits, seats or packs, show the next size up with the per-unit saving at the moment someone picks a pack.
17. Nitro: cross-sells inside the in-app purchase flow
What they did: Nitro's in-app purchase flow for Nitro Pro used to sell one license.
Nitro added software assurance and a backup CD as cross-sells, plus the option to buy multiple licenses.
Why it works: SaaS buyers often need protection, support or more seats, but will not go looking for them.
Offering them inside the purchase step makes them part of the order.
Result: +21% average order value, and cross-sell revenue rose 22%.
Nitro ran it. It is a before-and-after comparison from 2014. See the test.
Copy this: Add one service add-on and a seat-count option to your SaaS checkout.
18. eteamz: a trial-end upgrade page that leads with value
What they did: When the eteamz premium trial ended, the old page asked users to upgrade before explaining why.
The new page stacked benefit bullets with images, then one CTA with a one-time offer and a note that premium data would stay locked.
Why it works: A plan upsell has to make the case first.
Users who have used the premium features respond to a reminder of what they would lose.
Result: +43% upgrades at 95% confidence (the source says almost 43%).
MECLABS ran it. Several changes shipped together. See the test.
Copy this: On your trial-end or upgrade screen, list what the user used during the trial before you show the button.
19. Preu AI: an annual plan next to the monthly one
What they did: Exam prep app Preu AI's paywall offered one monthly plan.
The test added an annual plan, described as priced at 60% of the monthly price.
Why it works: An annual option is the subscription version of an upsell: more money up front, framed as a saving.
Result: +80% proceeds per user and +22% trial conversion.
Preu AI ran it, published as a Superwall case study.
Sample size and prices were not shared. See the test.
Copy this: If you only sell monthly, add a discounted annual plan beside it and measure revenue per user, not just conversion.
Patterns across the winners
Cart: the copy and the pick matter more than the widget
Two of the biggest cart results changed only the heading (PerTronix, Jackson's).
Another changed only which product was offered (the watch strap).
Test the words and the product before buying a new cart app.
See Shopify upsell apps if you do need one.
Product page: put it next to the buy button and explain it
Add-ons that sat near the button and said why the buyer needs them won (WedgeGuys, Stone Creek on desktop, Bear Mattress).
Bundled offers are a close cousin; see our product bundle examples.
Post-purchase: same product beats new product
Yensa, The Perfect Jean and BOOM! all point the same way: after checkout, offer more of what they just bought, in one unit, with the discount written plainly.
Consumables do even better as a subscription (Hydroviv).
Our list of post-purchase upsell apps covers the tools.
SaaS and apps: the upsell is a bigger plan or pack
In subscription businesses, the upsell is usually the annual plan, more seats or a bigger credit pack.
Preu AI, Paltalk and Nitro show it works when the bigger option is framed as a saving.
More plan tests are in our app paywall examples.
Examples that lost (what not to copy)
- Too many related items in the cart. A premium bag brand added rows of matching straps to the cart drawer and revenue per visitor fell: -10% to -12% revenue. Order value held; fewer people bought. Convertibles ran it.
- A deal that asks for a big commitment. A buy 3, get 1 free golf ball offer in the cart drawer cost -2.8% revenue/visitor, and only 2.3% of shoppers took it. Convertibles ran it.
- A paid customization step at checkout. Loop Earplugs offered an engraved case for an extra €6 and conversion rate fell -3.04%, even though 6.89% of users clicked the option. Dash of CX ran it.
- Asking for two units instead of one. BOOM! offered two more Trios after checkout instead of one and lost: $5.32 vs $3.91 per visit. Ezra Firestone ran it.
- Upsells that cover the cart. Pvolve's cart upsells partly hid the shopper's items. Removing them raised +11.8% profit per session. Visually ran it.
The common thread: upsells lose when they add a decision, a step or clutter right before payment.
How to test this on your own site
- Start with one placement. The cart or post-purchase page is lowest risk, because the buyer has already chosen.
- Pick the most relevant single item. Use order data for what is bought together, or survey answers for what people ask for.
- Measure revenue per visitor and conversion together. A high take rate can hide a drop in orders, as the bag strap test showed.
- Split by device. Stone Creek's desktop and mobile went in opposite directions.
Use our AOV calculator to size the opportunity, and the average order value guide for other levers.
For every upsell test we hold, see the upsells element page or browse all examples.
For more cart and upsell tests, see Oliver Kenyon's tests.
Other teams behind these results include Blend Commerce, Convertibles and GrowthRock.
Real results
@CROguy_
Zipify OCU Plus team
@stoykovallex1
ConvertiblesWatch out for
- Pre-ticked paid add-ons. They inflate take rates and are restricted in the EU and some other markets.
- App stacking. Each upsell app adds scripts. Pvolve's win came from removing upsells, not adding them.
- Discount dependence. If every upsell is discounted, buyers learn to wait. Test a full-price relevant add-on too.