The best copy A/B tests in our archive share one move: they replace vague or clever wording with plain words about the outcome the visitor wants.
The top pick is a benefit-led product title ('Achieve [outcome] with [product]') that lifted revenue per visitor 16.92%, followed by a cart message that turned free shipping into a reward (+13.93%) and one added word, 'supplement', that lifted purchases 89.97%.
How we ranked them: we pulled the copy tests from the LiftArchive research database and ranked them by how cleanly the test isolated the wording, how strong the evidence is (A/B test with reported numbers first) and how useful the lesson is across sites.
Lift size came last, because small samples produce the biggest numbers.
Three losses are included because they teach as much as the wins.
Ranked table: 20 copywriting tests at a glance
| Rank | Test | Change | Result | Who ran it |
|---|---|---|---|---|
| 1 | Benefit-led product title | Plain name to 'Achieve [outcome] with [product]' | +16.92% revenue per visitor | Dave Diederen |
| 2 | Free shipping cart message | 'Will be applied at checkout' to 'You unlocked FREE shipping!' | +13.93% revenue per visitor | Dave Diederen |
| 3 | Movexa headline | Added the word 'supplement' | +89.97% purchases | Source Patrol |
| 4 | Newsletter headline | Question headline to a customer quote | +24.31% sign-ups | Laura Roeder Social Media |
| 5 | Product page benefit copy | General claims to pain-point copy | +$159K per month (forecast) | Oliver Kenyon |
| 6 | BestSelf Self Journal headline | Headline aimed at professionals | +27% sales | SplitBase |
| 7 | Corcentric demo CTA | 'Get a Demo' to 'Free Demo' | +99.42% clicks | Greenlane Search Marketing |
| 8 | Add-to-cart button copy | Creative labels vs 'Add to Cart' | 'Add to Cart' won every time | Dave Diederen |
| 9 | Fear-angle description | Risks of not using the product | +41.90% revenue per visitor | Dave Diederen |
| 10 | Infinite Age hero headline | Slogans to an outcome-led headline | +35% purchases | Vasta |
| 11 | Chêne Gear product names | One benefit word added | up to +27.1% conversion | Visually |
| 12 | Showpo newsletter checkbox | Default label to playful 'EXCLUSIVE' line | +175.52% sign-ups | VWO |
| 13 | PriceCharting button | 'Download' with price to 'Price Guide' | +620.9% clicks | PriceCharting |
| 14 | VenueSphere form subheadline | Process copy to 'we'll do the hard work for free' | +69% leads | VenueSphere |
| 15 | AppVault subject line | First name to the prospect's ATS | 13% to 66% opens | AppVault |
| 16 | Mojo annual price copy | Yearly price to '$X/month, billed annually' | +45% Brazil, +26% Mexico, +8% USA | Mojo |
| 17 | Clear 30 paywall | Feature carousel to a 7-day preview | 20% → 30% conversion (+50%) | Clear 30 |
| 18 | Free trial button | 'Start Free Trial' to 'Try Now' | -26% clickthrough (lost) | MECLABS |
| 19 | PPC ad headline | Question to statement | CTR fell from 2.71% to 1.99% (lost) | MarketingExperiments |
| 20 | Donation page | Added value copy before the ask | -28% donations (lost) | NextAfter |
The tests cover headlines, CTA copy, value propositions, microcopy, email subject lines and app paywalls across ecommerce, SaaS, lead gen and mobile apps.
For the theory behind them, start with our copywriting guide and the guide to writing headlines.
The 20 best copywriting A/B tests
1. Product title rewritten as 'Achieve [outcome] with [product]'
The change: A plain product title in the style of 'Resistance Band Set' became a title that leads with the result the buyer wants, then names the product, in the style of 'Build Strength at Home with the Resistance Band Set'.
The result: +16.92% revenue per visitor ($1.10 to $1.28) and +15.61% conversion rate (1.79% to 2.06%) in an A/B test run by Dave Diederen.
Average order value was effectively flat (+1.14%). The exact title and sample size were not shared.
Why it likely worked: The title is the first line most shoppers read on a product page.
Leading with the outcome answers 'what does this do for me' before they scroll.
Try it if: Your product names are nouns or brand names that say nothing about the result.
2. Cart copy: 'Congrats! You unlocked FREE shipping!'
The change: Once the cart passed the free shipping threshold, the line 'Free shipping will be applied at checkout' was replaced with 'Congrats! You unlocked FREE shipping!'
The result: +13.93% revenue per visitor ($3.07 to $3.49) and +11.08% conversion rate (2.82% to 3.14%), A/B test by Dave Diederen.
The ranges were wide and crossed zero, so the true lift may be smaller.
Why it likely worked: The same fact reads as a reward the shopper earned instead of a note about a later step.
It also confirms the benefit now, in the cart, rather than promising it at checkout.
Try it if: You have a free shipping threshold and the cart message is flat or procedural.
3. Adding the word 'supplement' to a headline
The change: The Movexa landing page headline named the joint product but not what kind of product it was. The variant added one word: 'supplement'.
The result: +89.97% purchases at 98% significance, in an A/B test run by Source Patrol and published as a VWO case study.
Sample size and conversion rates were not shared.
Why it likely worked: A coined brand name gives no clue about the category.
One plain word told visitors what they were looking at before they read anything else.
Try it if: Your product or brand name is invented and your headline assumes people already know what it is.
4. A customer quote as the headline
The change: The homepage headline 'Could You Use a Free Plan For Promoting With Social Media?' was replaced with a line from customer feedback: 'Yours is the only newsletter that I actually read.'
The result: +24.31% newsletter sign-ups, with a 97% chance to beat the original after 12 days, in an A/B test by Laura Roeder Social Media (VWO case study). Sample size was not shared.
Why it likely worked: Every newsletter claims to be useful.
A reader saying it is the only one they open is a specific, believable claim the brand could not make about itself.
Try it if: You collect emails and have real feedback from subscribers or customers you can quote word for word.
5. Pain-point benefit copy instead of general claims
The change: Under the price, short generic model descriptions that explained what the product was were replaced with a short paragraph and a ticked list about the pain points it fixes, framing it as key to performance and recovery.
The result: Revenue per session up $0.604, forecast at about $159,125 extra per month, in an A/B test by Oliver Kenyon.
The baseline was not shared, so the percentage lift is unknown, and the monthly figure is a forecast.
Why it likely worked: Shoppers buy the fix, not the spec.
Copy that names the problem in their words makes the price feel tied to something they want gone.
Try it if: Your product page describes features and models, and your reviews keep mentioning a problem the copy never names.
6. A headline aimed at one audience
SplitBaseThe change: BestSelf Co.'s Self Journal product page got a new headline that addressed professionals by name, promised help reaching their goals and added how many people use the journal.
The result: +27% Self Journal sales against the original page, A/B test by SplitBase.
Desktop conversion rose 74%, but that is a segment result, so it is larger than the overall lift.
Why it likely worked: Research showed visitors were not scrolling and did not grasp what the journal would do for them.
Naming the reader and the goal in the first line did that job above the fold.
Try it if: Most of your buyers share one role or situation and your headline speaks to everyone.
7. 'Get a Demo' changed to 'Free Demo'
The change: On Corcentric's homepage, the yellow button stayed the same.
Only the copy changed, from 'Get a Demo' to 'Free Demo'.
The result: +99.42% CTA click-through rate, A/B test by Greenlane Search Marketing (VWO case study).
The sample was close to 800 visitors over 2 weeks, and completed demo requests were not shared.
Why it likely worked: B2B visitors often assume a demo is the start of a sales process with a cost attached. 'Free' answers that worry on the button itself.
Try it if: You sell to businesses and your main CTA asks for a demo or call. Measure booked demos, not just clicks.
8. Plain 'Add to Cart' against creative button copy
The change: The add-to-cart button was tested with creative alternatives: 'Try Risk Free Today', 'Get Yours Now' and 'Start Your Journey'.
The result: "Add to Cart" won every time, in A/B tests by Dave Diederen. No per-variant numbers were published.
Why it likely worked: Shoppers scan for the button they expect.
Clever copy makes them stop and read, and some phrases ('Start Your Journey') hide what clicking will actually do.
Try it if: Someone on your team wants to make the buy button 'more on brand'.
Test it, but expect the plain label to hold.
9. Product description rewritten with a fear angle
The change: The description was rewritten to spell out the health risks a shopper faces if they skip the product.
Two other angles, value and problem, ran in the same test.
The result: +41.90% revenue per visitor ($2.67 to $3.79) and conversion rate 3.24% to 4.31% (+32.75%), A/B test by Dave Diederen.
The value and problem angles also beat control, by less.
Confidence ranges were wide (-33% to +201% for revenue per visitor), so treat the size with caution.
Why it likely worked: Loss framing makes the cost of doing nothing concrete.
It works best when the risk is real and the product genuinely reduces it.
Try it if: Your product prevents a real problem (health, damage, loss) and the copy only lists benefits.
Keep it truthful; scare copy that overstates risk can backfire.
10. Outcome-led hero headline under 12 words
VastaThe change: Broad wellness slogans in Infinite Age's landing page hero were replaced with two headlines, each under 12 words, that led with the supplement's main outcome and added a light urgency cue.
The result: +35% completed purchases at 95% confidence over a 10-day even split, A/B test by Vasta.
The source does not say which of the two headlines won.
Why it likely worked: Slogans sound nice but make visitors work out what the product does.
A short outcome statement does that work for them.
Try it if: Your hero headline could appear on any competitor's site without anyone noticing.
11. One benefit word added to product names
VisuallyThe change: Chêne Gear's branded product names stayed, but each title gained one word naming a benefit for the hunter.
The result: Up to +27.1% conversion rate in the best single test, by Visually.
Visually says every variant beat its original, but the other results were not shared, so 27.1% is the high end.
Why it likely worked: Branded names mean something to the team and nothing to new visitors.
One functional word turns a name into a reason to click.
Try it if: You have a catalog of named products where the name does not say what the item does.
12. Newsletter checkbox copy with personality
The change: Showpo's checkbox read 'Join our Newsletter.' The winning variant read 'We wouldn't say EXCLUSIVE if we didn't mean it! Join our newsletter to find out.'
The result: +175.52% newsletter sign-ups on the sign-in page and +146.78% on the checkout page, credited to VWO's case study. The metric is sign-ups, not sales.
Why it likely worked: 'Join our newsletter' offers nothing.
The new line hints at a reward and addresses the reader's doubt about whether 'exclusive' means anything.
Try it if: You have an opt-in checkbox or field with default wording. See more microcopy ideas.
13. 'Price Guide' instead of 'Download' with a price
The change: PriceCharting's button read 'Download' with the fee beside it, for example 'Sega Genesis Price List for $2'.
The variant read 'Price Guide' with the price removed and led to a page to build a custom guide.
The result: +620.9% button clickthroughs at 99.9% significance after about a week, A/B test by PriceCharting (VWO case study). Purchases were not reported.
Why it likely worked: The label named the thing people wanted instead of the action, and hiding the price until the next step removed the reason to hesitate.
Some of the lift likely came from the hidden price, not the words alone.
Try it if: Your CTA names a mechanical action ('Download', 'Submit') rather than what the visitor gets. Track the downstream purchase too.
14. 'We'll do the hard work for you for free' above a form
The change: VenueSphere's subheadline above the enquiry form changed from 'Call us or fill in a form to speak to a dedicated venue coordinator' to 'Stop right now!
Call us or fill in the form and we'll do the hard work for you for free.'
The result: +69% enquiry form submissions, A/B test by VenueSphere (VWO case study).
Close to 1,200 visitors took part, a small sample for a lift this large.
Why it likely worked: The original described a process.
The new copy promised that the visitor would do less work and pay nothing, which is the real reason to use a venue finder.
Try it if: You run a lead-gen form and the copy around it explains steps instead of what the visitor saves.
15. Naming the prospect's software in the subject line
The change: A B2B email subject line personalized with the recipient's first name was replaced with one naming the applicant tracking system installed on the prospect's careers site.
The result: Unique open rate went from 13% to 66%, in a test by AppVault reported by MarketingSherpa.
Only opens were reported; no clicks, meetings or sample size.
Why it likely worked: First-name tokens are everywhere and easy to ignore.
Naming the tool a prospect actually uses signals the email is about their setup, not a blast.
Try it if: You sell B2B and have firmographic or technographic data you could put in the subject line.
16. Annual plan shown as '$X/month, billed annually'
The change: Mojo's paywall showed the annual plan as one yearly price.
The variant showed it as a monthly figure with yearly billing stated.
The result: New revenue per paywall impression up 45% in Brazil, 26% in Mexico and 8% in the USA, in an A/B test by Mojo reported on RevenueCat's blog. Sample size and confidence were not shared.
Why it likely worked: The same price written per month is easier to compare with a monthly plan and feels smaller.
The effect was much weaker in the higher-income market.
Try it if: You sell subscriptions with an annual plan, especially in price-sensitive markets.
Check app store rules on how prices can be shown.
17. Paywall copy that previews the first 7 days
The change: Clear 30 removed the feature carousel (peer support, accountability texts, progress tracking) from its paywall and instead outlined what the first seven days of the program would look like, day by day.
The result: Paywall conversion went from 20% to 30% (+50% relative), reported by Clear 30's founders on Superwall's blog.
This was a before-and-after change, not a split test.
Why it likely worked: Features describe the app.
A day-by-day plan lets the user picture themselves doing it, which is what they are actually buying.
Try it if: Your paywall lists features and your product is a program, course or habit.
Run it as a proper split test if you can.
18. 'Try Now' lost to 'Start Free Trial'
The change: A banner button reading 'Start Free Trial' was tested against 'Try Now' and other phrases.
The result: Lost: -26% clickthrough for 'Try Now', in a test by MECLABS published by MarketingExperiments.
In the same test, 'Get Started Now' lifted clickthrough by more than 11%. Only clicks were measured.
Why it likely worked: 'Try Now' drops the word 'free' and says less about what happens next.
Shorter is not better when it removes the value.
Try it if: You are tempted to shorten a CTA. Keep the words that carry the offer.
19. Statement headline lost to a question in a PPC ad
MarketingExperimentsThe change: A pay-per-click ad shown on searches for 'free content' had its question headline swapped for a statement.
The result: Lost: clickthrough rate fell from 2.71% to 1.99%, in a test by MarketingExperiments.
The question ad did 36% better. No impression counts or significance were shared.
Why it likely worked: A question that mirrors what the searcher is wondering can feel more relevant than a claim.
This is one 2008 ad test, so treat it as an idea, not a rule.
Try it if: You write search ads. Test one question headline against your best statement.
20. Value copy before a donation ask
The change: A donation page that went straight to the ask was tested against one that first laid out the seminary's value proposition and quotes from well-known figures.
The result: Lost: donation conversion fell 28%, statistically significant per the source, in a test by NextAfter. Sample size was not shared.
Why it likely worked: Visitors on this page were already motivated to give.
More persuasion added reading and friction between them and the form.
Try it if: You plan to add more copy to a page people reach with their mind made up. Test a shorter version first.
What the winners have in common
They say what the thing is and what it does
Movexa's one-word fix and Chêne Gear's benefit words both solved the same problem: a name that meant nothing to a new visitor.
If your headline or product title would confuse someone who has never heard of you, fix that first.
Our headlines and copy element page collects more of these.
They lead with the outcome, not the feature
The 'Achieve [outcome] with [product]' title, Infinite Age's headline, Oliver Kenyon's pain-point copy and Clear 30's 7-day preview all moved from describing the product to describing the result.
See features vs benefits for how to make that switch.
CTA copy wins when it removes a worry or names the reward
'Free Demo' answered a cost worry and 'Price Guide' named what people wanted.
But plain 'Add to Cart' beat creative labels, and 'Try Now' lost to 'Start Free Trial'.
The pattern: keep the words that carry the offer, cut the ones that only add style.
More in CTA copy and our CTA button examples.
Specific beats generic
A customer's exact words beat a marketing question.
A prospect's own software in the subject line beat their first name. Showpo's specific promise beat 'Join our Newsletter'.
More copy is not always better
NextAfter's donation page lost 28% when value copy was added for visitors who had already decided.
Copy should match how ready the reader is.
Our list of best practices that failed has more cases like this.
What to test next
- Rewrite your main product titles or hero headline to lead with the outcome, and keep the product name in it.
- Add one plain word that says what your product is if the name is invented.
- Replace a generic opt-in or form label with a specific promise about what the reader gets.
- Test your primary CTA for a missing reassurance word ('free', 'no card needed') rather than a clever rewrite.
- Swap a marketing-written headline for a real customer quote on a sign-up page.
- Rewrite system messages (cart, shipping, confirmation) as benefits the visitor has earned.
- On paywalls and pricing pages, test how the price is written, not just the price itself.
Measure the metric that pays the bills. Several tests above report clicks or opens only, and a click lift does not always become a sale.
Our losing A/B tests page shows how often promising ideas fail, and first A/B tests to run helps you pick an order.
More copy tests from the archive
@DaveDiederen@DaveDiederen
@kasper_dolkWatch out for
- Small samples, big numbers. The largest lifts here (+620.9%, +99.42%, +69%) came from short tests or small samples. Rerun a big winner before you treat it as a rule.
- Clicks are not sales. A CTA that doubles clicks can still send the same number of buyers. Track the final conversion.
- Scare copy and fake urgency. The fear-angle description worked because it described real risks. Overstated risks or invented deadlines erode trust and can break advertising rules.
If you want help finding the copy changes worth testing on your own site, Oliver Kenyon runs research-led copy and page tests for ecommerce and subscription brands.