Find the page that sells more.
Test complete page versions yourself, or let Lexsis run the comparison and name a winner once the results are clear.
Page decisions shouldn't be made by opinion.
A redesign can look stronger and still sell less. Test the whole buying journey before every visitor sees it.
The favourite design goes live for everyone.
- All traffic sees an unproven redesign
- One page has to serve every campaign angle
- Setup and reporting scattered across tools
Complete pages compete against a stable control.
- A share of traffic tries each version, the rest is protected
- Each visitor keeps seeing the same version
- Pages, traffic and results in one place
No winner until the results are clear.
Early leads come and go. Lexsis waits until there is enough time and traffic, and the store checks pass, before it calls a winner. Step through a test below.
Whole page directions, not button colours.
Compare the decisions that change whether someone buys: what you say, what you offer, and what they see first.
Message
A benefit-led headline against a problem-led one.
Offer
A single product against a bundle or a different incentive.
Hero
A lifestyle image against a product-first layout.
Section order
Reviews first, or the bundle first, or the comparison first.
Page type
A campaign page against the product page, or a quiz against a collection.
Everything a complete-page test needs.
Complete-page comparisons
Every version is a full page, built from your products and brand.
As many versions as you need
Two pages or several, against the same stable control.
Fixed or adaptive traffic
Split traffic evenly, or shift more of it toward the versions that are doing better.
Stable visitor assignment
A visitor who saw version B keeps seeing version B.
Metrics and safeguards
Revenue and conversion, with guardrails that protect against a bad version.
Manual or managed
Run tests and make the call yourself, or let Lexsis manage them and decide.
Judged on revenue, not just clicks.
Every test reports the numbers that decide whether a page is better for the business, with Shopify orders behind them.
- Conversion ratePrimary
- Revenue per visitorPrimary
- Average order valueSupporting
- CTA clicksSupporting
- Section reach and scroll depthSupporting
- Eligible sessionsReadiness
No early winners
A version can lead on day two and lose by day ten. Winners are only called once time, traffic and store checks are all met.
Guardrails
Guardrail metrics run next to the main result, so a version that starts hurting the store shows up before it costs you.
Your call, or ours
Publish the winner yourself, or let Lexsis do it once the decision is clear.
One page, one question, one decision.
- Step 01
Define the test
Pick the page, the question you want answered and the metric that decides it.
- Step 02
Create the versions
Lexsis builds each challenger from your products and brand. You can edit any of them.
- Step 03 · Your review
Review and launch
Check every version on desktop and mobile, set the traffic split, and start.
- Step 04
Decide and publish
When the results are clear, publish the winner, and start the next test from what you learned.
Active tests on CRO & A/B Testing plans
Agentic Commerce plans include 5, 15 or unlimited.Test anything the rest of Lexsis builds.
Reach more visitors
Test the pages that visitors from AI answers land on.
Explore →02 · Campaign pagesConvert more clicks
Test a campaign page against the product page the ad used to send to.
Explore →03 · Shopping agent · QuizzesHelp visitors choose
Test a page with the agent or a quiz against one without.
Explore →How many versions can I test at once?
As many as the test needs, each against the same control. How many tests you can run at the same time depends on your plan.
How is traffic split?
Evenly, or adaptively, where more traffic goes to the versions that are doing better. Either way, each visitor keeps seeing the same version.
How do you avoid calling a winner too early?
A winner is only named once the test has run long enough, every version has had enough traffic, and the store checks pass.
Who decides which version wins?
You choose. Run tests manually and make the call yourself, or let Lexsis manage the test and publish the winner once results are clear.
Does it work with Shopify?
Yes. Every version uses your live Shopify products, cart and checkout, and results are measured against real Shopify orders.
Bring one campaign and one page decision.
Book a demo and we'll set up the versions and a controlled test with you.