FOUNDING STORES
$0during early access
no card · onboarding in waves
- ▸ Every product-page test we've shipped, nothing held back
- ▸ A direct line to the founders, and a human answers
- ▸ Fair warning and better terms when pricing starts
[ 01 ] A/B TESTING, GHOSTWRITTEN
Your next winning test is already written.
GhostBot reads your product pages and writes the A/B tests you never get around to running. You approve anything before it goes live, and every result is settled on the revenue it brought in.
FREE DURING EARLY ACCESS · NO CARD · US SHOPIFY STORES FIRST
Animated demo of the GhostBot pipeline: the AI drafts a test variant, storefront traffic splits, purchases are attributed to variants, and a statistically significant winner is declared. All numbers simulated.
[ 02 ] THE PROBLEM
Testing works. Nobody has time to run it.
You already know A/B testing pays. Running one is the problem. The tools hand you an empty editor and wish you luck: pick what's worth testing, write the variant, then squint at a significance calculator.
So the page ships on instinct and you never learn what the last redesign cost you. GhostBot does the part nobody has time for. It writes the tests and it handles the statistics.
[ 03 ] WHAT'S IN THE BOX
What the ghost actually does.
DRAFTS · YOU APPROVE
Test ghostwriter
It reads the product pages you actually have, then writes variants for the things that tend to move sales: the headline, the button, the line that answers the obvious objection. Every suggestion comes out of what's on the page.
GUARDRAILED
You hold the pen
Nothing reaches your storefront until you say so. Anything unsafe gets thrown out before it lands in your queue, so a test can't slip in a script, link off your site, or break a layout.
MEASURED IN $
Revenue verdicts
When an order comes in, we tie it back to the variant that earned it, server side from Shopify's own webhook. Winners are called on money rather than clicks, with a proper stop condition on the test.
CONSENT-AWARE
Privacy, built in
The script asks your store's consent settings through Shopify's Customer Privacy API and stays quiet when a shopper hasn't agreed. Deletion requests carry through to our analytics on their own. It weighs about 3KB and loads asynchronously.
[ 04 ] HOW IT WORKS
Three steps to your first verdict.
01
Install
Add the Shopify app, then switch on the theme embed. A checklist walks you through it and there's no code to paste anywhere.
ghostbot ~ app embed enabled ✓
02
Approve
GhostBot drafts an experiment for one of your pages. You see the current line next to the new one, and you either approve it or you don't.
variant B: "Free US shipping" · approve? [y/n]
03
Collect
Traffic splits on its own. Once enough of it has come through to trust the result, you get a verdict in dollars, and whatever's losing gets paused without you having to watch for it.
verdict: B · +$412 revenue · significant ✓
[ 05 ] THE MATH
What counts as a win.
01
Exposure
We record which variant each visitor saw, from first paint. A test can't be judged on traffic that never reached it.
02
Conversion
Add-to-carts and checkouts get tied back to the visitor who saw the variant, counted per person rather than per pageview.
03
Revenue
Orders reach us server side from Shopify's order webhook, so the figure against each variant is money you actually took, in your own currency.
04
Verdict
A stop condition decides when the test is finished. Whatever's underperforming gets paused for you.
There's no dashboard nudging you to call it early, and nothing parked at “trending toward significance” for three weeks. A test finishes when the numbers hold up, and the answer comes back in dollars.
[ 06 ] EARLY ACCESS
Free while we onboard founding stores.
THE FINE PRINT, UP FRONT
We're letting stores in a wave at a time, US Shopify only for now. Founding stores get the whole product, a direct line to the people building it, and plenty of warning plus better terms whenever we do start charging.
[ 07 ] FAQ
Questions, answered straight.
What exactly does GhostBot change on my store?
Only what you approve. It changes wording and emphasis on pages you already have, applied by a small script as the page renders. Unsafe edits are rejected before they ever reach your review queue, so scripts, off-site links and layout-breaking styles never get that far. Switch the embed off and your theme is exactly as it was.
Will it slow my storefront down?
The script is around 3KB compressed and loads asynchronously, so it never holds up your page rendering. There's no session replay and no pile of third-party trackers coming along for the ride.
How does GhostBot handle privacy and GDPR?
The script asks Shopify's Customer Privacy API first and only collects anything when consent allows it. If a shopper declines, tracking stays off and any identifiers we'd stored are wiped. When Shopify sends a deletion request for a store or a customer, it carries through our analytics automatically.
What does “free during early access” mean?
The whole product, free, while we onboard founding stores. We'll charge eventually. When we do, early stores get advance notice and better terms than whatever the public price turns out to be. Nothing to cancel and no card now.
Do I need to understand statistics?
No. GhostBot runs the split and handles the statistics, and it only calls a winner when the result holds up. What you read at the end is a plain sentence telling you which version made more money.
Does Shopify have built-in A/B testing?
Yes, as of early 2026. Rollouts sits in your admin on the Grow plan and above, and it splits traffic between two versions of your theme. It works at the section level, and it stops short of what most conversion rate optimization (CRO) work actually needs: it won't write the variant for you, it doesn't report revenue per visitor, and it never calls a winner, so you read confidence intervals and decide for yourself. GhostBot writes the test, you approve it, and the result comes back as attributed revenue with a stop condition on it.
Can I A/B test my Shopify product pages?
Yes, and product pages are where GhostBot starts, because that's where the buying decision actually happens. It tests the headline, the button, and the line that handles the obvious objection, then attributes orders back to whichever version produced them. Collection pages and the cart come later.
How much traffic do I need before A/B testing is worth it?
It depends on your conversion rate and the size of the improvement you're hoping to find. A store converting at 2% needs considerably more traffic to detect a 5% lift than one converting at 10%. GhostBot won't call a winner before the numbers support it, so on a smaller store a test runs longer rather than handing you a confident answer that happens to be wrong.
Who can install it today?
US Shopify stores, which is where early access is focused for now. EU and UK are planned. Join the waitlist and we'll write to you when your region opens.
[ 08 ] EARLY ACCESS
Let the ghost write the next one.
NO CARD · ONBOARDING IN WAVES · SUPPORT AT GHOSTBOT.AI