Ozone Landing Machine Case study: Save a Playa

A/B testing for your whole store

Ozone Landing Machine shows each new visitor the next of up to five complete designs, from the home page to checkout, and counts only paid first orders. At Save a Playa, five storefronts share the same traffic.

Proud to have worked with Magmic Grubbrr Digital Knack TGI Fridays Domestika
The challenge 5.39%

is what a beauty and personal care website converts, on average. Every redesign is a bet on beating it.

Source: Dynamic Yield by Mastercard, eCommerce conversion rate benchmarks.

Stop redesigning on opinion. Let your buyers vote.

Save a Playa pays for much of its traffic through Meta and TikTok ads. Every visit that lands on a weaker page is ad money spent for less.

The brief to Ozone was not a button-color test. Build complete new versions of the store from customer research, split new visitors evenly between them, count only first orders that were paid, and make switching to the winner one click. That became Ozone Landing Machine.

The test

Live on saveaplaya.org now.

Complete designs in the test 5

The current store plus four new designs, each with its own home page, product pages, cart and checkout.

Of new visitors per design 1 in 5

Round robin: visitor 1 sees design 1, and visitor 6 sees design 1 again.

What counts as a sale First orders

Paid orders from new customers. Bots, browser prefetches and staff are never recorded.

When a design wins 95%

Its chance to beat the current page, once the current page has 300+ visits.

What we built

Six things Ozone Landing Machine does, all running on saveaplaya.org.

A WordPress plugin on WooCommerce. The designs are sub-themes of the store’s own theme, so they share its products, prices, stock and checkout.

Hands each new visitor the next design in line.

An atomic counter assigns the designs in turn, so the split stays even when visitors arrive in the same second. In testing, twenty visitors arriving at once split exactly four per design.

Keeps every shopper on one design.

A returning visitor sees the design they first saw for 90 days, on every page in the test, so no buyer compares two versions of your store.

Tests complete storefronts, not buttons.

Each design can bring its own home page, product pages, cart and checkout. A page a design does not change falls back to your current one.

Counts only paid first orders.

A conversion is a paid order from a customer with no earlier order under the same billing email. Bots, browser prefetches and logged-in staff are never recorded.

Shows each design's chance to win.

The dashboard runs 20,000 simulated draws from each design's visits and orders and shows its probability of beating your current page, next to the rule for when to switch.

Makes the winner your default in one click.

One click on the dashboard makes the winning design your default page. Search engines and bots always get the default, so Google indexes one version of your store.

See it on a real store first. Save a Playa: brand, WooCommerce store, 200,000+ units sold, and the test that decides its next home page.
What it means for you

Hire the developer, and the platforms come with an open license.

In the agentic era e-commerce companies need a solution tailored to their own store to get ahead of the industry’s average metrics, not another tool with the same defaults as everyone else.

Full-time developer
$3,900/mo
160 hours a month, month to month
One senior full-stack developer, exclusively yours, on US time.
An open license to all three platforms: Revive, Second Order and ROAS Link. Source code, database schema and prompts, yours to change while the seat runs.
Part-time developer
From$2,500/mo
80 hours a month, month to month
One senior full-stack developer, on your hours, on US time.
An open license to one platform of your choice: Revive, Second Order or ROAS Link, with the same terms.
The right seat for a store that wants one platform running well before adding the rest.
Keep it after the seat
$900a month, flat

The platform core with its updates, and every customization made for your account, kept running on our server or yours. Hosting and support included. No development hours; when you need them, $60 an hour. Month to month.

Not sure your store has the traffic for a test? Book a call and we will tell you, or tell you to wait.

Both seats: one month minimum, paid in advance via PayPal or Payoneer, no recruiting fees, cancel any time after the first month.

Getting started

Your test starts from a running platform, not a blank page.

The same plugin that runs on saveaplaya.org, set up on your store by the team that built it.

Week one

We install Ozone Landing Machine, connect it to your WooCommerce orders and choose the pages it rotates: home, product pages, cart and checkout.

Week two

Our designers build the new versions from your customer research, as sub-themes of your own theme, and check each one on phone and desktop.

Week three onward

The test runs. You follow visits, first orders and each design's chance to win on the dashboard, and one click makes the winner your default page.

Questions

A/B testing on a live store, explained.

The rule to switch needs 300+ visits on your current design and a design with a 95% chance to beat it. With five designs, each one gets one in five new visitors, so plan for at least 1,500 new visitors before the first decision. Fewer designs need less traffic.

Stop guessing which page sells. Let your buyers decide.

Up to five designs, the same traffic, and only paid first orders counted. The winner becomes your default page.

Part time from $2,500 a month (80 hours, one platform of your choice) or full time at $3,900 (160 hours, all three platforms), on US time, month to month, no recruiting fees.