Create your first experience

Playbook sits on top of the store you already have. Your catalog, checkout, and theme stay exactly where they are — what changes is what a shopper sees the moment they arrive, matched to the ad or audience that brought them. We call each of those an experience: a landing page for a specific shopper, built on your live store.

This guide takes you from a blank composer to your first live, tested experience.

Create an experience

Everything starts on the Create page — one composer, not a multi-step wizard.

1. Pick a destination. This is the one required choice: the page your experience is for — a product, a collection, or a page. It's what the experience targets and what grounds the AI in real products. Until you set it, the button reads "Pick a destination."
2. Describe what you want (optional). Type a line like "turn a cold ad click into a first order" or "make my bestseller impossible to scroll past." Leave it blank and Playbook works from your brand and the destination alone.
3. Add context (optional). Attach a Meta Ad (its images, copy, and audience feed the draft), a Template to start from a proven layout, or file it under a Play. You can also drop in a reference photo.
4. Sharpen your prompt. As you type, Playbook may offer a few clarifying chips (a goal, a primary CTA). Answering one adds a small pill to your prompt — it never blocks you, and you can skip it.
5. Generate. Playbook drafts the whole experience — sections, copy, and imagery — and drops you into the editor, where the sections stream in. Edit anything: text, images, layout, colors.

That's a draft. Nothing is live yet — which is exactly what you want before your first test.

Take it live with a test

An experience is a draft page until you put it in a test. Starting a test is what makes it live and what lets you learn whether it actually works. In the New test dialog you choose how visitors see it:

• Split test (50/50) — half your visitors see the new experience, half see your comparison. The fastest way to know which one sells better.
• Send all traffic to the new experience — everyone sees it right away, no comparison. Pick this when you just want to ship it.

For a split test, you compare against your current page (the original, untouched) or a different experience (a head-to-head between two ideas). Both sides target the same page. Hit Start test (in send-all mode the button reads Send it live) and Playbook begins serving it.

Going live needs an active Playbook plan — you'll be prompted to choose one the first time you start a test. Reading your results is always free; only serving to shoppers needs a plan.

Going live is where testing stops being a guess. You're not betting — you're finding out, with real shoppers and real numbers.

The best first tests to run

Don't overthink your first one. The tests that teach you the most, fastest:

Start with a page that already gets traffic

Your bestseller's product page, or the page your top ad points at. Testing where visitors already land means you gather a verdict in days, not months — and any lift there is worth real money.

Change one clear thing

A split test answers one question well. Rework the hero, sharpen the headline, reframe the offer — then run it against your current page. When it wins, you know exactly why.

Match the experience to the traffic

If an ad promises "clinically-tested hydration," the page it lands on should say so above the fold. Ad-to-page match is one of the highest-leverage things you can test — and one Playbook is built to make effortless.

Reading your results

Two places tell you what happened:

• Optimize (Analytics) — your portfolio view, headlined "Your wins": revenue won, impressions, conversions, and what's live.
• A single experience's analytics — the drill-down, with three numbers up top: revenue lift, conversion-rate lift, and win probability.

A few habits keep you honest:

• Let it gather enough visitors. Playbook shows a readiness / confidence signal. A day-one number is noise; wait for the signal to firm up before you act.
• Watch win probability. Above 50% and climbing means the experience is genuinely ahead, not just lucky.
• Understand "revenue won." It's the extra revenue your experience earned over the original page — incremental lift, not gross sales. It starts counting toward your total the moment your experience earns more per visitor than the original — that's positive lift, and it's what puts dollars on the board. Win probability is a separate read on how sure we are, not a gate on the money — a small, early lead already counts. The winner keeps running for everyone, and its earnings keep adding up.

Best practices, in one breath

• Start with your money pages.
• Change one thing at a time.
• Give every test enough traffic to be sure.
• Keep the winner running — that's where the revenue compounds.
• Set your brand voice and play context once, so every draft starts on-brand.

The fastest way to understand any of this is to do it once. Open Create and ship your first one.

Next: set the foundation that makes every draft sound like you — Play context and brand voice.