# QA & Preview | Tailor AI

> QA your Tailor AI experiments before launch. Preview variants, share links, test behind login walls, and force yourself into treatment groups.

Source: https://tailorhq.ai/docs/qa-preview

[Docs](/docs)

Toggle navigation

# QA & Preview

Verify your tailored experiences before they go live. Preview, share, and validate without sending real traffic.

## One-Click Verification Before Launch

When you hit **Start Test**, Tailor shows a confirmation screen that summarizes everything about your experiment in one place: changes, conversion goals, targeting, URL, and traffic split. You can verify all of this in under 30 seconds before going live.

[Image: Confirm Ramp Details dialog showing page name, changes, conversion goals (click and Amplitude), base URL, targeting params, and variant traffic split before starting a test]

This screen confirms your changes, goals (including Amplitude goals), base URL, targeting parameters, and how traffic is split across variants.

## Pre-Launch QA Checklist

For a more thorough review, run through this checklist before starting any experiment:

### Preview the variant

Use the Tailor extension preview button or append ?preview\_mode=treatment to the URL.

### Check eligibility

Confirm only the intended segment matches your targeting rules. Watch for unintended overlap with other experiments.

### Inspect layout

Check for broken sections, spacing issues, and mobile regressions on both desktop and mobile.

### Speed sanity check

Confirm the tailored page loads quickly (Tailor is async and non-blocking, so load times should be the same).

### Verify tracking

Trigger the conversion goal on both control and variant. Confirm the event fires identically in your analytics.

### Click diagnostics (optional)

Verify CTA clicks are still tracked as expected after DOM changes.

## How Preview Mode Works

Preview mode lets you see any variant without being part of a live experiment. It replicates production behavior exactly, with one difference: preview bypasses CDN caching, so the initial load may be slightly slower than production.

#### Preview URL Parameters

`?preview_mode=treatment`

Shows the tailored version

`?preview_mode=control`

Shows the original version

`?preview_mode={variant_id}`

Shows a specific variant (for multi-variant tests)

Layout, content, targeting, and tracking all behave identically to production. Preview is the most reliable way to QA before launch.

## Sharing Preview Links

### No Extension Required

Preview links work for anyone. The viewer does not need the Tailor Chrome extension installed. Just share the URL with the `?preview_mode=treatment` parameter, and they'll see the tailored version.

The easiest way to generate a preview link: open the Tailor extension, find your variant, and click the preview button (external-link icon). This opens a new tab with the correct preview parameters already set.

## QA Behind Login or Paywall

Tailor operates at the DOM level after the page loads. It doesn't interact with authentication, sessions, or paywall logic. This means:

-   Log in normally, then use the Tailor extension or preview link to QA
-   No special setup needed for authenticated pages
-   Tailor is agnostic to whatever access restrictions your site has

## Forcing Yourself Into the Treatment Group

During testing, you may want to guarantee you see the tailored version:

### Option 1: Tailor Extension (easiest)

Open the extension, find your variant, click the preview button (external-link icon). A new tab opens with the variant loaded.

### Option 2: Manual URL parameter

Add `?preview_mode=treatment` or `?preview_mode={variant_id}` to the URL.

## Troubleshooting: Seeing Control When Expecting Treatment

If you're not seeing the variant you expect:

-   1Try an incognito window to avoid sticky assignment from a previous session
-   2Clear cache and hard refresh (Cmd+Shift+R / Ctrl+Shift+R)
-   3Verify the experiment is active and ramped (not deramped to 0%)
-   4Check that your targeting rules match your current context (UTMs, device, geo)

## Related Guides

### A/B Testing

Set up and manage experiments

[Read more →](/docs/ab-testing)

### Targeting Guide

Configure who sees what

[Read more →](/docs/targeting-guide)

### Troubleshooting

Fix common issues

[Read more →](/docs/troubleshooting)

Ask anything