Conversion Reports¶
A Conversion Report points a small named panel of synthetic shoppers at one page on your site and shows you where the funnel leaks. There is no second version, no winner, and no confidence or lift — it is a single-page diagnostic, not a comparison.
Use a Conversion Report when you want to understand how one page is performing. Use an A/B test when you already have a fix in mind and want to know whether it converts better than what's live.
Conversion Report vs A/B test¶
| Conversion Report | A/B test | |
|---|---|---|
| Question it answers | Where does this page lose shoppers? | Does version B beat version A? |
| Pages you supply | One | Two — Control page (A) and Variant page (B) |
| Panel size | 2–10 named shoppers you select by archetype | Large statistical sample |
| What you get back | A funnel, ranked fixes, leak-by-segment, and shopper quotes | A verdict, lift, and confidence |
| When to use it | You want to diagnose a single page | You have a change to test |
Your named panel¶
Instead of a large anonymous sample, a Conversion Report sends a small named panel — 2 to 10 shoppers — through your funnel. You choose the panel when you set up the report:
- Archetypes — pick from six shopping styles: Budget hunter, Deep researcher, Impulse buyer, Brand loyalist, Skeptic, Gift buyer. The panel is assembled from your chosen archetypes.
- Panel size — three tiers to match your need:
- Quick read — 3 shoppers. Fastest, cheapest, rough signal.
- Standard — 6 shoppers. Recommended for most reports.
- Deep panel — 10 shoppers. Slowest, most thorough.
Each shopper has a name and a hue. You can watch any one of them individually, so the panel gives you both aggregate signal and individual evidence.
What a finished report shows¶
Squoosh crawls the page you point it at to map the funnel it finds — for example product, add to cart, checkout, confirmation — then sends your named panel through that funnel to find where they drop off. The report reads top to bottom:
- The headline leak. The single step where the funnel loses the most shoppers, stated plainly: where the biggest break is, the drop percentage, and how many shoppers completed the funnel.
- The funnel step by step. Each step shown as a bar with its count and share. The largest consecutive drop is tagged Biggest leak.
- Leak by segment. A matrix showing how the leak pattern differs across Shopping style, Device, and Traffic source.
- Who dropped and why. A per-shopper ledger of where each shopper left and their reason, each with a Watch action to replay that session.
- What to fix first. Ranked fixes grounded in the real leak and friction, with an impact-versus-effort view to help you prioritize.
- Clustered verbatims. Shopper quotes grouped by the step where each shopper stalled.
- Run-over-run trend. How the biggest-leak percentage has changed across prior reports on the same page, so you can see whether the problem is improving.
The report counts only genuine drop-offs. A shopper whose browser session failed is excluded, not reported as a leak.
Note
A Conversion Report needs a funnel with more than one step. If Squoosh can't find a multi-step funnel on the page, the report tells you so rather than inventing one.
Run a Conversion Report¶
- In the sidebar, click Conversion Reports.
- Click New conversion report.
- Enter the Target URL — the page you want to diagnose. A Conversion Report never asks for a second version.
- Under Personas, pick the archetypes you want in your panel and set the count for each.
- Under Panel size, select a tier: Quick read (3), Standard (6, recommended), or Deep panel (10). The wizard shows the shopper count, credit cost, and estimated time.
- Under Shopper tasks, review the objectives Squoosh drafted from your page. Edit or remove any task; you need at least one.
- Click Launch report.
Squoosh crawls the page, assembles the named panel, opens browser sessions, and walks the panel through the funnel. When it finishes, the report opens with the headline leak and the full evidence.
Note: You must calibrate your synthetic shoppers before you can launch. If you haven't, the wizard prompts you with Set up shoppers. See Synthetic shoppers.
Cost and stopping¶
Each Conversion Report draws from your workspace's credit balance. The wizard shows the exact cost and your remaining balance before you launch. A larger panel takes longer and costs more; a smaller one is faster and rougher. See Credits and tests.
The report finishes when the full named panel has walked the funnel. There is no statistical stopping rule — with a panel of 2–10 shoppers, every shopper walks.
What to do after a report¶
A finished report has two actions:
- Re-run report runs the same Conversion Report again on the same page. Useful after you change the page or want to see whether the leak has improved.
- Export / Share (where available) lets you save or share the report.