Install on ShopifyYou installed virtual try-on. Sessions are climbing. But did try-on actually change shopping behavior — or did a sale, an ad campaign, or seasonality move the numbers? Without a control group, you cannot know.
This guide walks through how to run a clean A/B test on Shopify product pages using Stylab's built-in traffic buckets. No external testing tool required.
Why guessing fails
Common traps merchants fall into:
- Comparing this month to last month while running a promo
- Enabling try-on store-wide and attributing any cart lift to the widget
- Looking only at try-on session count (engagement ≠ purchase intent)
- Stopping after a week when traffic is too thin to learn anything
A/B testing splits visitors on the same product pages into two groups: one sees try-on, one does not. Same traffic source, same pricing, same photos — only the try-on button differs.
How Stylab A/B testing works
From Widget in your Shopify admin app:
- Toggle A/B testing on
- Set the traffic slider (10%–90% of visitors see try-on)
- Each visitor is assigned to a bucket consistently for the test period
The control group sees the normal product page. The test group sees the Try it on button. After enough time, compare buckets in your dashboard:
- Page views per bucket
- Try-on sessions (test group only)
- Add-to-cart count and rate per bucket
Pre-test checklist (do this before you start)
- Pick 2–5 products with steady weekly traffic — not new launches with zero history
- Confirm garment photos are acceptable (upgrade flat-lays on key SKUs if needed — see flat-lay guide)
- Freeze pricing, hero images, and major ad spend changes for 30 days
- Note your baseline add-to-cart rate from Shopify analytics for those URLs
- Estimate expected try-on volume with our usage calculator so you stay within plan quota
30-day test protocol
- Week 0 — Enable try-on on selected SKUs only. Run without A/B for 3–5 days to confirm the widget works and shoppers use it.
- Day 1 of test — Set A/B to 50/50. Do not touch the slider mid-test.
- Weeks 1–4 — Monitor weekly: try-on adoption (% of test-group visitors who start try-on), add-to-cart rate per bucket, failed generations.
- Day 30 — Review results using the decision framework below.
Low traffic? If a product gets fewer than ~500 PDP views in 30 days, extend the test or combine similar SKUs. Thin data produces noisy conclusions — not wrong answers, just unreliable ones.
Metrics that matter vs vanity metrics
Watch these
- Add-to-cart rate — test bucket vs control bucket on the same products
- Try-on adoption — are shoppers actually clicking? Low adoption may mean button placement, mobile UX, or wrong SKUs
- Try-on → cart path — sessions that led to add-to-cart (available in dashboard)
- Failed generation rate — high failures may indicate photo quality issues, not product-market fit
Do not over-index on these alone
- Raw try-on session count without a control comparison
- Store-wide revenue during a sale period
- Social media mentions or press
How to decide after 30 days
Use this framework — not gut feeling:
- Expand — test bucket shows equal or higher add-to-cart rate and meaningful try-on adoption. Roll out to more SKUs; consider 70/30 or 100% try-on on winners.
- Iterate — try-on adoption is strong but cart rate is flat. Improve flat-lay photos, button copy, or product selection before scaling.
- Pause on SKU — near-zero try-on usage after 30 days. The product may not need visualization, or traffic is too low.
- Extend test — inconclusive with thin traffic. Run 30 more days or add products.
Results vary by niche, photography, price point, and shopper demographics. Stylab helps you measure — it does not guarantee a specific lift.