Offering Product Customization Before Buying A/B Test

Requests for a personal touch tend to surface in support tickets about colors or engraving, on-site searches for personalized versions, or surveys where that kind of request keeps coming up; those are the cue for this test. If the idea holds, some visitors in the variant complete a customization, while those who want the standard product still add it without getting stuck. If add-to-cart is low because of price or missing product information, look into that first. Make sure before launch that every combination on offer can be produced and fulfilled from stock, and that the preview matches the product that will be made.

Add-to-cart rate

Compare all successful add-to-cart actions per assigned visitor, including standard product additions in the customization arm.

Also track

• Customization completion rateReport completion among visitors who start customizing and the number of completed customizations per assigned visitor.
• Return rateCompare returns over the same follow-up period and note any differences in return policy.
• Cart abandonment rateUse the same definition for abandonment among visitors who add an item to the cart.

• ChangeAdd one predefined product customization option to the variant.
• Keep constantKeep the price, base product option, CTA and other product information unchanged.
• BehaviorReview customization use alongside purchases.
• Purchase pathMake sure customization does not block adding or buying the standard product.
• Separate testsTest the number of steps, preview or default selection separately.

1. 1Judge the variant by the add-to-cart rate including standard product adds, and report customization completion both among starters and across all assigned visitors. If adds rise while cart abandonment rises too, customization may be losing people along the way, so the rise alone does not justify shipping the variant.
2. 2Explain the actual return terms before purchase, and do not assume every customized item is non-returnable. Give both arms the same time for returns to come in before comparing them, and note any policy difference in the result. The finding applies only to the customization option you tested.
3. 3A slow-loading customization flow can cause some variant visitors to drop out of tracking, so check that the traffic split matches the plan (sample ratio mismatch). If the add-to-cart gap is small or unclear, use the sample size calculator to see how many visitors per arm the effect you care about requires, and if your sample falls short, do not treat the result as settled. Try step count or the preview in separate experiments.