How to measure virtual try-on performance

Quick answer

to measure a virtual try-on, follow the full journey: how many visitors see the button, how many use it, for how long, what they try, how many buy and how many return items. To know whether the try-on *causes* improvements (and does not only attract more motivated buyers), use an A/B test with split traffic.

By Jason Leven, ConsultantUpdated 19 October 20267 min read

Dashboard with usage rate, conversion and returns for a virtual try-on

The virtual try-on funnel

Visitors to the product page with the try-on available.
Activation: those who press "Try on".
Effective use: they grant the camera and complete a try-on.
Interaction: they try several references, capture or share.
Purchase: they add to the basket and complete the purchase.
After-sales: returns, reviews, repeat purchase.

Each step has a rate and can be optimised.

Key metrics

Adoption

Activation rate: presses ÷ visitors to the product page.
Technical success rate: completed try-ons ÷ try-ons started (important for detecting camera or compatibility failures).
Percentage of sessions with the try-on.

Engagement

Average time of use.
Products tried per session.
Captures and shares.
Colour or variant changes.

Business

Conversion of sessions with the try-on and without the try-on.
Average order value.
Adds to basket from the try-on.
Return rate by product or category.
Attributable incremental revenue.

Experience

Satisfaction (a quick question after the try-on).
Complaints or issues from customer service.
Technical performance: load times, errors by device.

Watch out for bias

Someone who uses the try-on is usually already more motivated to buy. If you simply compare users with and without the try-on, the effect will look enormous, but part of the difference is due to prior intent.

How to control it

A/B test: randomly split the traffic; show the try-on to one group and not to the other.
Comparable cohorts: compare users with a similar profile (source, device, history).
Equivalent periods: take seasonality and promotions into account.
Segmentation: analyse by category and by customer type.

Designing an A/B test

Hypothesis: "The try-on will increase conversion for sunglasses."
Primary metric: conversion from the product page to purchase.
Secondary metrics: order value, returns.
Sample size and duration: enough to obtain reliable results.
Correct randomisation.
Analysis with statistical judgement.
Decision: expand, adjust or discard.

Technical analytics

Define events such as:

tryon_button_view
tryon_start
tryon_camera_granted / denied
tryon_product_change
tryon_capture / tryon_share
tryon_add_to_cart
tryon_error

Label them with product, category, device and browser.

Return on investment

ROI = (incremental profit − total cost) ÷ total cost

Benefit: incremental revenue from higher conversion and order value, and savings from fewer returns.
Cost: model production, licences, integration, maintenance and promotion.

Recommended monthly report

Section	Content
Adoption	Activation, technical success
Engagement	Time, products tried
Business	Conversion, AOV, returns
Quality	Errors, devices, satisfaction
Learnings	What works, what fails
Next steps	Improvements and expansion

Common mistakes

Comparing users with and without the try-on without controlling for bias.
Measuring only use, without the business side.
Not recording technical errors.
Evaluating too soon, or with too little traffic.
Not segmenting by category.

Key ideas

Measure the full funnel, from the view to the return.
Use an A/B test to avoid selection bias.
Also record technical quality.
Analyse by category and device.
Calculate ROI with incremental revenue and savings on returns.

Frequently asked questions

How much traffic do I need for an A/B test?

It depends on your conversion rate and the expected effect. With little traffic, extend the length of the test.

What do I do if I cannot run an A/B test?

Compare comparable periods and similar cohorts, and be cautious when interpreting.

What activation rate is good?

It depends on the button, the sector and the category. Compare it with your own improvements.

How do I measure the effect on returns?

By cross-referencing try-on use with returned orders and their reasons.

Related services

Virtual try-on