Skip to main content

​
How do I A/B test Looksy?

Split eligible traffic between a control experience and one defined Looksy variation, then compare a preselected primary metric. Use an experimentation tool that is compatible with your current Shopify setup.

​
Define the test before it starts

Record:
  • the question and hypothesis;
  • included products, traffic, and devices;
  • control and treatment experiences;
  • primary metric and guardrails;
  • planned duration and stopping rule;
  • exclusions and data-quality checks.

​
What can I test?

  • whether Looksy is shown for a selected product group;
  • product-image button wording, corner, or offset;
  • instructions before upload;
  • product-selection strategy;
  • mobile-specific presentation.
Change one main variable at a time. If several things change together, the result will not show which change mattered.

​
What should I measure?

Choose one primary metric, such as try-on completion, add-to-cart rate, or conversion rate. Monitor guardrails including page performance, errors, support questions, and returns.

​
How should I interpret the result?

Report counts, rates, sample size, uncertainty, dates, and implementation details. Check whether promotions, stock, traffic mix, or seasonality affected one group differently.
Example percentages are not expected Looksy results. The outcome must come from your own experiment.

​
Common mistakes

  • ending the test when an early result looks favourable;
  • testing too many changes at once;
  • comparing different product or traffic mixes;
  • ignoring device or variant failures;
  • treating correlation as causation;
  • reporting a percentage without underlying counts.

Conversion Tracking

Define consistent events before the test.

Calculating ROI

Translate a measured effect into a financial estimate.