How to Measure Whether a Size Recommender Is Working

The only way to know whether a size recommender works for your store is to compare your own returns before and after you add it. This guide shows what to count, how long to wait and what can fool you, so you end up with a number you can trust and not a feeling.

Last updated: November 2026.

Notebook comparing size-related returns before and after introducing a size recommender, alongside a clothing item and testing checklist.

Start with one number

The number that matters is size-related returns per 100 orders. The total return rate mixes in everything else that makes people send clothes back, such as colour, quality or a change of mind, so it moves for reasons a size recommender cannot touch.

This only works if your returns are logged with a reason. Shopify's own return flow and most returns apps can record one, so you can count the ones marked as too small or too large. If you do not record reasons yet, start now, because the before figure has to be written down before you change anything. Our article on why clothes come back as didn't fit explains how to find the real cause behind a return.

What to count

What to count How to count it What you want to see
Size-related returns per 100 orders Returns with a size reason, divided by orders, times 100, for the same number of days A lower number after the change
Total return rate All returns divided by all orders Little change, or a fall that matches the size returns
Exchanges for a different size Exchange requests that name a size, divided by orders A lower number after the change
Shoppers who used it The usage numbers in the app admin, compared with product page visits Enough use for the result to mean something

How to run the check

  1. Write down the before. Take the last four weeks, or longer if you have few orders. Note the orders, the returns with a size reason and the exchanges for another size.
  2. Add the size recommender and change nothing else on those products. New photos, a rewritten size chart or a discount during the test make it impossible to tell what worked.
  3. Wait the same number of days. Four weeks is a sensible minimum.
  4. Compare per 100 orders, not raw counts. Order numbers change from month to month, so a plain count misleads.
  5. Check that shoppers used it. If few did, the result says little about the tool.

What can fool you

  • Sales and holidays. Gift buyers and discount shoppers return clothes for different reasons. Our article on holiday clothing returns explains why a November and December comparison needs care.
  • New products. A new range has its own return pattern, so compare the same products where you can.
  • A change to your returns policy. A longer or shorter return window changes how many people send things back.
  • Small numbers. If you have 50 orders a month, three fewer returns can be chance. Wait longer, or look at your best sellers only.

What a good result looks like

Expect size-related returns to fall, not to vanish. Some people will still order two sizes or change their mind, and a recommender cannot fix a wrong size chart. If the number does not move, check the chart first. Our guide to how accurate a size recommender is lists the checks to run.

Frequently asked questions

How long should I wait? At least four weeks, and longer if your store has few orders. Compare the same number of days before and after.

What if size returns do not fall? Check that the size chart is right for the products that come back most, that the fit of each style is described, and that shoppers are using the tool.

Can I A/B test it? You can show it on some products and not others, but products differ, so comparing the same products over time is usually cleaner.

Do I need an analytics tool? No. A spreadsheet with orders, returns with a size reason and exchanges is enough.

Where do I see how many shoppers used it? In the AI Frame admin, which shows usage for your store such as Size Chart opens and add to cart.

SMART SIZE RECOMMENDATIONS

Help customers find the size that works for them.

Size uncertainty can make online shopping harder than it needs to be. AI Frame uses customer profile details and your product size charts to recommend a suitable size, helping shoppers make more informed purchasing decisions.

Explore Size Recommendations →

Where AI Frame fits

AI Frame adds a Size Chart and a size recommender to your Shopify product pages, and its admin shows how shoppers use them. You can see it on the AI Frame app listing. If you are not sure what a size recommender does, start with what a size recommender is.

Related guides

Back to blog