Questions to ask before choosing a size recommendation tool

When you choose a size recommendation tool, do not judge it only by the return rate of shoppers who followed the advice. Ask for the share of all sales that received a recommendation; the split between shoppers who followed it, ordered bigger or ordered smaller; the return rate in each of those buckets; and the return rate of sales that did not use the tool. The effect on your overall return rate is roughly the share of sales with a recommendation, times the follow rate, times the reduction for followers. If that share is small, even an accurate tool barely moves the total.
Model in a grey tailored suit with subtle size measurement and data overlays

How to choose a size recommendation engine: the figures to demand

Ask every vendor for these figures on your own orders, not as a single marketing percentage. If they cannot or will not show them, you cannot tell whether the tool is changing your overall return rate.

  1. The share of all sales (orders) that received a size recommendation. Why it matters: this is the reach of the tool. Advice that never appears on most orders cannot move your overall refund rate.
  2. Of those sales, the share that followed the recommended size, the share that ordered a bigger size, and the share that ordered a smaller size. Why it matters: only the followers are a clean test of the advice. A headline for “sales with a recommendation” mixes all three groups and hides the dilution.
  3. The return rate in each of those three buckets. Why it matters: you need to see whether followers really come back less often, and what happens when shoppers ignore the advice.
  4. The return rate of sales that did not use the tool. Why it matters: this is the baseline. Without it, a lower return rate among followers has nothing honest to be compared with.

Clothing Size Finder Pro is built to show whether shoppers follow the advice, not to hide it. Its size-demand dashboard covers size uptake, return rates and conversion uplift. When you talk to Prime AI, we will show you what we see on average across these figures, and how they can be moved.

How to compare the impact of size recommendation tools on return rates

The overall reduction in your return rate is, in simple terms, the share of sales with a recommendation × the follow rate × the return-rate reduction for followers.

The table below is an illustration, not a set of Prime AI results. It assumes a 30% baseline return rate for sales with no recommendation, and that shoppers who then ordered a bigger or a smaller size also return at 30%. The 35% reduction applied to followers is already published for Clothing Size Finder Pro: on average, shoppers who ordered exactly the recommended size have a 35% lower return rate than shoppers who did not use Size Finder. Here that is a follower return rate of 19.5%. The first row uses a split of 50% followed, 25% bigger and 25% smaller.

Illustration Share of sales with a recommendation Follow rate (of those sales) Share of all sales that followed Overall return rate Change vs 30% baseline
A 20% 50% followed (25% bigger, 25% smaller) 10% 28.95% −1.05 points
B 50% 50% 25% about 27.4% about −2.6 points
C 60% 70% 42% about 25.6% about −4.4 points

In illustration A, only one in ten of all sales is a followed recommendation, so the overall rate falls from 30% to 28.95%. In illustration C, the same 35% reduction for followers, applied to a much larger followed share, takes the overall rate to about 25.6%. The tool has not become more accurate in the headline sense; more sales were advised, and more shoppers followed the advice.

Why many vendors will not show these figures

A claim such as “followers return X% less” is easy to publish and hard to test. It says nothing about how many sales were advised, or how many shoppers sized differently. Many vendors will shy away from showing the split, because a small sales share or a low follow rate then becomes visible. Retailers should demand the figures anyway. Once you can read those variables, you can have an honest conversation about what is possible on your catalogue.

How to grow the share of sales that get size advice

There is no magic tool that you add to a product page and then forget. Meaningful impact needs sizing tied into the whole journey, not only a widget on the PDP: the product page, the cart, and systems such as CRM and loyalty, which Clothing Size Finder Pro can connect to through its API.

A quiz alone only reaches shoppers who complete it. A zero-input size finder gives known shoppers a size from their past purchases, with no quiz and no tape measurements, from launch day, via their user ID. That is how advice can reach people who never fill in a form, and why zero input is central to growing the influenced share of sales. On Clothing Size Finder Pro, 50%+ of sales get size advice. The follow rate then decides how much of that coverage changes the overall return rate. See also How Zero Input Size Finder drives 50%+ influence and reduces returns et Why you can’t reduce refunds with quiz-based size finders only.

Prime AI also works on the other groups, not only on followers. Dynamic fit messaging can tell a shopper to size up, size down or stay before checkout. Shoppers who skip the quiz can still receive advice through zero-input recommendations based on their purchase history. The work is to raise the share of sales with advice, raise the follow rate, and keep the advice accurate in every bucket.

Questions fréquemment posées

How do I choose a size recommendation engine?

Start by asking for the figures above, on your own orders. Then ask how the vendor grows that sales share, including whether known shoppers get a size with no quiz, and how the tool fits into the rest of the journey.

Why is a lower return rate on recommended sizes not enough?

Because it only describes the shoppers who both saw the advice and followed it. If that group is a small share of all sales, the overall return rate barely moves. You also need the sales share, the follow rate, and a baseline for shoppers who did not use the tool.

What is a zero-input size finder?

A zero-input size finder gives a known shopper a size recommendation from their past purchases, with no quiz and no tape measurements. On Clothing Size Finder Pro, this starts from launch day, once the shopper’s user ID is passed through. New shoppers still answer a short quiz once, and are remembered after that.

How do clothing size recommendation tools reduce overall returns?

They reduce overall returns only to the extent that advice reaches a large share of sales, shoppers follow it, and those orders come back less often than the baseline. In the illustration above, raising the advised share and the follow rate does more to the overall rate than the 35% follower reduction can do alone.

What is follow rate, and why does it matter?

Follow rate is the share of sales with a recommendation in which the shopper ordered exactly the recommended size. The rest of those sales typically came in bigger or smaller than advised. A higher follow rate means more of the advised sales can produce the lower return rate you were promised.

Talk to Prime AI about the figures that matter

If you want to see what these numbers look like in practice, talk to us. We will show you what we see on average, and how Clothing Size Finder Pro can move sales share, follow rate and the return rate in each group, including shoppers who never complete a quiz. See Clothing Size Finder Pro, ou request a demo on your own catalogue.

Parlez-nous aujourd'hui