How the Size Recommender Works

The size recommender turns a shopper’s numbers into a size suggestion — with the reason shown, not a black box. It’s included on every plan, including Free, and runs on the chart data you already entered: no extra setup beyond a normal chart.

What shoppers see

On the product page, the widget shows two tabs — Calculate my size (default) and Size chart:

1. The shopper opens Find my size.
2. They enter either:
   • Height and weight — the two numbers almost everyone knows, or
   • Their own measurements — bust, waist, hip (a link switches modes).
3. They see the best size, the reason (“bust 88 falls in M (86–91)”), a fit breakdown of every size, and — if the theme’s variants map — a Select size M button that picks the size in the variant picker. The shopper still reviews and adds to cart themselves; the app never adds to cart for them.

Everything runs in the shopper’s browser on the table already on the page — no server round-trip per recommendation.

Why height & weight first

Most shoppers don’t carry a tape measure. Height and weight get them a solid estimate immediately; those who know their measurements can switch to the exact mode for tighter results. Height & weight results are labelled Estimated from height & weight so the distinction is honest.

Fit breakdown: every size gets a verdict

Instead of one answer, the recommender scores every row of your chart:

Verdict	Meaning
Perfect	The shopper’s numbers fall inside this size’s ranges
A little loose	Slightly above the ranges — still wearable
Too loose	Well above the ranges
Too tight	Below the ranges — not wearable

The size with the highest score wins; ties go to the larger size (a slightly big garment beats a too-small one).

How scoring works (published method)

For each measurable column you enabled in the editor (“Size recommendations use”), the app compares the shopper’s number against the row’s range [lo, hi] and converts the distance into a score. The scoring is deliberately asymmetric:

• Being too tight is penalised roughly twice as hard as being too loose.
• Rationale: a loose garment can still be worn; a tight one can’t. Real-world returns follow this pattern.

For garment charts, numbers are normalised to body terms first: each measurement has ease subtracted (defaults: bust/chest −10 cm, waist −6 cm, hip −8 cm) and the range’s upper bound is stretched by 10% to account for fabric give. Body charts are matched as-is.

Note

We publish this method because a recommender that hides its math can’t be trusted — every recommendation on your storefront shows its measurements, ranges, and reasoning.

Setting it up

1. In the chart editor, make sure the table has at least one measurable column (cm, in, kg, lb) — only those can be scored.
2. Under Size recommendations use, tick the columns to score (e.g. Bust and Waist, but not “Fit note”).
3. Enable Find my size in Appearance → What shoppers can do.
4. Test with example shoppers in Fit Lab before publishing.

Privacy

The quiz asks for numbers only — no name, no email, no payment data. Recommendations are computed in the shopper’s browser; nothing is stored against a person.

Video Tutorial

Play

Related Docs

• Fit Lab — test recommendations before publishing
• Appearance and the master template
• Chart editor guide