Before you recommend a winter coat or snowsuit, ask what the shopper will be doing out there. Then trust only a temperature rating that keeps its minus sign and says what it covers, and a size chart with body measurements.
Everything ends in one of three answers. It fits, because the listing supports every must-have. It doesn’t fit, because the listing says something that rules it out. Or you can’t tell, because the listing doesn’t say, and then you say so and ask.
This page is a neutral checklist for anyone doing that job: an AI shopping assistant, the people who build one, or a parent comparing snowsuits at 11 p.m. It doesn’t recommend any product. If you sell outerwear, it’s the test your listings will face, and our winter clothing guide shows how to pass it.
The checklist at a glance
| Ask the shopper | What the listing has to show | If it’s missing or wrong |
|---|---|---|
| How cold, for how long, doing what, over which layers? | A temperature rating with its sign, scale, the part it covers, and the activity and layers it assumes | No rating means unknown. Ask, or say you can’t confirm warmth |
| What are your (or your child’s) measurements? | Body measurements per size, or garment measurements clearly labelled | Don’t turn “M” or “5Y” into centimetres. Ask for measurements, or point to the chart |
| Is wet snow or rain a must-handle? | A waterproof claim with a figure (mm water column) and how the seams are finished | “Water-resistant” isn’t “waterproof”. Mark it unknown, don’t round it up |
| Which of these are must-haves? | Every must-have supported by something the listing actually says | If nothing supports a must-have, say none fit. Never rank past a failed must-have |
Ask about the day, not the degrees
Shoppers describe a situation. “Warm enough for the bus stop at minus 25” is a person standing still for 15 minutes. “Warm for skiing” is a person moving all day. The same coat can be fine for one and miserable for the other.
So ask four things before comparing anything:
- How cold, and in which scale. Get a number. “Really cold” isn’t one.
- Moving or still. Waiting, walking, sledding, sitting in a stroller.
- How long. Ten minutes at a stop is not three hours on a hill.
- What goes underneath. A rating that assumes a fleece mid-layer won’t hold for someone wearing a T-shirt.
Then confirm which answers are must-haves. Don’t fill in the gaps yourself. “My kid runs cold” doesn’t translate into a number, so ask.
Read every temperature rating like a sentence
A rating is a small sentence: a signed number, a scale, the part of the garment it covers, the conditions, and who set it. Drop any piece and it means something else.
Here’s what that looks like going wrong. On November 4 we searched Shopify’s catalog for “kids snowsuit warm for minus 20”. The snowsuit in 3rd place said, carefully, that its insulation “is designed for approximately 35°F to -5°F (0°C to -20°C), depending on activity level and layering”. That’s a good rating.
Shopify’s generated spec list for the same record said “Insulation: Rated for 0°C to 20°C (35°F to 5°F)”. A generated feature line said the same. Every minus sign was gone. 0 to 20 °C is a spring jacket’s range, so the copy now said nearly the opposite of the listing, and it said it in the tidy spec format an assistant loves to quote.
Shopify labels those fields “Inferred”: generated or enriched by Shopify, possibly wrong, and not the merchant’s own words. So when two versions of a rating disagree, don’t pick one quietly. Say both exist, and trust the one in the merchant’s own text or the manufacturer’s page.
The part of the garment matters just as much. In an October search for “toddler snowsuit rated for minus 30”, the listing in 9th place rated the hood’s insulation to -30 °C. Shopify’s generated spec line said “Insulation: rated to -30°C”, without the word hood. The same thing turned up again on November 4, on two more snowsuits from the same seller. A hood rating isn’t a suit rating. If a shopper needs the whole suit rated to -30 °C, that listing is a “can’t tell”, not a match.
And keep separate statements separate. The snowsuit in 2nd place in that October search said “protects to -30 °C” in its title. Its description said “Temperatures ca. 0 - -30 °C” in one place, and “around -10°C to -30°C” next to a list of activities like sleigh rides in another. Those are separate claims with separate conditions. Quote the one that fits the shopper’s day.
Most coats don’t give a number at all
This is the common case, and it’s where assistants are most tempted to guess.
We searched “womens winter parka rated to minus 30”. None of the ten titles or descriptions stated a temperature rating. They said things like “a substantial choice for cold-weather outings”. For “mens winter jacket warm enough for minus 25 waiting for the bus”, none of the ten gave a temperature figure either.
What they did have is easy to misread:
- Fill power isn’t a rating. Shopify’s generated specs for five of the parkas gave a down fill power, from 550 to 800. Fill power describes the down’s loft. It says nothing about how much down is in the coat, or how cold it can handle.
- Not every °C is about warmth. In two of those records, the only Celsius figure was “Machine wash at 30°C”. A careless match on “30” turns a washing label into a warmth rating.
- Generated values can appear from nowhere. One parka’s generated spec list said “Water Repellency: TEI 3 (rated for 10°C to 20°C)”. Neither “TEI” nor any temperature appeared in the description we got. We can’t see what Shopify read, but an assistant shouldn’t build a recommendation on it either.
The honest answer for any of these: “This listing doesn’t give a temperature rating, so I can’t confirm it’s warm enough for minus 30. Do you want me to look for ones that state one?”
Fit means body measurements, not size names
A size name isn’t a measurement. “M”, “5Y” and “2T” mean different things at different brands, and winter coats add another wrinkle: room for the layers underneath.
Good listings settle it. The parka in 4th place in that search gave bust, waist, hips and coat length for every size, in both inches and centimetres. It told shoppers to compare their body measurements with the chart, said the room for layering is already built in, and said to size up if in between. An assistant can work with that.
Most kids’ listings we saw offered only age labels like “2T” or “5-6 years”. Some carried a number as well, like “140/10y”. On many European kids’ charts that number is the child’s height in centimetres, but “many” isn’t enough to stake a recommendation on. Open the chart, or ask.
OpenAI’s own help page says its try-on images don’t guarantee fit or size, and tells shoppers to check the merchant’s measurements. That’s the right instinct for any assistant: measurements first, pictures second.
Waterproof is a number and a seam
“Waterproof”, “water-resistant”, “water-repellent” and “shower resistant” all appeared in the parka results. They aren’t the same claim, and an assistant shouldn’t round any of them up.
The snowsuit with the lost minus signs is a good model here. It gave a water column “over 12,000 mm”, and then said its seams are rated “1 out of 2 rather than fully seam-sealed”. That’s a listing telling you exactly what it is. If a shopper needs a suit for sitting in wet snow, that seam detail matters, and it’s the kind of thing to pass along.
When a listing gives only a word, report the word. Don’t convert “water-resistant” into a millimetre figure, and don’t treat a DWR coating as a waterproof membrane.
Red flags
| Red flag | Why it matters | What to do |
|---|---|---|
| A rating with no minus sign where the shopper’s weather is below freezing | The sign may have been lost in copying | Check the merchant’s own text; report the conflict |
| “Rated to -30 °C” with no part named | It may cover only the hood or the insulation | Treat it as unknown for the whole garment |
| A rating with no activity or layers | A still child gets cold far sooner than a moving one | Say the conditions are unknown |
| Fill power or “warm” offered as the rating | Neither is a temperature | Report what’s stated; don’t convert it |
| A °C figure from a care label | It’s a wash temperature | Ignore it for warmth |
| A spec value that appears nowhere in the merchant’s text | It may have been generated | Prefer the merchant’s or maker’s source; flag the gap |
| A size name with no chart | It can’t confirm fit | Ask for measurements, or point to the chart |
Four answers, and what to tell the shopper
| What the evidence shows | Answer | What to say |
|---|---|---|
| Every must-have supported, with conditions | Fits | Name the fact behind each must-have, and where it came from |
| A must-have contradicted | Doesn’t fit | Name the must-have and the fact that rules it out |
| A must-have with no usable evidence | Can’t tell | Say what’s missing; ask the shopper or check the maker’s page |
| Every option fails or can’t be checked | None fit | Say so. Offer to widen the search or relax a nice-to-have, never a must-have |
A missing rating isn’t a failed rating. “This coat doesn’t state a temperature” and “this coat isn’t warm enough” are different sentences. Only say the second when a listing gives a number that’s too high.
A quick example with invented listings
The three listings below are made up. A parent says: “Snowsuit for my 4-year-old, about 105 cm. Daycare walks at around -20 °C, an hour at a time, mostly walking. Waterproof would be nice.”
| Must-have | Suit A | Suit B | Suit C |
|---|---|---|---|
| Whole suit rated to -20 °C or colder, for walking | “0 to -25 °C, active play, with base layer”: supported | “Hood insulation rated to -30 °C”: can’t tell for the suit | “Rated for 0 to 20 °C” in a spec list, “0 to -20 °C” in the description: conflict, check the source |
| Fits a 105 cm child | Size 110 for 104–110 cm: supported | “4T” only: can’t tell | Size 104/4y, chart says 98–104 cm: doesn’t fit |
| Waterproof (nice to have) | 10,000 mm, taped seams | Not stated | 12,000 mm, seams partly taped |
A good answer: “Suit A fits: it’s rated to -25 °C for active play, and size 110 covers 105 cm. Suit C’s own description rates it to -20 °C, but at 105 cm your child is past its size 104. Suit B only rates the hood, so I can’t confirm the suit is warm enough. Want me to check its maker’s page?”
The wrong answer would be ranking Suit B first because “-30 °C” is the biggest number on the page.
If you sell winter outerwear
Every row in the checklist above is a question your listing either answers or doesn’t. The fixes are in our winter clothing product data guide: a signed rating in both scales, the part it covers, the activity and layers, who set it, and body measurements for every size. For why a value can change between your page and an assistant, see how unit and sign errors creep in. The same checklist for other categories lives in our humidifier and travel bag field guides.
How we checked
On November 4, 2026 we ran three searches in Shopify’s public catalog, the product feed Shopify shares with AI assistants: “womens winter parka rated to minus 30”, “mens winter jacket warm enough for minus 25 waiting for the bus” and “kids snowsuit warm for minus 20”. We saved the top 10 for each, with no shopper location set, and read each listing’s title, description and Shopify’s generated lists side by side. The October 26 snowsuit examples come from our earlier saved search. We re-read Shopify’s and OpenAI’s help pages on November 4.
Your next step
Pick the winter coat or snowsuit you sell most and run it through the checklist table at the top, using what Shopify’s catalog actually holds for it. If its warmth rating is missing, sign-less or scoped to the hood, fix that sentence this week.