Shoes are the item people most often want and least often can name. They are also almost always the smallest thing in the picture.
Those two facts are related, and fixing the second one fixes most of the first.
Why do shoe searches usually fail?
In a normal full-body photo, a coat might take up a third of the image. The shoes take up a fortieth of it.
Snagfit describes each garment in words before it searches, and it can only describe what is visible. At a fortieth of the frame, a shoe is a shape and a colour. "Black low top sneaker" matches thousands of listings.
Crop so the shoes fill most of the picture and the same shoe becomes a toe shape, a sole thickness, a panel arrangement and a material. That description narrows to something you can actually shop.
Should you crop to just the shoes?
If the shoes are what you want, crop to just the shoes. Snagfit files footwear as its own category, so it does not need the rest of the outfit for context, and a crop containing only the feet produces a far better description than the same shoes buried in a full-body shot.
Search the rest of the outfit separately from a second, wider crop. Two searches from one screenshot is normal and takes about twenty seconds.
Which angle works best for shoes?
The silhouette identifies a shoe more than any other single thing, and the profile is where the silhouette lives.
A side view shows the toe shape, the sole thickness, the heel height and the ankle line all at once. A shot from directly above shows the laces and hides everything that matters.
If you are pulling a frame from a video, scrub for a moment where a foot is turned side on. It is worth ten seconds of scrubbing.
Do you need both shoes in the photo?
You do not need both feet. A single well-lit shoe in profile carries most of the information, while two shoes caught mid-stride and both blurred carry much less. Pick the frame where one foot is still and clear.
Will Snagfit name the brand?
Snagfit will name a brand when the branding is genuinely visible, or when the silhouette is distinctive enough to be identified on shape alone. Plenty of sneakers are.
When there is no visible logo and the shape is common, the brand is not in the photo. Snagfit describes the shoe rather than guessing a label, because a guessed brand sends you to the wrong shop with confidence, which is worse than sending you nowhere.
Why do all the shoe results look the same?
This is the usual complaint about shoe searches, and it has a real cause: an enormous number of shoes share a silhouette. "White leather low top sneaker" is a genuine description of thousands of products.
What narrows it is the distinguishing detail: a contrast sole colour, a specific panel, a visible stitch line, an unusual eyelet arrangement. Those only help if they are visible in your photo, which brings it back to the crop.
If the results are all near-identical, the answer is usually a tighter crop or a better frame rather than a different search.
Why are feet so badly lit in most photos?
A physical reason shoe searches disappoint, and it is not about the camera.
Light generally comes from above: ceiling lights, the sky, a window. That means the top of a person is lit and the bottom of them is in relative shadow. Add a dark floor absorbing what light does reach down there, and shoes are routinely the darkest object in an otherwise well-lit picture.
Dark means the colour reads wrong and the texture disappears. A brown suede boot in floor shadow describes as a black boot, and the search goes looking for the wrong thing.
So when choosing a frame, look specifically at the feet rather than at the overall exposure. A photo that looks too bright overall often has the best-lit shoes in it, and that is the one to use.
What goes into a shoe description?
Knowing what Snagfit is trying to capture tells you what your photo needs to show.
The silhouette. Low top or high top, the toe shape, how bulky the sole is, where the ankle line sits. This is the single most identifying feature and it lives in the profile.
The upper material. Leather, suede, canvas, mesh, knit. Each looks different and each narrows the results.
The sole. Thickness, colour, whether it is a contrast colour, whether there is a visible tread or a stacked heel. A contrast gum sole is one of the most useful details a photo can contain.
Panelling and hardware. Overlays, stitch lines, eyelet arrangement, straps, buckles. On a sneaker this is often the only thing separating one model from dozens of similar ones.
Colour. The main colour plus any contrast. Which is why a shoe in shadow is a much worse input than the same shoe in daylight.
The shoes naming guide covers the words shops actually use for silhouette and sole type.
Why check that both shoes match?
A small sanity check that saves a confusing result.
In a frame caught mid-stride, one foot can be turned so far that it reads as a different shape from the other. Snagfit describes the footwear from what it can see, and an odd angle can produce an odd description.
Pick a frame where at least one shoe is clearly and normally oriented, and crop to that one.
Which shoes are easy to find, and which are hard?
Not all footwear behaves the same way, and it is worth knowing which you have.
Easy: distinctive sneakers. Many sneaker models have a silhouette recognisable on shape alone, so even a logo-free side view gets somewhere.
Easy: statement boots and heels. An unusual heel shape, a specific shaft height, a distinctive toe. These narrow fast.
Harder: plain leather shoes. A black leather derby looks like every other black leather derby at 1280 pixels. Expect a lot of near-identical results and judge on price and shop rather than on finding the exact one.
Harder: minimal white sneakers. An enormous category of near-identical products. The distinguishing detail is usually tiny, so the crop has to be correspondingly tight.
Hardest: anything in motion at the bottom of a video frame. Feet move more than any other part of a person, so they are blurred more often.
Does a pair of shoes count as one item or two?
A common misunderstanding worth clearing up.
Snagfit is describing a garment, not counting objects. A photo of a pair returns one Footwear item, described from what it can see, because a pair of shoes is one product.
So there is no advantage to including both feet, and there is a real disadvantage if including both means each is smaller and one is blurred. One still, clear, side-on shoe is the better input every time.
How do you photograph your own shoes?
If you are searching a pair you already own, you control the conditions entirely, so use that.
Put the shoe on a plain surface against a plain background. Daylight near a window, not overhead artificial light, which flattens texture. Photograph it side on, at roughly the shoe's own height rather than looking down at it. Fill the frame.
That takes about fifteen seconds and it is the best possible input to a shoe search. Compare it to a frame grabbed from a video and the difference in the description you get back is obvious.
Does the search know your shoe size?
Worth stating because it causes real problems with footwear specifically.
Snagfit identifies the shoe. It has nothing to say about which size will fit you, and shoe sizing varies enormously between brands and between countries.
When you land on the retailer's page, read their own size guidance and any notes from buyers about whether the model runs large or small. On a shoe, this matters more than on almost any other garment, and it is the reason a shop with easy returns is often worth more than one that is cheaper. That is part of checking out a shop you do not recognise.
How do you replace a pair you already own?
This is one of the more practical uses. Photograph your own shoe side on, in daylight if possible, against a plain background. Crop tightly. Search it.
You get the garment description and current listings for the same silhouette, which is useful even when the original model is discontinued. That is the normal case for shoes, since models get replaced every year or two, and a close match is often the honest answer.