You can’t match a lipstick shade by eye. Find its name.

August 17, 2026 · 7 min read

The colour you are looking at is not the colour of the lipstick. So stop trying to match it. Find the name instead, from the on-screen text, the audio, a pinned comment, or the base of the tube when it turns toward the camera. If no name exists anywhere in the video, accept a shade family rather than a shade, and buy it somewhere that takes returns.

That is the answer. The rest of this is why colour-matching fails so reliably, and where the name is usually hiding.

Four distortions, stacked

People assume the video is a slightly imperfect photograph of a lipstick. It is not. Four separate things change the colour between the tube and your eye, and they compound.

• The light and the filter.Ring lights, window light, and warm presets each set a different white balance for the whole frame. This is why the same lipstick reads coral in one creator's video and brick in another's.
• The compression. Delivery video stores colour at lower resolution than brightness. In the 4:2:0 scheme used almost everywhere, colour is sampled at half the horizontal and half the vertical resolution of the luminance. Artefacts show up worst at high-contrast colour edges, and the border of a lip against skin is precisely that.
• The lips underneath.Pigment on a mouth is not pigment on paper. It is sitting on somebody else's natural lip tone, which is acting as a base coat you do not have.
• Your screen. Two phones side by side will render the same frame at different warmth and saturation.

None of those four steps is reversible from the finished video. You are not looking at a slightly wrong colour. You are looking at a colour that has been through four transformations, and no amount of staring undoes them.

It is very often not one product

Here is the part almost nobody writing about this mentions. The lip in the video is frequently a stack, not a lipstick.

The lip combo is its own named trend, popular enough to have a hashtag: a darker liner to define and reshape the lip line, a lipstick or tint for colour, and a gloss or oil on top for shine. Three products, layered, and the finished colour belongs to none of them individually.

So “what lipstick is that” can be the wrong question. You can identify the lipstick correctly, buy it, put it on, and still not get the thing you saw, because you are missing the liner underneath and the gloss on top. If the video shows a defined edge darker than the middle of the lip, assume a liner. If it catches the light in a wet band across the centre, assume a topper. Then go looking for three names instead of one.

Where the name is actually hiding

In rough order of how often it works, and how little time it costs:

• On-screen text. Free, instant, and routinely ignored because it flashes for half a second. Scrub frame by frame through the first three seconds and the last three.
• Captions and audio. Turn captions on and rewatch. Shade names get said out loud, fast, and often over background music. Fifteen seconds of your time.
• The pinned comment and the replies. Creators answer the shade question once and then stop, so the answer is usually sitting under an older video rather than this one. We wrote up why that happens in why creators never answer the link comment.
• The tube itself. The shade name or code is printed on a small label on the base of the container. Find the frame where they cap it, set it down, or hold it up, and pause there. The print is small and the surface is curved, so it fails as often as it works. When it works, it is the only piece of evidence in the whole video that is unambiguous.
• The arm swatch. If they swatched it before applying, that stripe is a better sample than the lip. It is pigment over a flatter, more even surface. Skin is still not paper, and the frame is still filtered, so this narrows the field rather than settling it.

If you must judge by colour, calibrate first

Sometimes there is genuinely no name anywhere. In that case, find something in the frame you already know the colour of. A sheet of paper, a white t-shirt, a plain ceramic mug, the whites of their eyes. If those render yellow, the whole frame is pushed warm and the real lipstick is cooler than it looks. If they render blue, the reverse.

This does not give you a shade. It gives you a direction of error, which is enough to strike half your candidate list. Two honest limits: filters do not shift every colour by the same amount, and beauty filters frequently touch lips specifically, which breaks the assumption that the whole frame moved together. Use it as a tiebreak between two shades, never as evidence on its own.

The same caution applies to running a screenshot through visual search. It is a real workflow and it often gets you the right tube. Shade is where it falls down, for the reasons set out in our piece on Lens and video.

A sibling shade is a wrong answer, not a close one

This is the part worth being blunt about. MAC's nudes are the standard illustration. Velvet Teddy is a deeper neutral beige. Honeylove is lighter, with a warm rose undertone. Kinda Sexy reads as a neutral pinky rose. Same brand, same matte finish, same tube silhouette, three visibly different results on a face.

Get it wrong and you find out at the moment you put it on, which is the one test that counts and also the point of no return at most sellers. Some beauty retailers do take opened products back, and a few are generous about it. Plenty of brand sites and marketplace sellers are not. Which camp you are in is decided before you order, not after, so read that policy while you are still deciding rather than while you are holding a shade you will not wear.

When to stop looking and take a match

At some point the evidence runs out. Do not keep staring at the frame. Switch to somebody else's swatches.

Swatch databases are built for exactly this. Temptalia's Dupe List, for instance, lets you search a product and see alternatives with the original and the dupe photographed side by side, and it distinguishes matches the site has swatched and tested itself from ones voted on by its community. Two swatches shot under one consistent light is far better evidence than your eye on a compressed frame, whatever your eye is telling you.

Then buy the shade family, not the shade, and buy it where you have read the returns policy first. That is a real answer to the question, just not the one people want.

Where Stealsy fits, and where it stops

Stealsy watches the whole video rather than one frame of it. It reads the on-screen text, listens to the audio, and takes the caption into account together, then hands back the products with prices and shop links that have been opened and checked before you see them, plus the cheaper dupes where they exist. Heart something and it watches the price and tells you when it falls.

What it will not do is invent a shade name the video never contained. If nobody says it, no text card shows it, and the tube never turns its base to the camera, the evidence for a specific shade does not exist, and the honest output is the product line. A confident wrong shade is worse than a blank one. You would act on it.

Get Stealsy for iPhone — share the video, get the products.