Glow Seoul: how one hero product convinces sensitive skin to buy

glowseoul.my

The Glow Seoul concept demo: warm paper, one vivid orange, and a headline that picks a fight with its own category. Everything after it exists to prove that one claim.

Picture the shopper: she's seen the ampoule on Xiaohongshu, she's interested, but her skin is sensitive and she's been burned before, literally. She won't buy on hype. She buys when the doubts run out. This page is one long, gentle removal of doubt.

01One product, one decision: total focus

This isn't a store with 50 SKUs. It's a landing built around one hero product. That single focus is a feature, not a limitation: every section, every photo, every line is aimed at the same outcome. No category menus, no "shop all" distractions. A confused shopper doesn't buy; a focused one does.

02The hero takes a side: fewer ingredients, not more

The headline is "Sensitive skin doesn't need more ingredients", and the line under it finishes the thought: it needs fewer of the wrong ones. That is not a description, it's a position, and it quietly makes every rival's twelve-active "complex" look like a liability. The useful side effect: it hands the shopper a ruler. Once she believes shorter is safer, she has a way to judge every bottle she sees, and this is the only one built to pass.

03Printed percentages, and a list of what was left out

Most ampoules name their actives and stop there, because the amount is where the story gets weak. This one prints the number next to each: Centella Asiatica 5%, Hyaluronic Acid 2%, Niacinamide 4%. Three lines, and the section heading is simply "Short enough to print" the formula is the proof of the headline.

Next to it sits the move that actually closes reactive-skin shoppers: an orange panel headed "Five things we refused", no fragrance added or "natural", no drying alcohol, no essential oils, no parabens, no dyes. For someone who has reacted before, the absence list matters more than the ingredient list. She is not looking for the thing that will fix her; she is looking for the thing that will not hurt her. The page admits the trade in plain words: fragrance makes a product feel like it's working, and it's also the most common reason reactive skin stops using one.

Below both sits the layer the old version never had: JAKIM halal certification and KKM registration, as four small bordered cards. In Malaysia these are not decoration. They are the two credentials a shopper can actually look up, and the cards say "certified batch by batch, not once in 2019" because a certificate with no date is a claim, not a proof.

04The before/after is the hardest thing on the page

The section is headed "Quiet skin is the result", and it underclaims on purpose: calmer is the claim (not flawless, not transformed) calmer. A sensitive-skin shopper has been oversold to for years. Promising less is what makes the rest readable.

Here is the part worth saying out loud, because it is where the build actually got difficult. A before/after has to be the same person, four weeks apart. Two stock models never are, and the first attempt at this page used exactly that, two unrelated photos side by side, one cool-lit and one warm. It read as a lie the moment you looked at it, because it was one.

So the pair was generated instead, with the subject, the framing, the distance and the lighting all locked, and only the skin allowed to change. The two frames come out within one RGB point of each other in average tone, which is the measurable version of "shot under the same light". Every image on the demo is AI-generated, and the page says so in plain words under the photos: these are not real people and this is not a clinical trial. That disclosure is not a caveat bolted on at the end. It is the reason the comparison is allowed to look as convincing as it does.

The reviews carry the same honesty: the section is headed "The long version is in the reviews" and the line under it admits "we left the four-star ones in". An all-five-star wall reads as curated. One four-star review does more for trust than the twenty fives around it.

05The FAQ asks what she would actually type

The heading is "The fine print, large", and the questions are written the way a worried shopper types them, not the way a brand phrases them. Will it sting?: first, because for reactive skin that is the real question, and the answer names the three things she reacts to and ends "if it does sting, stop and send it back". Then How do I actually use it?, When will I see something?, Is it really halal-certified?, What if it doesn't work for me?

The last answer is the one that actually moves the sale: send it back within 30 days, half empty is fine, full refund, return postage on us. And the "when will I see something" answer points straight at it, "if nothing has changed by week four, it isn't working for you; that's what the 30 days are for." A brand willing to say that out loud is telling you it expects to be kept. Risk reversal beats one more adjective, every time.

06Why warm paper and exactly one orange

Every colour has one job. The background is warm off-white #FDFAF5: paper, not screen. It carries the honesty of the copy: a page that prints its own percentages should look like a document, not a campaign. Soft pink would have been the safe, expected choice for skincare, and that is exactly the problem: the shopper has scrolled past a hundred pink ampoule pages and believed none of them.

Then one vivid orange #F26B1D carries every action and nothing else. Because there is only one accent on the page, the eye never has to choose where to look: the orange is the instruction. Buttons drop to a deeper #C94D08, and that is not a stylistic choice: white text on the bright orange sits around 3:1, which fails, while the deeper one clears 4.5:1 and passes. The reverse trick is on the orange panels, where the body copy is dark ink rather than white: ink on that orange measures 5.7:1, white would have failed. Accessibility decided the shade in both places, and the page looks better for it.

Type follows the same logic: Fraunces, a serif, for the headlines (it reads as editorial, like something written rather than advertised) with italics doing the emphasis. DM Sans handles the body, where the percentages and the certifications live and nothing should get in the way of reading them. The layout is deliberately asymmetric: a full-bleed photo band with a white product card floating on it, a split stat strip, a three-part bento. Nothing is centred, because centred is what a template looks like. (Contrast the restaurant case study's dark, fiery palette: same studio, opposite mood, because the customer's feeling is different.)