Camille Roux - Blush

Your Skin Barrier Is Not A Wall, And It Does Not Need Rebuilding Daily

Camille Roux — Thu, 27 Aug 2026 00:00:00 GMT

Four years ago almost nobody outside a dermatology clinic used the phrase “skin barrier”. It is now printed on body wash. The speed of that adoption should tell you something: a term that useful to marketing is rarely still doing the work it was coined to do.

What the phrase describes is the stratum corneum — the outermost fifteen or so layers of dead keratinocytes, held in a lipid matrix of ceramides, cholesterol and free fatty acids. The usual analogy is bricks and mortar, and that is where most of the confusion starts. A wall is either intact or breached. Your stratum corneum is neither. It is a gradient, shedding at the top and rebuilding from below, and it tolerates a great deal of daily insult without any of it counting as damage.

What a compromised barrier actually looks like

Not a bad skin day. Trans-epidermal water loss rises measurably when the lipid matrix is disrupted, and the visible consequences are specific: stinging on application of products you tolerated last month, a fine flaky texture that makeup catches on, tightness that has not settled twenty minutes after cleansing, and usually a low-grade redness that will not calm down.

Those four signs travel together. One of them on its own is almost always something else.

If you have all four, the useful response is subtraction. Stop the acids. Stop the retinoid. Stop anything carrying fragrance high in the ingredients list. Wash with something that barely foams, apply a plain moisturiser that combines an occlusive with a humectant, and wear sunscreen. Two weeks of that fixes most of what people spend two hundred pounds trying to fix.

The part the category gets right

Ceramide-containing moisturisers do work, and the evidence for them is better than the evidence for most things sold in the same aisle. Formulations that supply ceramides alongside cholesterol and fatty acids in roughly physiological proportions outperform ceramides alone — a finding now over thirty years old, and still ignored by half the products claiming it.

Occlusives work too, and they are cheap. Petrolatum reduces trans-epidermal water loss by upwards of 95 per cent, which nothing sold as a serum in any price bracket comes close to. That it is unglamorous is not a scientific objection.

The part it gets wrong

Barrier repair is now sold as maintenance rather than as treatment, and that is the sleight of hand. Skin that is not compromised does not benefit from being treated as though it is. The barrier-support routine most people are sold — a cleanser, an essence, a serum, a cream and an overnight mask, all making the same claim — introduces more variables and considerably more fragrance than the single moisturiser it replaced.

There is a circularity here worth naming out loud. A great many barriers are compromised in the first place by routines bought to protect them: too many actives, too often, layered by someone following a video rather than a face. The category manufactures a share of its own demand.

What we would actually keep

One gentle cleanser. One moisturiser with a sensible lipid blend, used consistently rather than dramatically. Sunscreen every morning. And one active, introduced slowly enough that you can still tell what it is doing.

Everything else is optional, and most of it is optional in the way a second dessert is optional. Pleasant. Not load-bearing.

Double Cleansing Only Earns Its Second Step If You Wear Sunscreen

Camille Roux — Mon, 27 Jul 2026 00:00:00 GMT

Double cleansing arrived from a tradition where it made obvious sense: a first step that dissolves what is oil-soluble, a second that lifts what is water-soluble, in a routine that assumed a full face of makeup and a mineral sunscreen. Detached from those assumptions it became a rule, and rules travel further than the reasons behind them.

So here is the reason. A surfactant cleanser emulsifies oil, but only so much of it, and only if it is given time. Silicone-heavy makeup, water-resistant sunscreen and zinc oxide are all difficult for a gentle face wash to shift in the thirty seconds people actually spend washing. An oil or balm first step dissolves those into something the second cleanser can carry away. That is the whole mechanism.

When the second step is doing work

If you wear a water-resistant or mineral sunscreen, you need it. Zinc oxide is deliberately designed to sit on the surface and resist water, and it does not come off with a foaming wash — it comes off partially, which is worse, because you finish with a film you cannot see and a moisturiser going on over it.

If you wear long-wear foundation, or any product marketed as transfer-proof, you need it. Those films are built to survive an evening.

If you have oily skin and a heavy sebum day, an oil-based first step genuinely outperforms a second pass with the same gel cleanser, because like dissolves like.

When it is a wasted step

Bare skin, or bare skin plus a moisturiser and a chemical sunscreen. One well-chosen cleanser removes all of that comfortably, and the second pass is thirty extra seconds of surfactant contact with your face for no return.

This is not a neutral cost. Repeated cleansing raises skin pH and strips intercellular lipid; the more often you do it and the longer the contact, the more the barrier has to spend the night catching up. Anyone with the specific pattern of tightness-then-oiliness that gets diagnosed as “dehydrated combination skin” should try halving their cleansing before they buy anything for it.

Doing it properly if you are doing it

Both steps go onto dry skin, or the first one does. An oil or balm applied to a wet face partially emulsifies on contact and stops dissolving anything. Sixty seconds of dry massage, then wet hands, then emulsify.

Choose the second cleanser for gentleness, not for a second round of cleaning. Its job is to remove the first product, and the most common mistake is pairing a good balm with an aggressive foaming wash — at which point you have done the harsh cleanse anyway, with an extra step in front of it.

And skip the whole thing in the morning. There is nothing on your face at seven in the morning that requires two cleansers, and for most skin types there is nothing that requires one.

The uncomfortable summary

Double cleansing is a technique for removing specific things. It is not a virtue, it is not a baseline, and it is not more thorough in a way that helps if there is nothing there to remove.

Look at what is actually on your face at nine in the evening. That answers it.

Retinol Is A Long Game And Your Face Keeps The Score

Camille Roux — Mon, 13 Jul 2026 00:00:00 GMT

Retinoids are the second-best-evidenced thing in this magazine, after sunscreen. They are also the category with the highest abandonment rate, and those two facts are related: a product that takes three months to demonstrate anything is competing for shelf space against products that appear to do something on the first night.

The conversion problem

Not all retinoids are the same molecule and the difference is not a marketing tier. Tretinoin is retinoic acid — the active form, which binds directly to nuclear receptors and needs no conversion. Everything sold over the counter is a precursor that has to be converted in the skin to get there.

Retinaldehyde converts in one step. Retinol takes two. Retinyl palmitate and the other esters take three, and each step loses a great deal of material. The practical consequence is that potency runs roughly tretinoin, then retinaldehyde, then retinol, then the esters — and that a 1 per cent retinyl palmitate is not remotely comparable to a 0.3 per cent retinol, whatever the number on the front suggests.

The newer retinoid esters that bind directly, granactive-type compounds among them, sit somewhere between retinol and retinaldehyde and are considerably better tolerated. They are a reasonable starting point for anyone who has failed at retinol twice.

The schedule is the product

Two nights a week for four weeks. Three nights a week for four weeks. Then, only if there has been no persistent flaking or stinging, every other night.

Most people go nightly on day one, spend two weeks visibly peeling, conclude their skin cannot tolerate retinoids, and stop. The retinisation period is real — the initial adjustment involves genuine barrier disruption — but it is proportional to how fast you push, and the endpoint is the same either way. Going slowly costs six weeks and saves the attempt.

Apply to dry skin, at least twenty minutes after cleansing. Damp skin increases penetration, which sounds like an advantage and is in fact the most common cause of irritation on a first attempt.

What to expect, and when

Weeks one to four: nothing good. Possibly some purging, which is genuine — accelerated turnover brings existing microcomedones to the surface — and which is distinguishable from a reaction because it happens where you normally break out.

Weeks four to twelve: texture. This is the first real change, and it is the one people miss because it is gradual and they are looking for something else.

Weeks twelve to twenty-four: tone, and the beginning of fine-line change. The collagen data underlying the anti-wrinkle claims comes from studies running six months to a year, not eight weeks.

If you have not given it twelve weeks, you do not have an opinion about it yet.

The three rules that matter

Sunscreen, non-negotiably, every morning. Retinoids thin the stratum corneum initially and photosensitivity is real.

Do not layer it with an exfoliating acid on the same night. There is no additive benefit and there is a well-documented additive irritation.

And buy the opaque, airless packaging. Retinol oxidises in light and air, and a clear jar is a product that will be inert before you finish it — which is, incidentally, the cheapest way for a brand to sell a retinoid that never irritates anybody.

How To Patch Test Properly, And Why Almost Nobody Does

Camille Roux — Mon, 22 Jun 2026 00:00:00 GMT

Every product with any potential to irritate carries an instruction to patch test, and the instruction is almost always too vague to follow. So people put a smear on the inside of a wrist, wait for the length of a kettle boil, see nothing, and use it on their face.

That tests for one thing only: immediate contact urticaria, which is rare. It does not test for the two problems people actually have.

The two things worth testing for

Irritant contact dermatitis is dose-dependent and cumulative. It is not an allergy; it is a chemical insult, and it appears after repeated exposure rather than after one. A retinoid, a high-percentage acid or a strong surfactant will pass a single ten-minute test and cause a problem on day nine.

Allergic contact dermatitis is immune-mediated and delayed. A first exposure sensitises silently; the reaction appears on a later exposure, typically twenty-four to seventy-two hours after contact. This is why fragrance allergies seem to arrive out of nowhere with a product someone has used happily for months.

Both of those need days, not minutes. Which is the real reason nobody patch tests properly — the useful protocol is tedious and the product is sitting right there.

The protocol

Choose somewhere thin-skinned and out of the way. The inner forearm near the elbow crease is the standard, and it is a reasonable proxy for facial skin. Behind the ear is better for anything you will use on the face; the skin there is closer in thickness and it is where a reaction shows early.

Apply a small amount, roughly what you would use on that area of your face, and leave it. Do not wash it off.

Repeat once a day, on the same spot, for five days. This is the part everyone skips and it is the part that catches irritant reactions, which need accumulation to show.

Look at it at twenty-four, forty-eight and seventy-two hours after the first application. A delayed allergic reaction usually appears in that window: not a general redness but a defined patch, itchy rather than stinging, sometimes with small bumps, and — the giveaway — sharply bounded at the edge of where you applied it.

What a pass and a fail look like

A pass is nothing, or a faint transient pinkness that has gone within an hour of each application.

A fail is any of: itching that persists, a defined bordered patch, small raised bumps, or a burning that gets worse with each successive day rather than better. Stinging on application that settles in two minutes and does not escalate over five days is usually an irritant threshold rather than an allergy — it means go slower, not never.

The exceptions

Anything containing a hair dye precursor — PPD and its relatives — genuinely needs the full forty-eight hours and genuinely can cause a serious reaction. This is the one patch test with a documented history of people being hospitalised for skipping it, and box dyes say so in small print for good reason.

Anything you have reacted to before, in any form. Cross-reactivity between related fragrance materials and between related preservatives is common enough that a previous reaction is a reason to test the new product for the full five days, not a reason to assume this one is different.

Everything else: five days, behind the ear, and a note on your phone of when you started. It is dull and it is a great deal less dull than three weeks of a face you cannot leave the house with.

Facial Mists Do Something. It Just Is Not Hydration

Camille Roux — Mon, 01 Jun 2026 00:00:00 GMT

The physics here are not complicated and they are not in dispute. Water applied to the skin surface evaporates. Evaporation is endothermic and it takes with it not only the water you applied but some of what was already in the stratum corneum. Applied repeatedly without an occlusive over the top, misting leaves skin measurably drier than it started.

Every dermatologist says this. Every brand selling a hydrating mist ignores it. And yet the category keeps growing, which suggests people are getting something from it — so it is worth asking what.

What a mist can genuinely do

Deliver a humectant, if it contains one and you seal it. A mist with glycerin, panthenol or a low-molecular-weight hyaluronic acid, applied and then covered with a moisturiser within thirty seconds, is a functional hydrating step. The mechanism is not the water; it is the humectant left behind and the occlusive over it. Without the second half, the humectant sits on a drying surface and can pull water out of the skin rather than into it.

Reduce skin temperature. A cool mist genuinely calms flushing, and for anyone with heat-triggered rosacea that is not a trivial benefit. It is thermal, not hydrating, and it works.

Reset powder. This is the one makeup artists have always known. A fine mist over a powdered base melts the powder slightly into the layers beneath and removes the flat, dusty quality of a heavily set face. It is a finishing technique and it is genuinely transformative on camera.

Make you stop. Not nothing. A mist is thirty seconds of pause with your eyes closed, and a great deal of what people describe as calming skincare is that.

What it cannot do

Hydrate on its own, in the sense of raising water content in the stratum corneum for any useful length of time. The film is gone in ninety seconds.

Set makeup, unless it contains a film former — most do not, and the ones that do say so.

Deliver actives meaningfully. The dwell time is too short and the dose is too small. A mist containing niacinamide is a mist, with a marketing claim attached.

How to use one properly

If it contains humectants: mist, then moisturise immediately, before it dries. That is the whole instruction and it converts a useless step into a useful one.

If it is thermal water or a plain botanical water: use it for temperature, not for hydration, and do not expect it to replace anything.

If you are using it over makeup: hold it further away than feels right — thirty centimetres, not fifteen — and go across the face rather than at it. Large droplets land as spots and disturb the base; a fine even fall is what melts powder.

The one worth owning

An unscented thermal water in a can, kept in the fridge, for hot days and for skin that flushes. It costs very little, the claims on it are modest, and it does exactly what it says.

Everything sold as a hydrating mist at three times the price is either that with glycerin in it, or that with a fragrance in it. The first is worth having if you seal it. The second is a room spray you are aiming at your face.

Your Bathroom Shelf Is Too Crowded To Tell You Anything

Camille Roux — Mon, 18 May 2026 00:00:00 GMT

The most common question we get is a version of the same one: something changed, I do not know what caused it, what should I do. And almost every time, the reason it cannot be answered is that four things changed at once.

Skincare is one of the few consumer categories where the product only works if you can attribute the result to it, and where nearly everyone makes attribution impossible on purpose.

Why the confound is so easy to create

New products arrive in clusters. A recommendation, a sale, an empty bottle and a restock, and suddenly the routine has three items in it that were not there a fortnight ago.

Skin also has a long lag. Cell turnover in adult skin runs roughly forty days, and most actives have a meaningful lead time — twelve weeks for a retinoid, six to eight for a vitamin C, four for a good moisturiser. A change introduced in week one shows up in week eight, by which time two more things have been added and the timeline is unrecoverable.

And skin varies on its own. Season, hormonal cycle, sleep, illness, water hardness after a house move. Any of those will produce a swing large enough to be mistaken for a product effect.

The method

Change one thing. Wait the appropriate length of time for that thing. Then change the next.

For a moisturiser or a cleanser: four weeks. For an active: twelve. Those are not conservative numbers, they are the intervals at which the published trials for these ingredients measured anything.

Photograph, in the same place, at the same time of day, without makeup. Human memory for a face is dreadful, and the improvement you are trying to detect is the size of the difference between two mornings. Photographs are the only reason my own experiments have ever produced a number I trusted.

Write down the date you started. Every single time. It costs four seconds and it is the difference between an experiment and an anecdote.

What to do with a shelf that is already crowded

Strip it back. Cleanser, moisturiser, sunscreen, for two weeks. This is not a detox — skin does not detox and nothing is being purged — it is a baseline, and it is the only way to find out what your skin is like without eight variables on it.

A great many people discover at this point that two of the eight were causing the problem they were buying the other six to fix.

Then reintroduce, one at a time, at the intervals above. It will take four months to rebuild a routine this way. You will end it knowing what each item does, which is more than almost anyone can say about the shelf they have now.

The uncomfortable part

This method makes it much harder to buy things, which is presumably why nobody selling anything recommends it.

A routine you can test is a routine that grows slowly, rejects most of what it tries, and ends up short. That is not an accident of the method. It is the finding.