Do hair repair products actually work? An honest answer

Repair products do something real and they do not do what the word suggests. The best of them make damaged hair smoother, easier to comb, shinier and less likely to break in your hands. None of them heal a strand, because there is nothing alive in a strand to heal.

That distinction sounds like a technicality. It decides how you spend your money, so it is worth five minutes.

Why repair cannot mean what it sounds like

StatPearls describes the hair shaft above the skin as a cylinder of non-living keratinised cells. No blood supply, no cells doing maintenance, no immune response. A 2023 review of hair care chemistry in Polymers puts the consequence in one line: hair is made of dead cells, so it cannot be biologically renovated, and the only way to restore its properties is to repair it physically with a conditioner.

Compare that with skin. Cut your finger and living cells close the wound. Bleach your hair and the fibre stays exactly as bleached as it was, for as long as that length stays on your head.

So every honest claim on a repair bottle is about one of three things: adding something to the surface, adding something into the worn spaces, or changing how the fibre behaves for a while. Those are all worth having. None of them is healing.

What a conditioner actually does, measured

This is the part with the strongest evidence behind it, and it is the least exciting part of the shelf.

A 2015 review of shampoos and conditioners in the Indian Journal of Dermatology sets out the mechanism: hair carries a negative charge, conditioners carry a positive one, so the conditioner is drawn onto the hair and deposits most where the hair is worn. The same review says damaged hair is more negatively charged than healthy hair, which is a neat piece of chemistry. The product goes where it is needed without anyone aiming it.

Once deposited it does three measurable things. It flattens the cuticle scales. It reduces the friction between hair fibres. It increases how much light the surface reflects, which the review links directly to shine and colour.

That is the whole effect, and it is enough to explain most of what people notice after a good mask. Less snagging. Easier combing. A visible change in shine at the window. All of it sits on the outside and all of it washes off, which is why the routine matters more than any single treatment. The guide to deep conditioning and hair masks covers how often to use one and how long to leave it.

Protein: the size decides where it goes

Protein is the ingredient most often behind the word repair, and there is real work on it.

A 2020 study in the International Journal of Cosmetic Science tested hydrolysed keratins of three different sizes on textured hair. The smallest pieces, around 221 daltons, went deep into the cortex and did nothing useful for stability. The middle size, around 2,577 daltons, went deep into the cortex as well. The largest, around 75,440 daltons, mostly sat on the outside. Both the middle and large sizes made the hair stiffer and reduced breakage at two different humidity levels.

Two honest notes about that study. It was supported by an ingredient supplier and a hair testing institute, which does not make it wrong and does mean the people funding it sell the ingredient. And it measured hair fibres in a laboratory, not hair on a head going through a week of real life.

What it tells you as a buyer is simple. Protein in a bottle is not one thing. A product with small protein fragments and a product with large ones behave differently, and no label tells you which you are holding.

Where oils and silicones fit, and why neither is repair

Oil is the oldest hair treatment there is, and it belongs in a different box from repair.

The 2015 Indian Journal of Dermatology review notes that silicones such as dimethicone are superior to natural oils for film formation, shine and lustre. That is a direct comparison, and it explains why so many products people describe as smoothing contain one.

Both work the same way. They lay a thin film along the fibre, which smooths the lifted scales, cuts the friction between hairs and raises how much light bounces off. Hair that has been oiled combs more easily and looks better in a photograph on the same day.

What neither does is enter the worn structure and hold it. An oil sits on the outside by design, and that is its whole value. Judge an oil on how the hair behaves after it, not on a promise about the inside of the strand.

One practical note. Oil applied to the ends before a wash reduces how much water the fibre takes on, which matters because the 2023 Polymers review links water swelling to damaged hair becoming more fragile. That is a prevention step, not a repair step, and prevention is where oil earns its place.

Bond building: what has been shown, and what has not

The bond building category claims something different from a conditioner: work inside the fibre, re-forming the chemical links that bleach and chemical services break.

Here is what I found when I went looking for evidence that does not come from a manufacturer.

A 2024 study in the journal Cosmetics, by researchers at a Portuguese university, bleached hair samples to a set protocol and then tested two commercial repair products on them. Both raised the brightness and lightness of the bleached hair. One of them, a water-based product, produced what the authors describe as a visible inner sheet filling empty spaces, with an increased number of disulfide bonds measured afterwards. The other, with oil-based ingredients, formed a layer on the outside of the fibre instead. Two products, two different mechanisms, and the difference was measurable.

A 2025 paper in Pharmaceuticals took a separate approach and tested three experimental thiol cross-linking compounds on bleached hair strands. One of them restored tensile strength and elasticity better than the others. The authors declare no funding and no competing financial interests, which makes this the most independent evidence in the pile. It also tested new compounds rather than anything you can buy.

So the honest position. There is laboratory evidence that some products get inside the fibre and change what can be measured there. There is no long-term evidence I could find, from a source with no stake in the answer, showing how long that holds on a real head through months of washing. Where a brand's own studies are the only support for a claim, treat the claim as the manufacturer's until somebody independent repeats it.

Bond building is the first repair claim in decades that has a plausible mechanism and a measurable lab result behind it. That is genuinely more than the category used to have, and it is still not the same as proof about the bottle in your hand. The Kurli team

Split ends: held together, not mended

The clearest example of the gap between the word and the effect.

The 2015 Indian Journal of Dermatology review states that protein-derived substances in conditioners can temporarily mend split ends, which dermatology calls trichoptilosis, and that conditioners temporarily seal them. Both sentences carry the word temporarily.

What happens physically is that the product glues the two halves of the split together and smooths the outside. The hair looks whole and behaves better. Then you wash it and the split is a split again, because nothing joined it. Meanwhile the split keeps working its way up the strand.

That makes a sealing product useful for an event and a poor plan for a year. The permanent removal is scissors, and the guide to split ends and breakage explains how much to take and how often. If the question is how much length you have to give up, we worked through that in should I cut off damaged hair.

What a repair label should mean to you

Turn the word into a question before you pay.

Ask what is being claimed. "Repairs damage" is not a claim anybody can check. "Reduces breakage during combing" is, because it names a thing that can be counted. Prefer labels that describe an effect over labels that describe a state.

Ask where it works. Surface products, the coating and smoothing kind, give you the visible improvement immediately and rinse out. Internal products claim a slower, longer-lasting change. You need the first kind whatever else you use.

Ask how long it stays. A rinse-out conditioner does its work in the shower. A leave-in stays until the next wash. A treatment that claims months of effect is making the biggest claim in the room and should carry the most evidence.

Ask what it is for. Protein and moisture are separate problems with separate signs, and using the wrong one makes hair worse rather than better. That question is worth its own read, and the deep conditioning guide covers how to tell them apart.

Ignore the before-and-after photographs. Wet hair, a blow dry under tension and a different light can produce that difference on any head, damaged or not.

What no product does

Worth saying plainly, because the marketing is loud.

No product rejoins a hair that has broken. No product reverses a split permanently. No product makes bleached hair what it was before it was bleached. No product changes the growth rate at the follicle, which sits at about 0.35 mm a day for most people, according to StatPearls. And no product protects hair from a habit that is still happening.

That last one is the one that costs people the most. The American Academy of Dermatology's list of ten habits that damage hair includes daily heat styling, rubbing shampoo down the length, brushing wet, tight ponytails and stacking chemical services. A treatment applied on Sunday does not survive a flat iron on Monday.

How to test one on your own hair in four weeks

Buy one product, not three. Changing several things at once tells you nothing about any of them.

Use it as directed for four weeks, and change nothing else. Same shampoo, same heat, same brush.

Judge it on three things you can check. Does the comb meet less resistance on wet hair. Is there less hair in the brush. Does the surface reflect a band of light at a window that it did not reflect before. Those are the effects a good repair product produces, and they are all visible without a laboratory.

If you want the reading done on your own hair rather than on a category, the Kurli AI Specialist works from three photographs of your hair, says what it sees in the ends, the texture and the scalp, and describes what to buy by what the label should say rather than by brand. The signs to look for first are in how do I know if my hair is damaged.

And the honest last word. The most effective repair in hair care costs nothing and nobody sells it: stop the thing that is causing the damage, then let the length you keep be the length that grew after you stopped.