Minoxidil for beard growth: what it is and what it is not

Minoxidil did not begin as a hair product. It was developed as an oral medicine for high blood pressure, working as a vasodilator — a compound that relaxes the smooth muscle in blood vessel walls so that vessels widen and pressure falls. That is what it was for, that is the pharmacology it was characterised against, and that is still what the oral forms are fundamentally doing in the body.

The hair connection arrived as a side effect. Patients taking it systemically were observed to grow hair in places they had not previously grown it, a phenomenon clinicians call hypertrichosis. That observation was unexpected, reproducible, and commercially interesting, and it led eventually to topical formulations developed specifically for pattern hair loss of the scalp. The drug found its way to hair through an accident of observation rather than through a theory about follicles, which is worth remembering when someone describes its mechanism as though it were designed for the purpose.

So the compound has two distinct lives. In one it is a systemic cardiovascular drug. In the other it is a topical applied to skin for an approved hair indication on the scalp. Beard use is a third thing, and it is neither of those.

"Off-label" gets used loosely, usually to mean either "dodgy" or "fine, don't worry about it", and it means neither. A drug is approved by a regulator for a specific indication, in a specific form, on the basis of a specific evidence package. The label is the document that results — the summary of what the medicine is for, who it is for, what is known about its risks, and what the manufacturer is permitted to claim.

Using an approved drug for a different purpose than the label describes is off-label use. It is legal and commonplace in clinical practice; doctors prescribe off-label across medicine, often for good reasons, and it is frequently how evidence for a new indication starts accumulating in the first place. What it specifically does not mean is that the use has been evaluated. Facial hair is not an approved indication for minoxidil in any jurisdiction we are aware of. The trials that produced the label did not study it, the manufacturer does not claim it, and the safety information attached to the product was written about a different application site and a different population.

The practical consequence is that the usual chain of accountability is thinner. When someone applies a scalp product to their face on the strength of forum photographs, the questions that would normally have been answered by a regulatory review — who should not use this, what happens with this application site, what the interaction profile looks like — have simply not been asked in that context. That is not a reason to panic. It is a reason to route the decision through someone qualified rather than through a comment thread.

Here is where most explanations quietly overreach. The honest position is that minoxidil's effect on hair follicles is understood in outline and not in full, and the research literature says so openly. What can be stated with reasonable confidence: it is a potassium channel opener with vasodilatory activity, and its influence on the follicle appears to involve lengthening the active growth phase of the hair cycle and pushing finer, weaker hairs toward a thicker and more pigmented state.

To make sense of that you need the cycle. A hair follicle is not a continuously producing factory. It runs a repeating sequence: an active growing phase, a short transitional phase in which production shuts down, and a resting phase after which the old hair is released and a new cycle begins. Crucially, human follicles cycle asynchronously — your neighbours in the mirror are all at different points at once, which is why humans do not moult. The length of the growing phase is the main thing that determines how long a hair can get before it is shed, and it differs by body region. That is why the hair on a scalp reaches lengths the hair on an arm never does.

If a compound extends the growing phase, more follicles are in production at any given moment and each hair has longer to develop before it goes. That is the general shape of the claim. The vascular story — more blood flow, more nutrients — is the version that spreads most easily because it is intuitive, and it is probably too simple; the fact that the drug's activity on hair does not track neatly with its blood-pressure activity is one of the reasons researchers stopped treating vasodilation as a complete explanation.

One of the most useful things to understand about facial hair is that it responds to androgens in the opposite direction from scalp hair in pattern loss. Beard growth develops under androgen exposure: hairs in the beard region convert from fine, pale vellus hair into the coarse, pigmented terminal hair that makes a beard visible. In pattern hair loss of the scalp, genetically susceptible follicles do the reverse under the same hormonal signalling, miniaturising over successive cycles until the hair no longer covers.

Same hormone family, opposite outcome, depending on the follicle's local programming. This is not a paradox so much as a reminder that follicles are regionally specified: they carry positional identity and behave according to it. It also explains a pattern people find confusing — that the same person can be losing scalp hair and growing a heavier beard over the same years.

Why does this matter for a minoxidil discussion? Because it means beard density is substantially a matter of how many follicles in that region are programmed to make terminal hair and how they are responding, and no topical rewrites that programme. A compound that acts on the cycle of follicles that exist is not a compound that creates follicles that do not. Anyone promising a beard where the underlying follicle distribution does not support one is selling something. The surgical side of hair has the same hard constraint, which we go through in the piece on FUE versus FUT hair transplants: donor supply is finite, and no technique conjures follicles.

There is a specific and interesting piece of pharmacology here that explains a lot of the variance people report. Minoxidil as applied is understood to be a prodrug — it needs converting into an active sulfated form, and that conversion is carried out by sulfotransferase enzymes present in the skin and follicle. Enzyme activity varies between individuals. Two people applying exactly the same thing to exactly the same area are not necessarily generating the same amount of active compound where it matters.

This is one of the cleanest explanations available for why anecdote is such a poor guide in this space. When a person reports a strong response and another reports nothing, the natural instinct is to look for a behavioural difference — they were more consistent, they used a different product, they did something extra. Sometimes that is true. But a genuine biochemical difference in how much active drug each person's skin produces is sitting underneath, and it is not something you can see, guess or control from the outside.

The lesson generalises past this one compound. Wherever an intervention's effect depends on an individual biological variable that nobody measures, the testimonials will be spectacularly heterogeneous and the loudest ones will be from the strongest responders, because those are the people motivated to post. That is a selection effect, not evidence.

This point gets flattened constantly by internet advice written from one country's perspective. Minoxidil is a medicine, and how you are permitted to obtain it depends entirely on where you are. In some jurisdictions topical formulations for the approved scalp indication are sold over the counter. In others the same product is pharmacy-only, requiring a conversation with a pharmacist. In others again it requires a prescription. Oral forms are prescription medicines.

Someone in a country where the topical is on an open shelf will describe getting hold of it as trivial, and they are not lying; they are describing their regulatory environment and mistaking it for the world. Someone else reading that then orders from an overseas seller, which introduces a separate set of problems: no pharmacist, no label in a language they read, no assurance about what is in the bottle, and no regulated supply chain behind it.

We are not going to tell you what applies where you live, because it changes and because getting it wrong would be worse than saying nothing. The point is that "you can just buy it" is a claim about a jurisdiction, not a fact about the substance, and that the regulatory hurdle — where one exists — is there because the thing on the other side of it is a drug.

A product formulated for the scalp was formulated for scalp skin. Facial skin in the beard region is different tissue in a different place: thinner in areas, more exposed, more frequently shaved or abraded, and immediately adjacent to the eyes, the lips and the mucous membranes of the nose and mouth. None of those neighbours are relevant when you apply something to the top of a head.

There is also the vehicle problem. The active compound has to be dissolved in something to be delivered, and the solvent systems used in topical hair products are themselves a recognised source of skin irritation and contact reactions. When people report irritation, it is not always the drug doing it — sometimes it is the carrier, which is why formulation differences produce different tolerability even at matched drug content. That is a genuinely useful thing to know before you conclude that a reaction tells you something about the active ingredient.

And because minoxidil is systemically a cardiovascular drug, the question of how much crosses the skin and reaches circulation is a real one rather than a theoretical one. Absorption depends on the site, the skin's condition, the area covered and the formulation. That is precisely the sort of question a doctor is equipped to weigh against your medical history and precisely the sort of question a website should refuse to answer generically. We are refusing.

Three questions dominate every discussion of this compound, and all three are clinical. We will describe what they are, because understanding the shape of a question makes a consultation more productive, and then we will stop.

Shedding. An increase in hair shedding after starting is a widely reported phenomenon, and the usual explanation offered is mechanistic rather than alarming: if a compound acts on the growth cycle, follicles sitting in the resting phase may be pushed into releasing their old hair as a new cycle begins, and a normally invisible background process becomes briefly visible and synchronised. That is the standard account. Whether shedding in a specific person reflects that, or reflects something else entirely, is not something a stranger can determine from a description, and it is exactly the moment to talk to a doctor rather than to a forum.

Continuation. The general understanding of drugs that act on the hair cycle is that they act while they are acting — the effect is maintained by continued use rather than banked. That framing turns a cosmetic decision into an ongoing commitment, and the honest version of the question is not "does it work" but "am I prepared for this to be indefinite, and what does indefinite mean for cost, tolerability and monitoring".

Stopping. What happens on discontinuation follows from the same logic, and it is one of the most consequential things to understand before starting anything, not after. Ask about it first.

We are not giving a strength, a frequency, an application method or a timeframe for any of this, and you should be suspicious of any source that does so without knowing anything about you. The same caution applies to the drugs used for scalp pattern loss, which we treat the same way in the transplant article: mechanism explained, decision routed to a clinician.

It is worth being blunt about our own limits, because this topic invites an assumption we cannot meet. RealSmile measures facial geometry from a photograph. From 68 facial landmarks the scan produces two parallel outputs: a set of 17 geometry metrics, which is a proportion map rather than a ranking, and — separately, from a different model — a validated Face Score percentile. Those are two outputs of one scan, not one feeding the other.

That distinction is not pedantry, it is a result. We tested a blended composite of the geometry metrics against averaged human ratings and it failed — no positive correlation, published on our research page rather than quietly buried. The percentile that we do report comes from a separate model built on a different representation of the face, cross-validated against human ratings. So the 17 metrics are measurements you can read, and the percentile is a ranking; they sit side by side and we do not pretend one produces the other.

Now the part that matters for this article. The standard 68-landmark set maps the jaw outline, the brows, the eyes, the nose and the mouth. It contains no hair points at all. We do not measure hair density. We do not count follicles. We cannot see the hairline, the scalp, or the condition of skin. We cannot tell you whether a beard is filling in, thinning, patchy or improving, and we would be inventing it if we tried.

There is a subtler limitation too, and it is specific to bearded faces. When a landmark detector traces a jaw outline on a face with a full beard, part of what it is tracing is the silhouette of the hair rather than the bone beneath it. That means a beard can move the geometry read without anything about the underlying face having changed — which is worth knowing if you run the jawline test before and after a grooming change and wonder why the numbers moved.

Almost the entire popular case for using anything on a beard rests on paired photographs. This is the part of the topic we can speak to directly, because it is a measurement problem rather than a medical one, and it is our actual field.

Start from the right premise: what a photograph records is not the amount of hair on a face. It is the amount of hair a particular camera resolved, under a particular light, at a particular growth stage, at a particular angle, against a particular skin tone, after a particular grooming history, and then processed by a pipeline that had opinions about all of it. Hair is close to the worst possible photographic subject for honest comparison — it is fine, it is high-frequency detail near the limit of what a sensor resolves, it is directional, it casts and receives shadow, and its apparent quantity is dominated by how light strikes it.

Compare that to tooth colour, which we pulled apart in the whitening article. Tooth shade in a photograph is already unreliable enough that dentists refuse to judge it from images and use physical reference tabs under controlled light instead. Beard density is worse, because it adds growth stage and grooming to the same list of optical problems that colour already had. The next four sections take the confounds one at a time.

This is the single largest problem, and it is almost never controlled. A face photographed the day after a close shave and the same face photographed after a stretch of uninterrupted growth are not two data points about density. They are one face at two points in a growth process, and the difference between them is enormous regardless of whether anything else changed.

It compounds. Longer hairs lie over gaps. A hair that has had time to grow drapes across the skin next to it and covers the sparse patch behind it, so apparent coverage increases faster than actual hair count does. Length also changes how the hair takes light: a very short hair is essentially a dark dot at the skin surface, while a longer one is a line that catches a highlight along its length. And colour reads differently at different lengths, because a stubble field is mostly skin with dark points in it while a grown beard is mostly hair.

Because the follicles are cycling asynchronously and independently, there is also genuine natural variation in what is present at any moment, quite apart from length. Two honestly-taken photographs a few weeks apart with nothing done in between will not be identical. If your comparison is not anchored to a defined growth point — the same number of days since the same kind of shave, at minimum — the growth-stage difference will swamp everything you were trying to measure.

Fine hair is visible in a photograph mainly because of two effects: it interrupts the light falling on the skin behind it, and it catches specular highlights along its own length. Both are properties of the lighting, not of the hair.

Light raking across a face at a shallow angle throws every short hair into relief and casts a tiny shadow beside each one, which reads as texture and density. The same face under flat, frontal, diffuse light — a soft overcast window, a large softbox straight on — loses most of that micro-shadowing, and sparse growth can nearly disappear. A hard light behind or to the side produces rim lighting along the hairs themselves and makes a thin beard look substantial. A phone flash fired straight at a face flattens everything and tends to blow out skin, which reduces the contrast between skin and hair and makes coverage look worse than it is.

So the direction and quality of light alone can take the same face from "patchy" to "full" and back, with no biology involved. Anyone assembling a persuasive pair — deliberately or not — is choosing flat light for the before and directional light for the after, and the result will look like a transformation. We wrote the general version of this argument in the before-and-after teardown for whitening, and the same logic drives the capture guidance behind the 17-metric scan: control the light first, then look at the subject.

A beard is a styled object, and styling is a large, cheap, instantaneous lever on how it reads. Cleaning up the cheek line and the neck line produces a defined edge, and a defined edge reads as density because the eye interprets a crisp boundary as intentional rather than sparse. Combing hair across a thin area covers it. Combing it down instead of out changes the outline entirely. Trimming to an even length removes the stray long hairs that make an uneven beard look uneven.

Then there is the whole category of products whose entire purpose is to darken, thicken or fill: tinting, fibres, pencils and shading applied at the skin. These exist, they work photographically, and their presence is essentially never declared in a comparison. A pair of photographs cannot distinguish between more hair and better presented hair, and a viewer has no way to know which they are looking at.

This is the same effect we describe for brows in the eyebrow shaping guide: shaping a boundary changes perceived density far more than changing the number of hairs does, and it changes it instantly. If your before photo is ungroomed and your after photo is groomed, you have measured grooming.

Camera geometry changes what a beard covers in the frame. A camera held low looks up along the underside of the jaw and sees the neck and jawline hair at a grazing angle, where hairs overlap each other and coverage looks maximal. A camera held above eye level looks down onto the cheeks, sees the skin between hairs more directly, and reveals every gap. Neither is more truthful; they are different projections.

Distance matters through perspective. Holding a phone close for a selfie enlarges whatever is nearest the lens and compresses the rest, which changes the apparent proportion of the lower face — and therefore how much beard area there appears to be relative to everything else. Step back and use the same framing from a distance and the relationships change. Switching between the front and rear camera is worse again, because those are different sensors with different resolutions and different processing.

There is also a specific effect worth naming: a beard produces shadow under the jaw and around the neck, and that shadow deepens the apparent boundary of the lower face. It can make a jaw look sharper in a photograph without any change to the bone. If you are running geometric measurements, that is a real interaction — the landmark detector is working from a silhouette, and the silhouette has hair in it. It is the reason we would rather people scan clean-shaven or hold the beard state constant across comparisons, and the reason the jawline measurement should never be read as a claim about bone.

We can put a number on the general problem, from our own measurements, and it is a sobering one. We looked at pairs of scans from the same person taken on different calendar days less than fourteen days apart — n = 103 pairs, close enough in time that no meaningful change to the underlying face is plausible. The median difference between the two scores in a pair was 4 points. Around 40% of the pairs moved by 5 points or more.

Those faces did not change in under two weeks. What changed was the photograph: the light, the camera height, the angle, the expression, the distance. That is enough to shift where landmarks land and therefore what the geometry reports. We published it because it constrains what we are allowed to claim about our own product — a score from one photograph is a reading of that photograph, not a fixed property of a person.

Now transfer the lesson. Geometry is the most stable thing you can extract from a face. Bone does not move. If a geometric reading drifts that much between two casually taken photographs, then a judgement about fine hair — which additionally depends on light direction, growth stage, grooming and the resolving power of the sensor, none of which affect geometry — is far less stable again. An informal beard before-and-after pair is not a weak measurement. It is not a measurement.

You cannot control everything with a phone, but you can hold enough still that the remaining difference means something. This protocol is useful whatever you are or are not doing, because it is the only way to find out for yourself instead of relying on somebody else's photographs.

1. Anchor the growth stage. This is the one most people skip and the one that matters most. Shoot at a fixed, defined point after the same kind of shave or trim, and record what that point was. Comparing different growth stages tells you nothing.
2. Same grooming state, or none. Either both photographs are freshly groomed to the same lines, or neither is. No tinting, fibres, pencils or shading on either side. Declare it to yourself in writing if you have to.
3. One room, one light, one time of day. Kill every other light source. Mixed colour temperatures destabilise the camera's white balance and change how skin and hair separate tonally.
4. Fix the light direction deliberately. Pick one — flat frontal or a consistent side angle — and reproduce it exactly. Mark where the lamp or window is. Direction is the biggest single lever on apparent density.
5. Lock exposure and white balance. Auto anything re-decides between shots. On most phones a long press locks focus and exposure; a manual camera app is better.
6. Same distance, same camera height, same lens. Mark the floor. Use the rear camera both times. Do not switch between front and rear.
7. Shoot a set of angles, not one. Straight on, both three-quarters, and one from slightly above looking down at the cheeks. A single flattering angle is how people fool themselves.
8. Neutral expression, mouth closed, jaw relaxed. Clenching, smiling and jutting the chin all change the jaw silhouette and the tension of the skin the hair sits in.
9. No filters, no beauty mode, no portrait mode. Skin smoothing actively removes fine detail — which is to say, it removes exactly the hair you are trying to count. Export the originals.
10. Compare on one screen, side by side, same session. Screen brightness auto-adjusts, and comparing an image on a phone today against a memory of one last month is not comparing anything.

Ten steps sounds like a lot; it takes about five minutes to set up once and thirty seconds to repeat. The alternative is a folder of images taken under different conditions that will confirm whatever you already believed, which is worse than having no images at all — because it feels like knowing.

Even if every individual photograph were honest, the collection you encounter would still mislead you, because of who posts and when. People post when something worked. Someone who tried a thing and saw nothing rarely assembles a photo set documenting the absence of a result, and if they do it gets no engagement and sinks. What survives in a feed is the upper tail.

Add the timing effect. Posts cluster at the moment someone is most enthusiastic, which is not necessarily the moment the outcome is most representative. And add the commercial layer: some fraction of what looks like personal experience is marketing, sometimes disclosed and sometimes not, and photographs in that category are produced by people who understand lighting professionally.

So you are looking at a filtered sample, of an unreliable instrument, measuring a variable with genuine person-to-person biological variance underneath it. Three independent problems stacked. This is why we keep pushing people toward their own controlled baseline rather than toward better forums: your own well-taken pair, whatever it shows, is worth more than a thousand of somebody else's.

We cannot tell you what to do and we are not qualified to. What we can do is describe the shape of the conversation so it is less opaque when you have it. A clinician looking at facial hair concerns is generally interested in the cause before the cosmetics: whether the pattern is simply constitutional, whether anything is actively causing hair loss in the area, whether there is a skin condition present, what your medical history and current medications are, and whether the thing you want is achievable at all given the follicles you have.

Questions that tend to produce a clear answer:

• Given my history, is this compound appropriate for me at all, and is there any reason it would not be?
• What is its regulatory status here, and what does that mean for how I would obtain it and be monitored?
• What is known about applying it to facial skin specifically, as opposed to the approved site?
• What side effects are recognised, how common are they, and which ones would mean stopping immediately?
• If shedding increases after starting, what is the expected explanation and when should I come back?
• What happens if I stop, and is this a commitment I should treat as ongoing?
• How should the outcome be assessed, and can we record a controlled baseline rather than relying on memory?
• Is there anything about my situation that suggests the underlying pattern will not respond regardless?

That second-to-last question is the one people forget, and it is the one this whole article has been building toward. If nobody records a baseline under controlled conditions, then in a few weeks the only comparison available is between a new photograph and your recollection of your old face, and recollection of your own appearance is unreliable in a specific direction: toward whatever you were hoping for.

Beard density is one input into how a face reads, and it is neither the largest nor the one most under your control. Grooming — the cheek line, the neck line, the length, the shape relative to the jaw — is instant, free, reversible and carries no pharmacology at all. It is strange to reach for a drug before exhausting the lever that requires nothing but a mirror and ten minutes. The same ordering holds for skin, which is why the skincare routine guide starts with the two or three things that actually matter rather than with a shelf of products.

There is also a constraint worth accepting rather than fighting. Facial hair distribution varies enormously between people and between populations, it is substantially heritable, and it changes over a lifetime on its own schedule. A pattern that is simply how your face is built is not a defect awaiting correction, and no topical rearranges the underlying distribution of follicles.

If you take one thing from this: the reason this topic is so full of confident nonsense is that its evidence base is photographs, and photographs of hair are the least trustworthy images in the entire appearance space. Control your capture conditions, keep a real baseline, judge changes against a fixed reference rather than against memory, and route the drug questions to a doctor. Do that and you will actually know something. Skip it and you will have a strong opinion and a camera roll.