Collagen is the most commercially successful ingestible beauty product in existence and the one with the strangest evidence base. The strange part is not that it is all nonsense. It is that the evidence is better than the sceptics say and much weaker than the tubs imply.
Here is what is defensible, what is not, and where the money goes if you decide to try it.
The objection you have probably heard
You cannot eat collagen and have it become skin collagen. Your digestive system breaks protein into amino acids and short peptides, and those get used wherever your body wants them. Swallowing collagen does not deliver it to your face any more than eating a steak builds bicep specifically.
That objection is mechanically correct and it is not the whole picture. The current hypothesis is not that ingested collagen becomes skin collagen. It is that specific short peptides survive digestion, circulate, and act as a signal that prompts fibroblasts to produce more collagen themselves. That is a plausible mechanism and there is laboratory work behind it.
What the trials show
There is a reasonable body of randomised trials on hydrolysed collagen peptides and skin, and a lot of them report improvements in hydration and elasticity over eight to twelve weeks. Meta-analyses pooling them have generally come out mildly positive.
Then you look at who paid for them. A large share of this literature is funded by collagen manufacturers, and the effect sizes reported are small, measured with instruments rather than visible in a mirror, and often assessed on parameters chosen after the fact. That is not fraud, and it is exactly the pattern that tends to shrink when independent groups run bigger trials.
My reading is that there is probably a real small effect on skin hydration, that the elasticity findings are shakier, and that anyone promising you visible change in four weeks is selling rather than reporting.
Dose is where products fail
Trials showing anything generally use between two and a half and ten grams of hydrolysed collagen peptides daily, with most sitting around two and a half to five grams of specific bioactive peptides, taken for at least eight weeks.
So check two things on any tub. First, that it is hydrolysed peptides rather than gelatin or unhydrolysed collagen. Second, the grams per serving. A beauty blend containing four hundred milligrams of collagen alongside eleven other ingredients and a picture of a woman touching her cheekbone is not delivering a studied dose, and no amount of the other eleven ingredients compensates.
This is why plain unflavoured collagen peptides in a large tub usually outperform the prettier products. They contain one thing, at a real dose, at a fraction of the price.
Marine versus bovine
Marine collagen is mostly type one, which is the dominant type in skin, and it is absorbed slightly faster in some studies. Bovine gives you type one and three. Marine costs more.
The difference is not large enough to matter for most people, and I would let price, dietary preference and how the powder tastes decide it. Marine collagen has a distinct flavour that some people cannot get past in water.
What I would tell a friend
- If you want to try it, buy plain hydrolysed collagen peptides, five to ten grams a day, and give it twelve weeks. Take a photo at the start in the same light, because your memory of your own skin is unreliable.
- Do not buy a beauty blend with a proprietary formula and no per-ingredient doses. That phrasing exists to stop you doing the arithmetic.
- Do not expect anything dramatic. The honest ceiling here is slightly better hydration, which is also what a decent moisturiser and drinking more water achieve for less money.
- If your budget only stretches to one intervention, sunscreen has better evidence for skin than every ingestible on this site combined, and it is not a close comparison.
