What Marinades Actually Do (and Do Not)
A marinade penetrates one to three millimetres. Acid does not tenderise the way it is said to, and the only thing that really travels inward is salt.
A marinade stays on the surface
Cut into meat marinated overnight and the colour reaches one to three millimetres in. Most aroma molecules are large and fat-soluble, and they do not travel between muscle fibres.
So "I marinated it for 24 hours, so it is seasoned through" is not true. What long marinating mostly buys is a stronger-flavoured surface, and past a point the extra time returns nothing.
There is one exception, and it is the point of this guide.
Salt is the thing that gets in
Salt is a small ion. It dissolves and diffuses into the muscle β and once inside it does real work, loosening protein structure so the meat holds on to water and loses less of it in cooking (salt and seasoning).
Which means a good marinade is effectively a brine with surface flavour on top. If you keep one ingredient, keep the salt; a marinade without it is little more than aroma painted on the outside.
It is why the sauce on bulgogi or jeyuk bokkeum works in thirty minutes: soy sauce is doing the salt's job.
Which means some marinades are marinades in name only
Southern Fried Chicken is known for its overnight buttermilk soak, and the thing actually doing the work is the salt in it. Leave the salt out and buttermilk alone does almost nothing; salted water alone does most of the job. The buttermilk is there for other reasons β its acid and milk protein, below.
German Eisbein takes the logic to the end: three to five days in brine, and then boiled. Boiling water leaches the soluble flavour out of meat, so putting the seasoning inside before that happens is the whole technique of the dish.
It does not have to be a liquid
What salt needs in order to travel is water, and the meat already has it. So a dry rub is a brine too. Texas Smoked Brisket is salt and pepper left on overnight, and Suya uses a powdered rub with no liquid in it at all. For the first half hour the salt pulls water out of the surface; after that it is reabsorbed along with the salt (salt and seasoning).
Acid does not tenderise
This is the most widespread misconception. Vinegar, lemon and wine denature surface proteins, which can read as tender, but they do not reach the interior β and given time they:
- leave the surface dry and chalky;
- drive out moisture, so the meat browns poorly;
- firm up white fish and prawns as though cooked, within tens of minutes. Ceviche is built on exactly that, and it is cooking, not tenderising.
Acid has its own proper place β dishes like vindaloo, where vinegar is the preservative and the flavour axis. Adding acid for tenderness is aiming at the wrong thing.
So what are the four days for?
German Sauerbraten marinates beef in vinegar and wine for three to four days, and the folk explanation is exactly the misconception above: the vinegar tenderises it. What actually tenderises it is the three hours of gentle moist heat afterwards.
Which leaves two real answers for the four days.
- The salt genuinely gets in, exactly as in the section above.
- The marinade becomes the sauce. A recipe that throws the liquid away is not this dish; it is strained and reduced and poured over. So the four days are not time spent doing something to the meat, they are time spent moving the meat's flavour into the liquid. Historically the vinegar was there to preserve, not to soften.
That is not in conflict with the rule at the end of this guide. Boiling it is the condition, and sauerbraten meets it.
Enzymes really do break protein down, but they are hard to stop
Pineapple (bromelain), kiwi (actinidin), pear and fig carry enzymes that genuinely cut protein. They work β the trouble is that they do not stop.
- Thirty minutes is enough; a few hours turns the surface to mush.
- They act unevenly, so parts stay intact while parts collapse.
- Grated pear in Korean galbijjim is this mechanism in use, which is why the quantity is small and the time short.
Canned pineapple has been heat-treated and its enzymes are dead, so it is useless for this.
Yoghurt works differently
Tandoori chicken and Indian marinades use yoghurt for more than its acidity. Yoghurt combines mild acid, calcium and milk protein, so it acts far more gently than vinegar, and its thickness keeps it clinging to the surface. Its lactose and protein also feed the Maillard reaction, so it colours well (browning and maillard).
A yoghurt marinade can therefore sit overnight without wrecking the surface. Do not put it on the same timetable as a vinegar one.
In buttermilk, the acid's real job is the crust
The buttermilk in Fried Chicken is essentially thin yoghurt and survives overnight for the reasons above. But in that dish the acid's real work is not on the meat at all β it is on the coating.
Buttermilk is thick and mildly acidic, so worked into flour it does not disperse smoothly; it clumps. Those lumps fry as lumps, and that is the broken craggy crust. It is why step three of that recipe deliberately drizzles leftover brine into the dredge: a smooth batter film cannot produce it.
So a marinade has one more job beyond the oil and sugar above β being an ingredient of the coating that comes next.
What oil and sugar are for
- Oil distributes fat-soluble aromas β chilli, spices, herbs β evenly across the surface and carries heat in the pan. It has nothing to do with penetration.
- Sugar colours fast but burns. A sweet marinade over high heat scorches before the inside is done. Something like jerk chicken wants medium heat and time, or the sugar brushed on at the end.
Times, and two rules
Set the time by thickness, since the action is at the surface.
- Thin-sliced meat, prawns: 15-30 minutes
- Chicken pieces, 2cm-thick cuts: 2-4 hours
- Whole legs, thick joints: overnight, if it is salt-based
- Strongly acidic marinades: under 2 hours, always
- Enzymatic (pear, kiwi, pineapple): 30 minutes
- Salt-based, and going on to a long cook: days β Eisbein three to five, Sauerbraten three to four. The surface does suffer, and it does not matter, because both then spend hours boiling or braising. Never use this row for something you intend to grill quickly.
And two things without exception:
- Pat the surface dry before it hits the pan. A wet surface steams instead of browning.
- Do not serve the marinade raw as a sauce. Boil it first if you want to use it.
What "a wet surface" is, and a coating is not one
Both of those hold. The first is about free liquid β a film of water sitting on the meat, which has to boil off before the surface can pass 100Β°C, and until it does nothing browns (browning). What it is not about is the marinade staying on, and this guide already says as much a few paragraphs above: yoghurt's "thickness keeps it clinging to the surface." Those two sentences have never been put side by side.
Kebab puts them side by side, and its wording is the instruction: scrape the excess yoghurt off and leave the thin film that is left β that is what chars. "A skewer that goes on dripping steams itself grey." What comes off is what would run, not what is stuck to the meat.
- The film is not an obstacle to browning; it is the substrate. Tandoori chicken says what the yoghurt does once the piece is in the oven: it clings to the surface and colours it, "its lactose and protein feeding the Maillard reaction." That is the route browning calls what the surface is made of before it meets heat, and it is why that guide names a yoghurt-marinated kebab as charring while a bare skewer is still pale. A coating carrying sugar and protein browns sooner than clean meat, not later.
- A paste never had free liquid in it to begin with. Jerk chicken blends its marinade to a paste, works it into the slashes, and roasts for three-quarters of an hour with all of it on before the char at the end. Nothing is patted off because nothing would drip. Where the piece sits matters for the same reason: tandoori goes on a rack over a tray, so what does come out leaves instead of pooling underneath and steaming the piece from below.
- And at the far end, a dish wets the surface deliberately and keeps doing it. Anticuchos brushes oil stirred into the leftover marinade on at every turn, to "stop the surface drying" and layer the flavour β over a fire fierce enough that what lands leaves again as fast. That page also states the condition that makes it safe: "there is no sugar in it, so it will not burn." That is the bound on all of this. Where there is sugar, the rule in grilling takes over instead: jerk chicken pays for its sweet paste with medium heat and time, and satay keeps its brushing to the last minute over the fire.
So the question at the pan is not wet or dry. It is whether what is on the surface would run, and whether it brought anything of its own to brown with.
Thickness is a proxy; the variable underneath it is how much surface there is
The rule above is right and the table states it one dimension short. If the action happens at the surface and stops at one to three millimetres, then what decides how much of the meat a marinade ever reaches is the amount of surface per unit of meat β and that is set with a knife, before the marinade is mixed. Thickness moves it in one direction only, and the knife has more than one way to move it.
- Cut the piece up. Souvlaki goes into the bowl as 2-3cm cubes, so the seasoned layer is most of every bite where the same weight left as a chop would have an unseasoned middle. Char siu writes the arithmetic down: cut into long strips, "a 600g piece has several times the surface a joint would, which is what lets a three-hour marinade reach a useful depth and what gives the glaze somewhere to sit."
- Cut into the piece and leave it whole. Jerk chicken says "the slashes are what make the overnight marinade worth anything" β deep cuts into the flesh with the paste worked into them. The table above would file that bird under "whole legs, thick joints" and hand you the time; the slashes are what buy the surface, which is the half the row does not mention. Galbi does the same to the thick patches of a short rib.
- Take the thickness to where the question disappears. Bulgogi puts it plainly β "a 2mm slice has no inside to worry about." Every part of the meat is then inside the reach, the limit stops binding altogether, and thirty minutes really is enough.
None of this depends on there being a liquid, which is why it governs the dry rub above just as much. Suya slices the beef thin, then lays the slices between sheets and beats them thinner still: "a dry seasoning only ever sits on the surface, so more surface and less thickness is the seasoning."
Knife skills makes the case that uniform size is a cooking decision rather than a cosmetic one. This is the same cut answering a question that guide does not ask β not how evenly it cooks, but how much of it you can season at all. So when a marinade is not reaching far enough, the useful move is usually not a longer time. Past the limit, time returns nothing and surface returns in proportion.
There is also a direction that soaks after cooking
This guide has assumed raw, then marinate, then cook. Some dishes reverse it, and in doing so they get past the one-to-three-millimetre limit by another route entirely.
Leave something already cooked and porous sitting in liquid and, as it cools and contracts, it draws that liquid inward. Japanese Oden is built on it: thirty minutes of simmering plus a night of cooling beats ninety minutes of simmering, because while it is heating water is being driven out of the solids.
Jangjorim reverses the order for a different reason. Soy sauce added at the start firms the muscle fibre and gets in the way of the collagen converting, so the beef is simmered in plain water until fully tender and the soy goes in after that. Same ingredients, same pot, only the order differs.