I'd personally see it a similar way to how housing is planned to be built.
1) Have recipes, but those recipes are basically just a guideline to reach an overall result. For example, root A grinded to powder with plant B equals healing poultice. What strength and attributes that healing poultice would have, no idea yet, but you know the end result would be something akin to an healing poultice. That would essentially be the "foundation" of your alchemic creation.
2) Have each individual potential alchemic ingredient give different effects to the creations they're used in, depending on how they're used. For example, bark C is used inside the aforementioned creation, when grinded to a powder with a mortar & pestle it increases cauterization, but when distilled using an alembic it instead decreases blood loss. However, if reduced to ashes, it might instead increase blood loss, an effect you would likely not want to have. This is where experimentation comes into play and where some very interesting combinations could burst forth, or where side-effects could show up. Basically, adding or modifying to an existing recipe would have the same kind of effect as building a house in wood (flammability, vulnerable to sieges, low cost, etc.) versus building your house in stone (resistant to impacts & weather, long-term durability, costly, etc.); they're still both houses, but the materials used change completely how they behave.
This system doesn't necessarily ensures everything remains secret and no one share interesting information on internet or stuff like that, but it does ensure quite a few interesting things from a gameplay perspective :
It firstly ensures that alchemy has a very strong differentiation between every region it is used into; an healing poultice from the village in a specific forest could be immensely powerful at stopping blood loss, but it could skip cauterization entirely and make it so the user still requires medical attention afterward, whereas another poultice from elsewhere could be a more end-all-treat-all solution to most basic forms of external injuries but an incredibly painful burning sensation could prevent anyone using it from fighting for a moment. This would create regional differences between the same recipes and most importantly, as a result, it would foster trade and economic ventures with not only the alchemic ingredients but also the finished products, the potions, themselves.
Second, this system also brings some extent of secrecy and unknown to the more rare ingredients out there. Say for instance a lich's phylactery could be considered an ingredient in alchemy, something so rare would obviously not have been experimented upon that much in the past and, as a result, it means pretty much nothing is known about its effect in specific potion types. This very fact pretty much means it opens room for legendary ingredients to exist and for legendary potions, created using such ingredients, to exist too, and more importantly than one such potion would most likely end up completely different from another legendary potion someone else made, even if the ingredient itself might've been the exact same. Even in more mundane play, a rare ingredient, a simple plant or something only found in one or two biomes, could end with major impacts on the overall game. Take for example a type of cactus only found in the arid deserts, and a kingdom, on that server, has a monopoly on all the arid deserts located on the server. Such a kingdom could very well decide that unique cactus is now to be used only by the military for warfare purposes, and use or trade outside of that is strictly forbidden. This leads to military secrets, to a kingdom having a potential ace in the hole for warfare, to black markets and smuggling rings for that item, to enemies paying the high price in exchange for experimentation information on that ingredient, etc.
Finally, it also sits very well with the already-planned research & technology system CoE plans to have. Everything mentioned in point 1) could be subject to such research (for example, to discover a brand-new recipe for a new type of item, or to discover substitute primary ingredients for an already-existing recipe), and everything in point 2) could still stand firmly above such research endeavors as the way to customize those further.