Milky color might be soluable oil (used in machine shop opertions) that is a stable emulsion of water and light oil. In machining, the oil fraction lubes the cutting tool and the water fraction carries heat away from the process making the cutting edge last a little longer. It is an old trick to put some of this into a radiator to 'lube' the water pump - of dubious value since there is nothing in the water pump exposed to the water except the carbon/ceramic seal element and that is self lubricating due the the nature of carbon molecule. The first thing you must deteremine is what part of the heat transfer process is the chemical supposed to address. Heat transfer is a fairly complex process but the physics is straight forward and well understood. There are several issues on the air side of the tubes and in the transfer of heat through the metal of the tubes. Since the additive does not enter these domains, you can stick to the water flow part of the heat transfer mechanism. Even there, there is only a small part of the water flow that could be positively affected by the additive. There is a boundary layer of molecules laying against the metal tube wall that don't move. Ranging out from that boundary layer are a few molecule layers that begin to move faster and faster until only a very few fractions of a 1/1000th of an inch the bulk water is flowing. Flow in that bulk water flow region is turbulent, i.e., water is being well mixed so it is absorbing heat uniformly. There is no additive in the world that could do anything in this region but hurt heat transfer, including ethelene glycol (anti-freeze) because there is no other substance which holds heat as well as plain water. Any additive, solid, liquid, or gasseous will lessen the ability of the flowing stream to carry heat away, compared to water.
That leaves the only area that these additives can improve the heat transfer process is in the very thin boundary layer next to the tube wall. In fact, this is an area of poor heat transfer because molecules are slow moving and heat is transferred by conduction only, not the efficient mixing action of convection. Unfotunately, unless there is a serious fouling on the tubing surface that the additive is cleaning off (that may be the source of the 30F temperature drops people see), there isn't enough heat transfer improvement available in this boundary layer to account for big improvements in cooling. In fact, the water side is the easiest part of the equation to fix. As long as you have reasonable water flow rate and the tubes are fairly clean, the water side can be ruled out as the problem. That side of the heat exchanger is MANY TIMES more efficient than the air side. Again, assuming the water side is in pretty good shape, most problems in cooling are on the air flow side. Every overheating hot rod I have ever had I spent virtually 0% of my time fiddling with the water and kept beating on the air flow problem and eventually got them to cool right down. Clean fins, radiator shroud, BIG air flow PULLER fan, no obstructions to air flow ahead of or behind the fan, and isolated compartments ahead of and behind the radiator (prevent hot air from recirculating from behind the radiator to in front of it!) result in solving cooling problems.
Just my observation. That is the one thing I am the world's foermost authority in - my opinion!