Ha ha... I do, actually. It is nothing more than extremely low viscosity silicone-based lubricant/hydraulic fluid--used as ferro-fluid, lubricant, TT dampers and lifts, and "shock oil" in RC toys. You're getting the fluid for cheap, but packaging and shipping an ounce or two is what kills you.
Yes, not much magic there, but the formulation is slightly more complicated than this and I wouldn't necessarily use random fluid. I looked into cooking up a batch of ferrofluid for speakers which I still have not rebuilt. I, too, was put off by the expense for an ounce of the stuff and also by the fact that not knowing what is in it, I'm not sure I trust the formulation to last.
The basic idea behind ferrofluid is creating a colloidal suspension of tiny magnetic particles (mostly hematite and magnetite) in a constant-viscosity oil plus some surfactant as a dispersant.
Surfactant: The surfactant is very important, as it functions as a dispersion agent by chemically binding to the magnetic particles, surrounding each one and repelling similarly surrounded particles. This prevents clumping. So the surfactant must be chosen to support binding to the iron and being compatible with the silicone.
Oil: Which brings us to the oil viscosity issue, which really matters. The original ferrofluids used conventional hydrocarbon oils, but these tended to suffer lowered viscosity with temperature and also tended to crosslink into sludge over time. (Some oils had hydrocarbon thinning agents which rapidly evaporated, but as a general rule oil subjected to oxygen in the air and heat will, over time, polymerize into a tacky mess like the cooking oil deposited on kitchen cabinets) That's why the move was made to silicone oils as these were more temperature stable and tended to not as readily crosslink. The whole point is to cool the tweeter, so the temperature is hot enough to matter. But the downside of low-weight silicone oil is that it tends to evaporate.
Particle Formulation and Size: Then there's the particle diameter, and smaller is better. Sometimes instead of pure iron particles ferrite blends with manganese and cobalt are used. These cost a lot more. Pennies for us in the quantities we use. But the quality of the formulation depends upon what particle and what size. Anything larger than 10 nm tends to not stay in solution.
TL/DR: all ferrofluids are not identical and it behooves one to do some investigation as to quality.
Anyway, I doubt anyone is going to cook up a batch, but it likely matters where one purchases the replacement fluid, as the manufacturer may not have made it for speakers or may not have properly chosen the oil, surfactant, or particle size. One never knows who cooks up this stuff, and not everyone respects the chemistry. So only purchase the finest blue ferrofluid.