You are attempting to make a conceptual argument using a misunderstanding of systems.

The statement that private individual consumer choice does not have a material effect on child labor is demonstrable using math. It is you who are proposing magical thinking. You are also incorrect in your assessment of my analysis.

I am in no way assuming industries are static objects. I am starting from the assumption that they are complex systems with feedback mechanisms and buffers. Let's take a ridiculous example: the ocean. Sea levels are rising. A single person choosing to spit into the ocean daily has no material effect on sea level. I don't think this is a controversial statement. But neither does it assume that the ocean is a static boulder. The global water system is a complex dynamic system, and as such, it exhibits dynamic equilibrium. One person's daily spit will not rise above the level of random noise already in the system.

Similarly is large global economic systems. If you want to talk about the material effect of a single moral protestor choosing to change their private behavior, it materially affects the retailer, specifically in their gross sales. When gross sales are affected, the retailer makes choices about how much of that product to order on the next cycle. That's the effect. It hasn't reached child labor yet. But before we move forward, we have to ask the question of whether the retailed will actually change their wholesale purchasing behavior. And depending on where you live, like for example a major city, the number of people new people moving in and the number of existing people moving out, daily, is going to have a bigger effect on gross sales at that retailer than one moral protestor. Street construction is going to have a bigger impact. The weather is going to have a bigger impact. The retailer is a buffer. Depending on their size (let's assume a non-chain local large retailer), they are not going to notice your choice not to buy. Hell, my whole adult life I have never been able to sustain stable housing longer than 4 years. My constant moving has more of an effect on retailers than my moral choices.

But that's for a single retailer with a single location. When we move to the conglomerate level, individual choice matters even less. Like the water cycle constantly moving water in and out of various locales, this dwarfing the contribution of your spot, conglomerate retailers are constantly reallocating inventory based on demand. Now your individual choice needs not only to rise about the random noise of consumer behavior at a single location, it needs to rise above the random noise of consumer choices around the whole territory of the conglomerate. That noise includes the rise and fall of populations in entire states, it includes natural disasters, it includes national marketing and trends.

So until your individual choice has a material effect on the amount of chocolate candy that a retailer purchases at wholesale, your individual choice has demonstrably no affect on child labor.

And that's only the first and second buffers in the system! The food companies that make chocolate candies also make other products that involve cocoa, and they operate under different brands. If you managed to have an effect on the wholesale purchase numbers of a single purchaser from the food manufacturer, that wholesale change would need to rise above the random noise of the thousands of wholesale purchase numbers around the world. That buffer absorbs major regional changes very well, by reallocating inventory to growing markets and clawing it back from shrinking markets, without ever having to change how much cacao they purchase, therefore not having an affect on childa labor.

This is how highly complex systems maintain over long periods of time in highly variable environments. They build feedback mechanisms and layers of quantitative buffers so that small changes get aggregated into larger trends allowing larger trends to be anticipated and adapted to without threatening the system.

This is how systems work.