Industries grow and shrink as a result of changes in consumption. The change in outcome is a sum of all the changes in the conditions. All changes, whether we have the tools to measure and distinguish them or not.
You’re inventing a mechanism where a change being too small to be considered important means its effects are erased from reality. An organized boycott is still composed of many individual immeasurable purchase decisions. If the small actions had no way to exert their effects, they couldn’t make up for that with numbers. Capitalists would be able to will a boycott out of existence just by refusing to see it as a trend. You could will away a dune based on the impossibility of following a single grain of sand.
Again, industries are not boulders that sit static until given a big enough shove. They are dynamic processes that continually fluctuate as a result of all the transactions that did or did not happen at each stage of production and distribution. The friction of whether or not to make a change is already always being overcome by other market forces.
Industries are governed by thousands of people making endless small adjustments to the operation of the node they manage in response to the changes made by the managers of the nodes they interact with. At no point is the manager who decides by how many children the cacao company can expand its workforce going to hear about a movement becoming organized enough to be considered a boycott. They will make their decision as a result of the decisions of their nearest organizational neighbors. The layers and layers of decision makers aggregate and obfuscate the decisions made leading all the way back to the consumer(s). In the same way that it’s impossible to trace the impact of a single purchase forward from a store, it’s also impossible for the person hiring the children to trace back and selectively be influenced only by coordinated efforts, but not by silent individuals. If it is possible for a boycott to succeed, it is only because the individuals who participated each had a small impact.
You can say that it’s a negligible effect, you can say it’s unimportant when compared to the effect of [X], but your claim was “an individual choosing to stop consuming something has no material effect on child labor”. That is not true. It’s a tempting simplification that offers comforting reassurance morally. In many cases it’s close enough to true to accept as a premise. But if you take it to be literally true, then you’ve necessarily buried some magical thinking into your understanding of the world.