Yup. Unless all those variables are controlled, testing feed seems silly to me. One data point does not a conclusion make. Even variability within batches. Why don't people just buy another feed?
Some people do just buy a different feed. And if their birds start laying better, they still don't know whether the feed caused the change, or the daylength, or something else. If the feed is just as good, it doesn't really matter. But if the new feed is more expensive, or harder to get, or has some other disadvantage, then it would be helpful to know whether the original feed really was the problem or not.
If there really is a contaminant in the feed (like when melamine was found in pet foods), then a test might find it, and then the company could recall the affected feed, and everyone's flock would benefit.
And, if the feed is not the problem, then something else IS the problem. It's a waste of time and money for everyone to try several brands of feed every year, if what they really need is supplemental light. Or if their hens are molting and nothing will make them lay until they finish. So what we learn this year can be helpful in later years too.
I think testing feed to get one data point, or a few data points, is a good idea.
And I think people changing feed to see if their flock lays better is also a good idea.
And if changing the food seems to help, I think changing back to see if the laying continues can be a good idea too.
There's no need to pick one "best" method and cut down the others. Different people trying different ideas, and all sharing the results, can be a good way to figure out what things help, what things hurt, and what things have no effect.