That's more or less what philosophers tended to do back in Classical Greece. Unfortunately, it doesn't actually work very well.

The less information you have, the less good a position you're in to come up with plausible hypotheses. If you try and draw a map of a city before you go out and see what the city actually looks like, your map will almost certainly be wrong. Experiments are the "go out and look" step.

What sounds to you like a rational hypothesis might be very obviously wrong to someone who'd actually made the relevant observations. You could end up doing a whole lot of work theorizing and putting together a coherent model, and then go out and do your experiments and find that your theory is completely wrong and gives you nothing useful to work with. You could have saved a lot of effort by going out and looking and finding what observations your theory needs to account for.

Drawing your conclusions before doing your experiments is not just putting the cart before the horse, it's worse. It's like putting the cart in front of the horse in a narrow alleyway where it's a nightmare to get them turned around again. Humans naturally get fixated on and attached to explanations, and once you've latched onto an idea, it tends to affect how you view all the new information you get. By the time you raise an explanation to the level of attention, you really should have enough information to home in on that particular explanation out of the whole space of possibilities.