Rational decisions serve to back up whether or not to go through with choices, after deciding on an emotional level. There is no data to answer the question, “Should I open a restaurant?’ Answering that would be the classic example of overthinking. “Do I want to open a restaurant?” is the only question that can be answered, and that’s an emotional choice. Once that’s decided, data can be used to decide, “Should I open THIS restaurant.” Where “THIS” is a reference to THIS location, THIS menu, THIS decor, THIS service type, etc. For those cases, rational decisions can evaluate things, but the decision to be a restauranteur itself is entirely non-rational.
Testing the viability of a restaurant could be done by handing out food samples on the street outside where the restaurant will be located. If 50% of the 200 people who sampled don’t like it, now there is some numbers. You’re not looking at the full town showing up, you’re looking at half the town. Can the half that likes it keep you in business? Depends on the size of the town. This is the usefulness of data, but it doesn’t changes the answer to that first question, it only fills in the details comparatively.