Collect fewer references. Study them more.

Saving a reference is easy. Explaining what makes it useful takes a little longer. A folder full of beautiful screens does not automatically tell me how to make the next design decision.

I find a reference more useful when I can name the question it helps answer. How does this page establish a reading order? How are similar choices distinguished? What happens when there is more content than the layout expects?

That changes the way I look at it. Instead of collecting the overall appearance, I can examine a specific relationship between content, hierarchy, and interaction.

I also want to look beyond the most polished state. A dashboard with carefully selected sample data may tell a different story once names become longer, lists are empty, or several actions compete for attention. Those states reveal how much of the design comes from a working structure.

Recreating a small part of a reference can be a useful exercise, provided it stays an exercise. Then I can change the content or constraint and see which choices still work. That is where understanding becomes more useful than resemblance.

The goal is not to avoid inspiration. It is to know what I am learning from it and whether that lesson belongs in the problem in front of me.

Before saving another screen, I want to be able to finish one sentence: this works here because…