I stopped designing features in my head

There's a brainstorming skill I use that I would normally roll my eyes at. It refuses to let you start building until it has walked you through clarifying questions one at a time, proposed two or three approaches with trade-offs, written up the decision in a document, and gotten you to explicitly approve it. For a portfolio. For a layout change. Overkill, right?

Because the alternative is what I've always done. Describe a feature in two sentences, start building something close to it, and notice three days later that it's subtly wrong in a way that's now hard to undo. The brainstorming process is the cheap forcing function that catches my own laziness.

For this site I used it to plan the blog reorganization and the featured-post system. Both times I ended up with decisions I wouldn't have made without it. Like the decision that blog essays should become a simple list, not grouped by category, because category grouping was a form of emphasis I didn't actually want.

Small changes. A color tweak. A typo fix. Anything where there's genuinely only one obvious answer. Forcing a brainstorming session on those wastes the same fifteen minutes that the process saves on bigger decisions. Use judgment.

Before building any feature that will take more than an hour, answer these questions: What exactly are we building? Who is it for? What are two or three different ways to approach it? What are the trade-offs of each? Write the answers down. Pick an approach. Then start building.

The writing part is the key. Ideas in your head feel clear until you try to write them down. The vagueness that was invisible in your head becomes obvious on paper. Fix the vagueness in the spec, not in the code.