Research and discovery: how to find out what to write before you write it

Why the research half is the leverage

Most content that fails is not badly written. It is written about the wrong question, or about a question the site already answered, or in a format the results page does not reward. None of those are writing problems, and none of them are fixable afterward by editing the prose.

The work that prevents them happens before the brief, costs about twenty minutes, and is almost entirely searches you can run by hand.

What the work actually is

Four questions, in order. Does the site already cover this? What does the results page for the query look like, and what does that mix say about what the searcher wants? What do people ask about this that the current results do not answer? And what have competitors already built around it?

Each of those has a manual method that costs nothing, and each has a point where volume makes a tool worth reaching for. The guides here cover both halves — the method first, then where it stops scaling.

Doing it by hand first

A tool that returns an answer without showing its reasoning teaches you nothing about the category you work in. Running the searches by hand for a few topics builds the judgment that makes the tool output readable later — you learn what a normal results page looks like for your queries, which means you can recognize an abnormal one.

The tools are worth it when the question is settled and the volume is the problem. Asking one query what its results look like is a search. Asking three hundred is a job.

Where research stops

Research has a stopping point, and it is earlier than most people think. Once you know the topic is uncovered, the intent is clear, and the competing pages are beatable, further searching does not improve the brief. At that point the honest next step is to publish and measure, not to keep reading.