Developing repeatable research criteria

Each prospect row carried the reason it had been kept or dropped, and those reasons repeated across lists. They covered relevance to the client's subject, the presence of a passage a link would fit, whether the site accepts contributions and whether an expected fee sits within a realistic campaign budget.

Collecting them into one detailed working document was a short step from there, and it turned a set of habits into something that could be stated explicitly.

The document set out each consideration in enough detail to be used as instructions, so it could be given to AI for removing clearly unsuitable opportunities from a discovery list.

The filtering handled the candidates that were obviously off topic or outside the client's situation. Everything that survived was reviewed manually.

Whether a page has a passage where a mention reads naturally is established by reading it. The same applies in forum research, where discussions can meet every stated check and still differ in whether a brand mention would be welcome, because that depends on what the participants are asking for.

The written considerations narrow a list to candidates worth reading. The reading is where the judgement happens.