Dismissing a post handles the immediate decision. Looking at several dismissals together can tell you why you keep making that decision. I think the second part is easy to skip when the feed keeps moving.

Set aside a short review at the end of a defined period. A week is one practical option. The aim is to find a pattern you can act on, while keeping the original conversations close enough to check your reasoning.

Gather a manageable set of decisions

Use the review information available in ICPSignal and any notes your team keeps. Include the original post, why you rejected it, and the profile that produced it. You don't need a large sample to start asking useful questions, but keep small-sample limits in mind.

Avoid rewriting every rejection as ‘poor fit.’ A resolved question, another vendor, an unsupported requirement, and missing information can all lead to dismissal for different reasons.

Sort by the reason the next action stopped

Look for repeated categories and then read examples within each. Are the same words attracting vendors? Is a service boundary unclear? Are useful questions arriving at a stage where you expected a purchase request?

I'd also revisit a few decisions that felt uncertain. You may discover that the profile is fine and your review criteria need clarification. The purpose is to learn from the batch, including your own judgment.

  • Wrong audience or another seller.
  • Unsupported requirement or service area.
  • Problem already resolved.
  • Relevant but too early for the reply you planned.
  • Important information missing.

Make one change tied to the evidence

Choose the most useful correction you can explain. Clarify a requirement, add a concrete exclusion, use more accurate problem language, or change the kind of reply you prepare for early questions.

Keep a note of the old brief and the adjustment. If you change several major conditions at once, you may improve the results without understanding why. A smaller revision can give the next review a clearer question to answer.

Look for the intended effect in the next batch

Review a comparable set of new results and inspect whether the targeted pattern changed. Check for useful posts the new boundary might exclude. Count relevant conversations alongside the remaining misses.

Keep the revision if it helps, refine it if it partly works, and reconsider it if the evidence goes the other way. I want the process to make your customer understanding more accurate over time. A growing list of exclusions is only useful when those boundaries reflect the people you can actually serve.