I want a lead score to help me decide what to read first. I still want to understand why the post received attention. A number by itself leaves too much hidden.

In ICPSignal, the fit explanation and original conversation give you something to inspect. Use them together. The purpose of review is to decide whether the evidence supports a useful next step.

Check the reasoning against the source

Read what the explanation says makes the person a potential fit. Then find the supporting detail in the post. Is the requirement explicit? Is the author seeking help, describing a solved issue, or offering a service themselves?

If the explanation depends on an ambiguous phrase, keep that uncertainty in your decision. You can choose to investigate further without treating the interpretation as a fact.

Find the requirement the score may not settle

A post can describe the right problem while leaving out a decisive constraint. Budget, integration needs, geography, or the type of help requested may still be unknown. None becomes known because the score is high.

For a hypothetical migration request, a named platform might make the topic relevant while a missing destination system prevents a confident fit decision. The next step could be one clarifying question, rather than a sales pitch.

  • Which detail supports the fit?
  • Which requirement could rule it out?
  • Is the problem still open?
  • What information would change the next action?

Keep the number within its purpose

A fit score isn't a probability that the person will buy. It also doesn't establish who is behind a username or whether the person wants contact through another channel. Those judgments need their own evidence.

I'd be cautious about choosing a cutoff from one good or bad example. Review a group of posts and look at the reasons. A repeated mistake in the interpretation may point to a clearer profile description or exclusion.

Use disagreement to improve the process

When you disagree with a match, write down why in your review notes and use the product's available review controls appropriately. Make the distinction concrete: wrong audience, unsupported requirement, resolved problem, or missing information.

Then inspect the pattern across results. If several good matches were missed by your own initial assumptions, learn from that too. I want the review to improve both the profile and the judgment of the person using it. A useful score starts a decision; the evidence makes that decision worth trusting.