How it works

What each article has to survive, in order

An article is dropped the first time it fails one of these — it never gets asked the later questions.

• 1. Is this actually about the account?Dropped as “Different company”
  A name match is not a company match. The most common reason a story gets rejected is that it names the right company but describes a completely different business. The story mentioned a name matching this account, but was actually about a different business. Common with short or shared brand names.
• 2. Does it actually matter to the business?Dropped as “Not commercially relevant”
  Plenty of real, factual news is set aside here because it has no bearing on revenue — a charity sponsorship, an award, a routine product launch. Only events with real commercial weight move on.
• 3. Is it a risk or an opportunity?Dropped as “Neither risk nor opportunity”
  Every surviving story is read against your definition of risk and opportunity. Whether this threatens the account or opens a door. Risk means something that could cost us the relationship; opportunity means something we could sell into. If it is neither, it is dropped no matter how interesting it is.
• 4. Is it urgent enough to interrupt someone?Dropped as “Not urgent enough”
  Rated at least 3 out of 5 for urgency — currently “Worth knowing.” You can make this stricter or looser on the Schedule screen. How urgently this needs attention. It is set when the article is first read, based on whether it changes what someone should do this week, this quarter, or not at all.
• 5. Was the reading reliable?Dropped as “Too uncertain”
  Every judgement also carries a confidence score. Below half, it is set aside rather than risk sending something misread. This floor is fixed in the system — nobody can turn it down to let more through. The system's own estimate of how well it read this story, not a measured accuracy score. Lower confidence usually means the article was short or vague, not that the event is unimportant.
• 6. Have we already told you about this?Dropped as “Already sent”
  The same story chasing the same account across ten publications only needs to reach you once. This story was included in an earlier brief, so it is not repeated here.

Most articles are discarded, and this is by design, not a shortcoming. Every one of them keeps the reason it was dropped, so the judgement can always be checked rather than taken on trust. See the articles from the last brief.

• Does it run? Every run is recorded with what it found and how long it took. A failure raises an alert by email, because nobody watches a dashboard.
• Does it land? Delivery is tracked per person, so “three of five received it” is visible and the other two can be retried.
• Is the judgement good? Readers rate items in one click. Already knew it is counted separately from not useful — the first means too slow, the second means wrong, and they need different fixes.