Eleven days of silence is a lost deal, not a slow one
A deal where the buyer has not initiated contact in eleven days while your rep has sent three follow-ups is not a slow deal, it is a lost one that has not been recorded yet. Activity counts hide this because they count messages without counting direction — two messages in a week looks healthy and is terminal if the rep sent both.
Why does the direction of a message matter more than the count?
Because an inbound reply is evidence that the buyer is still spending attention on you, and an outbound message is only evidence that your rep is. Activity reporting sums the two, which destroys the only part of the signal that predicts anything.
A thread with six messages in ten days reads as a healthy, engaged deal in every activity dashboard ever built. If all six were sent by your rep, it is a deal where somebody has stopped answering and nobody has said so.
What does a reply-latency trend tell you that a snapshot does not?
It tells you the buyer has re-ranked you. A contact who replied within four hours in March and takes nine days in June has made a decision about your priority, and they made it well before the close date slipped.
The absolute number is nearly meaningless across accounts — a procurement lead at a bank was never going to reply in an hour. The change within one relationship is the signal.
Why is a stage change a lagging indicator?
Because a rep moves a deal backwards only after they have accepted the bad news, which is weeks after the evidence arrived and usually after a forecast call has already been run on the old number.
There is also an incentive: moving a deal back is a conversation with your manager. Leaving it where it is costs nothing until the quarter closes.
What should a team measure instead?
Four things, all observable without asking anyone: days since the last inbound reply, the ratio of outbound to inbound over the last month, whether the named champion still appears on replies, and whether a new unfamiliar address has started answering on the buyer’s side.
Each is a fact about messages that already exist. None requires a rep to update a field, which is exactly why none of them decays.
The version of this that is actually useful
A list is not an insight. What changes a forecast call is a named deal, a risk, and the sentence the risk came from — so the conversation moves from "why is this still at 80%" to "what are we doing about these four".