With a fixed number of inboxes, scaling a winner means stopping something else. That trade is the actual decision, and it gets avoided when the two halves are managed separately.
Fixed capacity makes this a trade
Sending capacity is a fixed fleet of warmed inboxes. It does not expand because a campaign is working. So scaling a winner is only possible by taking capacity from something, and the honest version of the decision names what.
Managed as two separate activities, stopping losers and scaling winners, teams reliably do the second and avoid the first, and end up with everything running at half the volume it needs to produce a signal.
What scaling actually involves
Raising the cap, feeding it more leads, and giving it more inboxes. All three, or it is not scaling. Raising a cap on a campaign that is about to run out of leads does nothing, and adding leads without inboxes just lengthens the queue.
It confirms rather than re-derives
The numbers were already pulled and the verdict already reached by the monitor. This does not recompute all of it, which would be slow and would occasionally disagree with itself. It confirms the specific campaign with a fresh pull, then acts on your go.
Runs well with
Skills compound. These are the ones we usually install alongside it.