A campaign with no replies after 30 sends and one with no replies after 3,000 look identical on a counter. One is noise. The other is a conclusion. Acting on the first wastes a good campaign; ignoring the second wastes a month.
The signal check applies statistics to that question, and reports three things about a campaign:
- Delivery — is this mail reaching inboxes at all?
- Interest — is the offer producing genuine replies?
- Harm — are recipients actively rejecting it?
Each comes back as pass, fail, or inconclusive — and inconclusive is a real answer, meaning you do not yet have enough sends to tell. No more reading tea leaves in a sample of forty.
Two design choices worth knowing about:
Delivery is measured with auto-replies, not opens. Out-of-office replies and read receipts are suppressed by Gmail and Outlook for mail filed as spam, which makes the auto-reply rate a measure of placement alone — independent of your copy, your offer, and your audience. Reply rate cannot do that; it mixes placement with persuasion. So interest is not interpreted until delivery passes. If your mail is not arriving, the check tells you that rather than blaming your writing.
The baseline is measured from your own history, not hard-coded. Auto-reply rates swing by a factor of four across the year — roughly 1–3% in ordinary weeks against 8–12% in August and late December. A fixed threshold cries wolf every August and stays silent every February. Sellvance measures your workspace's own rate over the past six months and uses that, falling back to a configurable default when a workspace is too new to have one. The card tells you which of the two it used.
The thresholds are yours to set per workspace: target reply rate, maximum tolerable unsubscribe rate, and the expected auto-reply fallback. Verdicts appear on the campaign dashboard, on the workspace dashboard, and in the campaign list, and are available through the API.