Write down the denominator
A rate is incomplete without its denominator and time window. Decide whether a reply rate uses sent messages, delivered messages, or unique contacts, then keep that definition consistent. Record how duplicates and automated responses are treated.
Separate replies from useful conversations
An automatic response, a request to stop, and a qualified question are all different outcomes. Read the substance of replies alongside the counts. Do not describe every reply as positive interest or every opportunity as revenue.
Compare campaigns in context
Audience, offer, timing, and sending setup can all change the interpretation of a result. Compare groups with similar conditions and note meaningful differences. A small or early sample should remain a reason to investigate rather than a promise of future performance.
Turn the finding into a review
Choose the question the next campaign should answer, make one meaningful change, and set a review date. Keep notes about what changed so later results can be interpreted. Use the public calculators for arithmetic; use your team’s judgment for the conclusion.