Positive reply rate: classify interest before you calculate

Agree on the rubric before reading the results

The following is a proposed reporting rubric, not a universal industry standard or a description of an automatic classifier. Define it before comparing campaigns and keep it unchanged during the comparison. If you revise it, record the change and reclassify the affected sample consistently.

Interest means the recipient engages with your specific offer or accepts a relevant next step. A response that merely acknowledges receipt should remain separate. Keep ambiguous replies in a review queue rather than forcing them into the positive category.

Calculate at a declared contact level

For a prospect-level report, count each contacted person once. Use the same cohort definition for numerator and denominator, and name any exclusions. This guide uses all unique contacted people as the denominator. A delivery-adjusted measure is a different calculation and needs its own label.

In a synthetic cohort of 200 contacted people, 15 send a human response. Six express positive intent, five decline and four need review. Three other contacts send only automatic replies. A positive contact rate uses six, not every response event. The pending cases remain visible until reviewed.

If no one has been contacted, the rate is undefined, not zero. If one person replies several times, those messages do not become several positive people. If sentiment changes during a conversation, apply the team's declared observation rule, such as the latest reviewed state at the cutoff.

Positive contact reply rate = positive unique contacts / contacted unique contacts × 100
Example: 6 / 200 × 100 = 3%
Human contact reply rate: 15 / 200 × 100 = 7.5%
Pending classification: 4 people (not silently counted as positive)

Keep a review trail you can explain

Use a private worksheet with a stable contact identifier, cohort, classification, reason, reviewer and review time. Link back to the conversation only in a system that the reviewer is authorized to access. A public report should contain aggregates and carefully redacted examples, not raw customer conversations.

For disagreements, have a second reviewer inspect the same context without simply copying the first label. Record the resolved rule so the next similar response receives consistent treatment. A heuristic sentiment tag can help organize work, but it should not be presented as a verified positive-intent outcome.

1. Freeze the cohort and reporting cutoff.
2. Deduplicate contacts and separate automatic-only responses.
3. Apply the rubric to the full relevant conversation.
4. Resolve ambiguous cases or report the pending count explicitly.
5. Publish the numerator, denominator, exclusions and observation date alongside the percentage.

Keep positive replies separate from meetings and revenue

A request for more information can be positive without becoming a meeting. A meeting can occur without becoming a qualified opportunity. Keep those outcomes as separate fields so you can see where progress stops. Otherwise a promising reply rate can conceal a weak handoff or an offer that does not survive qualification.

Open tracking cannot fill these missing outcomes. Apple's privacy behavior is one reason an open is not proof of a person reading or agreeing with the message. Use the conversation and subsequent action as evidence for intent and progression.

Source: Apple: Mail Privacy Protection

Use the result to choose a concrete next step

If positive replies are scarce but objections repeatedly name the same mismatch, revise the audience or offer hypothesis before expanding volume. If replies show interest but few meetings happen, examine the response time and proposed next step. These are investigation paths, not conclusions established by one percentage.

Zintara's overview reply rate is message-based. This manually reviewed positive-contact calculation is a separate reporting method, not a new native metric added by this guide. Keep that distinction in any client report. For a full explanation of the units and time window, continue with the reply-rate versus open-rate guide.