For an influencer outreach reply rate, count each unique creator with a human reply once, then divide by the unique creators contacted in the same campaign group. Exclude automatic responses from the numerator. Keep positive replies and accepted offers as separate counts. If you use delivered recipients as the denominator, label that rate separately and use only replies from that delivered group.
These are recommended reporting definitions. Your email provider supplies message events; your team must define what counts as a creator response and an accepted offer.
Choose the unit before calculating
Use one creator-campaign record as the unit for recruitment reporting. Attach the creator's email addresses, manager contacts and message threads to that record. A creator with two email addresses and three follow-ups still contributes one creator to the denominator.
Keep a second, message-level report for sending operations. It answers different questions.
| Count | Recommended meaning | Unit |
|---|---|---|
| Sent | Messages your provider records as sent | Message |
| Delivered | Messages with the provider's delivery event | Message |
| Unique contacted | Creators with at least one sent message in this campaign group | Creator |
| Unique delivered | Contacted creators with at least one recorded delivery | Creator |
| Human replied | Creators with a qualifying human response by the cutoff | Creator |
| Positive replied | Creators whose response expresses interest in the proposed opportunity | Creator |
| Accepted offer | Creators who explicitly accepted the documented scope and commercial terms | Creator |
For a shared manager inbox, record which creator each request and response concerns. A manager's reply about one creator does not count as replies from everyone on their roster. An ambiguous response stays unassigned until a reviewer resolves it.
The accepted-offer label above is an operational stage. It does not decide whether a legally binding contract exists. Track agreement completion separately if your workflow requires it.
Keep delivery events separate from replies
Amazon SES defines a send as a successful send request. SES can record that event even when suppression prevents delivery. Its delivery event means the recipient's mail server received the email. Neither event establishes that the creator read it.
Map your provider's statuses before using them. Do not quietly rename "sent minus known bounces" as confirmed delivery. If delivery events are unavailable, use unique contacted creators as the denominator and state that delivery is unverified.
Authentication belongs in a separate sending check. Google's sender guidelines require SPF or DKIM for all senders to personal Gmail accounts, with additional requirements for bulk senders. Yahoo also sets SPF or DKIM as a minimum and recommends monitoring bounces. Passing an authentication check supplies no evidence of a human reply.
Classify incoming messages before counting them
First separate automatic mail from human responses. Then classify the human responses by intent.
RFC 3834 describes the Auto-Submitted header and recommends auto-replied for automatic responses. It also warns against relying on an "Auto:" subject prefix alone. Use headers and message content together; send unclear cases for review. A missing automation header should not automatically earn a human-reply label.
The following message descriptions are illustrative, not campaign records.
| Incoming item | Human reply? | Positive reply? |
|---|---|---|
| Automatic vacation notice | No | No |
| Automated acknowledgement from an agency inbox | No | No |
| Delivery failure notification | No | No |
| Creator declines the offer | Yes | No |
| Creator asks to stop receiving messages | Yes | No |
| Manager asks for the brief to assess a named creator's interest | Yes | Yes, under this report's definition |
| Creator requests clarification without expressing interest | Yes | No, pending clarification |
| Creator accepts the stated fee, scope and dates | Yes | Yes; also accepted offer |
A human-written request to stop receiving messages counts as a negative human reply. A one-click unsubscribe event does not count as a human message. Both belong in your opt-out handling process. Use the cross-list opt-out workflow to keep that action separate from the measurement label.
If an automatic vacation notice arrives first and a person replies later, count one human-replied creator when that later reply qualifies. Keep the automation event in the history. Do not add two people to the numerator.
For positive replies, write down whether you measure "ever expressed interest" or "currently interested." Use the first for the example below. A later decline changes the current stage, while the historical positive-reply event remains. Accepted offers still need their own evidence.
A hypothetical outreach report
Assume a campaign first contacted 100 distinct creators on the same day. The team measures replies during the next 14 days, including permitted follow-ups. Every human reply in this example belongs to a creator with a recorded delivery. These invented figures illustrate the arithmetic; they are not a benchmark.
| Hypothetical count | Total | What it includes |
|---|---|---|
| Sent messages | 180 | 100 first messages and 80 follow-ups |
| Delivered messages | 172 | Message-level delivery events |
| Unique contacted creators | 100 | Each creator counted once |
| Unique delivered creators | 94 | At least one delivered message per creator |
| Human-replied creators | 24 | 12 positive, 8 declines, 4 clarification-only replies |
| Automatic-only responders | 10 | No qualifying human reply by the cutoff |
| No incoming response | 66 | The remaining contacted creators |
| Positive-replied creators | 12 | Includes those who later accepted |
| Accepted-offer creators | 5 | Explicit acceptance of documented terms |
The incoming-response groups reconcile to 100 creators: 24 human-replied, 10 automatic-only and 66 with no response. Positive replies and accepted offers are subsets of the 24, so do not add them again.
| Hypothetical metric | Formula | Result |
|---|---|---|
| Human reply rate, contacted basis | 24 / 100 × 100 | 24.0% |
| Human reply rate, delivered basis | 24 / 94 × 100 | 25.5% |
| Positive reply rate, contacted basis | 12 / 100 × 100 | 12.0% |
| Positive share of human replies | 12 / 24 × 100 | 50.0% |
| Accepted-offer rate, contacted basis | 5 / 100 × 100 | 5.0% |
| Acceptance among positive responders | 5 / 12 × 100 | 41.7% |
Counting the ten automatic-only responders would inflate the contacted-basis reply rate to 34%. Dividing 24 people by 180 messages would produce 13.3%, a mixed-unit figure that changes when you send more follow-ups.
The 24.0% and 25.5% rates can both be correct because their denominators differ. Keep their labels attached whenever you copy them into a report.
Freeze the group and the observation window
Choose the cutoff before comparing results. For rolling outreach, give each creator the same observation period after first contact. Label newer records as incomplete rather than comparing yesterday's outreach with a fully observed campaign. Choose a reporting window before looking at results when planning that comparison.
A response after the cutoff stays in the conversation history but does not silently change the frozen report. Publish a later update with its own cutoff if needed. Replies arriving outside email belong in a separately labelled cross-channel report unless your definition already includes them.
Before sharing a rate:
- Deduplicate creator records, addresses and message events. SES warns that multiple notification methods can report the same event.
- Resolve uncertain human-versus-automatic classifications, or report the unresolved count.
- Check that every numerator record belongs to its denominator group.
- Keep declined, opted-out and bounced contacts in the historical contacted group. Removing them later rewrites the result.
- Show the counts beside percentages. If the denominator is zero, report the rate as unavailable.
Avoid attaching a universal performance target to this calculation. The Modash outreach survey declines to supply a general response-rate benchmark and describes several different outreach asks. Those asks require different levels of commitment from creators.
Before your next report, write one sentence naming the unit, qualifying reply, denominator and cutoff. Then recount one completed campaign against that sentence.
