# Event satisfaction score — editable analysis sheet

Event/edition: [ ]  Survey owner: [ ]  Field dates: [ ]  Eligible audience: [ ]  Invitations delivered: [ ]  Responses received: [ ]  Complete ratings for this question: [ ]

**Exact question:** [ ]  **Scale and anchors:** [e.g., 1 = very dissatisfied; 5 = very satisfied]  **Optional/required:** [ ]  **Collection channel:** [ ]

| Rating | 1 | 2 | 3 | 4 | 5 | Missing / skipped |
|---|---:|---:|---:|---:|---:|---:|
| Responses | | | | | | |
| Share of complete ratings | | | | | | n/a |

**Mean of complete ratings (if useful):** (1×count₁ + 2×count₂ + 3×count₃ + 4×count₄ + 5×count₅) ÷ complete ratings. Label it “mean / 5,” not a percentage of satisfied attendees.

**Response coverage:** responses received ÷ invitations delivered, if one invitation per eligible person and delivery count is known. State exclusions and duplicates. If an open QR code or social link was used, do not invent an invitation denominator.

**Worked arithmetic only:** Counts of 1, 0, 1, 1 and 2 across ratings 1–5 give five complete responses and a mean of 3.6/5. Two people chose 5, but the 1 and 3 remain visible in the distribution. Five voluntary responses are not enough to characterize all attendees.

| Segment or edition | Eligible / invited | Responded | Complete ratings | Distribution (1–5) | Mean, if used | Collection difference / caveat |
|---|---:|---:|---:|---|---:|---|
| | | | | | | |

Compare segments only when the same question, anchors and collection method were used, and only if reporting them does not expose small groups. An event NPS recommendation question on a 0–10 scale is a different measure; do not mix it with overall satisfaction ratings.

| Finding and evidence | Alternative explanation / missing voices | Follow-up source to check | Decision owner | Action and review date |
|---|---|---|---|---|
| | | | | |

Report the count, scale, distribution and response coverage alongside any average. Treat voluntary comments as clues, not votes from all attendees. No universal “good” threshold can replace the event’s own question, audience and past comparable editions.