Survey Results Analysis and Findings Deck
Turns a raw survey export into cleaned data, correct significance tests, and slide-ready findings with bases and caveats on every number.
Before & after
Same prompt, same agent — the only difference is the skill.
I can help you make sense of survey data. Statistical analysis involves looking at patterns, calculating averages and distributions, and identifying correlations between questions. Different questions call for different analysis methods. You could create visualizations to show the findings, and a presentation would help communicate results to stakeholders. I'd suggest organizing the data by customer segment if possible. Would you like me to start by reviewing the raw responses?
SURVEY ANALYSIS REPORT: CUSTOMER SATISFACTION (240 responses) DEMOGRAPHIC SPLIT: Enterprise 62%, Mid-market 28%, SMB 10%. KEY FINDINGS WITH CONFIDENCE INTERVALS (95%): Overall satisfaction: 73% (71-75% range). Enterprise leads at 81%, mid-market 65%, SMB 58%. Issue: price-value perception drops 18% in SMB tier. Problem ranking by segment impact: onboarding friction (Enterprise: 34% mention, Mid-market: 47%), feature gaps (28%, 31%), support response time (18%, 22%). STATISTICAL METHODS APPLIED: Likert-scale to ordinal conversion, Spearman rank correlation for segment clustering, Fisher's exact test on categorical sub-groups (n too small for chi-square). SECONDARY FINDING: 61% of mid-market respondents cite competitor comparison during evaluation. Enterprise shows 12% coefficient of variation in renewal likelihood by geography. ACTION PRIORITIES: Onboarding redesign affects 41% of sample. Price rationalization for SMB tier. Support SLA upgrade. DELIVERABLES: Statistical summary (attached), slide deck with 12 charts, segment-level deep dives, confidence bands on all metrics, methodology appendix.
About this skill
Your survey export is sitting in a spreadsheet and the deck is due Thursday. The usual result is a pile of percentages with no bases, segment "differences" that are noise, and NPS gaps nobody tested. This skill cleans the data, runs the right test for each question type, and builds a findings deck where every number shows its sample size and every claim survived a significance check.
What it does
- Profiles the export first: completes vs partials, speeders, straight-liners on grids, failed attention checks, and broken skip logic, with a removal log you can defend
- Sets the correct base for every question, including skip-logic subsets and multi-select items that do not sum to 100%
- Picks the test that fits the data: proportion tests, chi-square or Fisher's exact, Mann-Whitney for Likert, and the proper variance formula for NPS gaps
- Corrects for multiple comparisons across cross-tabs so 1 in 20 "significant" cells is not just chance
- Labels each finding Significant, Directional, or Unreliable (small base) instead of presenting everything as equally true
- Writes the deck as sentence headlines, one finding per slide, with chart type, base note, and caveat, plus pandas/scipy code you can run on your full file
What you get
A data-quality summary, a findings table with bases, intervals, and confidence labels, and a slide-by-slide deck outline led by the few results that actually hold up.
Best for
Marketers, product managers, founders, and researchers running customer, NPS, employee, or market surveys who need numbers they can defend in front of a skeptical exec or client.
SKILL.md preview
---
name: survey-results-analysis-deck
description: Turns raw survey exports into cleaned data, correctly based and tested statistics, and a findings deck where every number carries its n and a confidence label. Profiles data quality, sets bases per question, picks the right test per question type, corrects for multiple comparisons, and leads with only the findings that hold up. Use when the user says "analyze my survey results", "here's our NPS export, what does it say", "is this segment difference significant", "turn these survey results into slides", "cross-tab this by plan and region", "my survey percentages don't add up to 100", or "leadership keeps asking if these numbers are real".
… (sign up to view the full skill)More research & data skills
View all Research & Data skills →Source-Reconciled Research Brief
Turn supplied links into a concise brief with claim-level citations, reconciled source conflicts, and explicit evidence gaps.
Source-Traced Research Bundle
Compile many sources into one bundle where every claim is atomic, sourced, dated, tiered, and conflict-checked, in Markdown or JSON.
A/B Test Power and Validity Designer
Exact A/B sample sizes and MDE with math shown, plus the traps: unit mismatch, multiple variants, peeking, guardrail power, and SRM.
Research Routing and Cited Briefings
Route research to the right method, then produce a cited briefing with evidence gaps, disagreements, and a source audit.