/feedback

name: feedback
description: "Use this skill when the user needs to collect user feedback, run surveys, set up NPS, prioritize feature requests, close the feedback loop, or turn customer input into product decisions. Covers feedback collection methods, survey design, prioritization frameworks, and building a feedback-driven product culture."

User Feedback & Feature Requests

Feedback is abundant but insight is rare. Your job is not to build everything users ask for — it's to understand the problems behind the requests. This skill helps you collect, prioritize, and act on feedback without drowning in it.

Core Principles

• Feedback is a gift, but not all gifts are useful. Filter signal from noise.
• Users describe solutions. Your job is to find the problem underneath.
• "Build what customers ask for" is wrong. "Solve the problems customers reveal" is right.
• A feedback system you actually use beats a perfect one you ignore. Start simple.
• Closing the loop (telling users what you did with their feedback) is the most powerful retention tool you have.

Feedback Collection Methods

Ranked by Signal Quality

| Method | Signal Quality | Effort | Best For | |--------|---------------|--------|----------| | 1-on-1 conversations | Highest | High | Early stage, understanding "why" | | Support ticket analysis | High | Low | Finding recurring pain points | | In-app feedback widget | High | Low | Contextual, in-the-moment feedback | | NPS survey | Medium | Low | Tracking sentiment over time | | Cancellation survey | High | Low | Understanding churn drivers | | Feature request board | Medium | Low | Aggregating demand signals | | Social media mentions | Medium | Low | Unfiltered opinions | | Usage analytics | High | Medium | What users DO vs. what they SAY |

What to Use When

0-50 users:    Talk to every user. Email them. Get on calls. No tools needed.
50-200 users:  In-app feedback widget + cancellation survey + monthly NPS
200-500 users: Add a public feature request board + quarterly user interviews
500+ users:    All of the above + systematic ticket analysis

---

In-App Feedback

Simple Feedback Widget

**Tell AI:**

Add a feedback widget to my app. Requirements:
- Small "Feedback" button fixed to the bottom-right corner
- Clicking opens a simple form: text area + optional email
- Includes the current page URL and user ID automatically
- Saves to a [feedback] table in the database
- Shows a "Thank you" message after submission
- No third-party tool needed — just store it in the database

Contextual Feedback Triggers

Collect feedback at the right moment:

| Trigger | Question | |---------|----------| | After completing a key action | "How was that experience? (1-5)" | | After 7 days of usage | "What's the one thing you'd improve?" | | After upgrading | "What made you decide to upgrade?" | | After using a new feature | "Was this feature helpful? Yes / No / Needs improvement" | | After a support interaction | "Did we resolve your issue? (1-5)" |

---

NPS (Net Promoter Score)

How It Works

Ask one question: **"How likely are you to recommend [Product] to a friend? (0-10)"**

• **0-6:** Detractors (unhappy, at risk of churning)
• **7-8:** Passives (satisfied but not enthusiastic)
• **9-10:** Promoters (loyal, will refer others)

**NPS = % Promoters - % Detractors**

NPS Benchmarks for SaaS

| Score | Assessment | |-------|-----------| | < 0 | More detractors than promoters. Fix the product | | 0-30 | Average. Room to improve | | 30-50 | Good. Users like your product | | 50+ | Excellent. Strong word-of-mouth potential |

Running an NPS Survey

**Tell AI:**

Add a quarterly NPS survey to my app.
- Show to users who have been active for 30+ days
- Question: "How likely are you to recommend [Product]? (0-10)"
- Follow-up: "What's the main reason for your score?" (open text)
- Don't show to users who responded in the last 90 days
- Store responses with user ID and timestamp
- Dashboard showing NPS trend over time

**Follow up based on score:**

| Score | Action | |-------|--------| | 0-6 (Detractor) | Personal email: "I saw your feedback. Can I help?" | | 7-8 (Passive) | Ask: "What would make us a 9 or 10?" | | 9-10 (Promoter) | Ask for a review, testimonial, or referral |

---

Feature Request Management

The Feature Request Board

A public or internal board where requests are collected and prioritized.

**Simple options:**

• Canny (free tier available)
• Fider (open source, self-hosted)
• Notion board (free, manual)
• GitHub Discussions (free, developer-friendly)

**Tell AI:**

Help me set up a feature request system using [Notion / database table].
I need:
- Users can submit requests (title + description)
- Users can upvote existing requests
- I can tag requests by category and status
- Status options: Under Review, Planned, In Progress, Shipped, Won't Do
- A public-facing view and a private admin view

Processing Feature Requests

Not every request deserves action. Use this filter:

For each feature request, ask:
1. How many users requested this? (1 user = anecdote, 10+ = pattern)
2. What's the problem behind the request? (they want X, but WHY?)
3. Does it align with our product direction?
4. How much effort to build? (hours, not weeks)
5. Will it reduce churn, increase conversion, or expand revenue?

Score: Impact (1-5) × Confidence (1-5) / Effort (1-5) = Priority Score

What Users Say vs. What They Mean

| What They Say | What They Might Mean | |---------------|---------------------| | "Can you add a calendar view?" | "I need to see my tasks by date" (many solutions) | | "I want an API" | "I want to connect this to my other tools" (Zapier might work) | | "Make it faster" | "The dashboard takes too long to load" (specific page, specific fix) | | "Add more customization" | "The defaults don't fit my workflow" (better defaults might fix it) | | "Build a mobile app" | "I need to check one thing on my