Insights for teams who act on what they hear
Practical writing on survey design, customer and employee feedback programs, and research methodology — plus a running record of what AI actually changes about this work. No fluff pieces; everything here is meant to be used.
Build a survey, share it and read the answers, by asking Claude
You can now add SurveyRock to Claude. Ask in your own words, and Claude builds the survey, publishes it, gives you the link and reads you the answers. Here's what that looks like.
What Is a Good NPS Score?
There is no official good NPS score. Here's what the number means, what real published data exists, and how to judge your own score without bluffing.
The Eight-Week Employee Engagement Survey Timeline, Week by Week
It's October, and the survey has to run in January. Here's exactly what to do each week between now and then — including the week most teams skip, and pay for later.
The Annual Engagement Survey Is a Once-a-Year Confession, Not a Feedback System
Most companies ask employees how they feel once a year, then guess for the other eleven months. Once a year was never the best way to listen. It was the most anyone had time for, and that has changed.
How to Run Your First Employee Engagement Survey
Most guides about engagement surveys start with the questions. Start with three decisions instead — who owns the results, what anonymity actually covers, and how often you can keep doing this — and the questions get a lot easier to write.
What We Built So Your Survey Results Actually Lead Somewhere
Most feedback programs are good at asking questions and good at making reports. The part in between, deciding what matters and making sure someone does something about it, is where they fall apart. Here's what we built for that part.
The Complete Guide to Employee Engagement Surveys in 2026
How to run an employee engagement survey that actually changes something — cadence, the questions that matter, real anonymity, and the 30/60/90 plan for acting on results.
The Agent Glass Ceiling: Your Survey Tool Has an API. Your Agent Still Can't Use It.
We audited our own product: 86 things a human can do in the UI, ~15 an agent can do through the API. Naming the invisible limit every 'we have an API' claim hides — and what human-agent parity actually requires.
25 Employee Engagement Survey Questions That Work (and the 25 to Cut)
Every question list gives you fifty questions and no warnings. Here are 25 vetted questions organized by driver category — and the 25 popular ones that quietly wreck your data, grouped by exactly how they fail.
Survey Completion Rates Across 400,000 Real Surveys: What Actually Moves Them
We pulled real completion-rate data instead of repeating industry folklore. Completion falls off a cliff between 15 and 16 questions, most abandonment happens before the halfway point, and mobile only falls behind desktop on long surveys.
How Long Should a Survey Be? Response Rate Data and a Decision Rule
We pulled completion rates from real SurveyRock survey data. Completion falls off a cliff between 15 and 16 questions — here's the full data, the decision rule, and the test that tells you which questions to cut.
Mobile-First Survey Design: A Practical Guide for 2026
72.3% of real SurveyRock responses now come from a phone, and rising. The question types to stop using on mobile, the 5 design rules that matter, and how to preview properly before sending.
When Should You Actually Use a Likert Scale? (And When You Shouldn't)
The Likert scale is the most common survey question type — and the most overused. When it's actually the right tool, when it isn't, and the four mistakes that quietly ruin the data.
Get our best posts in your inbox
A short monthly email: what we published, what we shipped, and one thing we learned from the data. Unsubscribe anytime.