Stop 03 · User Research & DiscoveryTopics · 2/12

Surveys & Questionnaires

In short

Interviews tell you why. Surveys tell you how many. When you already have a hunch, a short survey to a few hundred people shows whether it's a big deal or a rare case.

A way to gather quantitative signal from a large number of users at once — best for validating a hypothesis you already have, not for discovering new ones.

The jaadu holds a clipboard while a crowd of students raise their hands to vote.

Think of it like this

The class farewell poll

The class rep can't chat with all 120 students about the farewell date. So she sends one Google Form with three choices. In a day she knows which date most people want. Fast, wide, but not deep.

Watch it happen

Running a useful survey

Goal

Questions

Pilot

Send

Analyse

1/5Write the one decision this survey will help with, like 'Should we build bill reminders first?'

The big ideas

What you need to know

Validate, don't discover

Surveys are best for checking a hunch you got from interviews. They rarely reveal brand-new problems.

Short wins

Every extra question loses people. Cut anything that won't change a decision.

Neutral wording

'How easy was checkout?' leans positive. 'How easy or difficult was checkout?' is fairer.

Who answers matters

200 answers from the wrong people mislead you. Check that your sample matches your real users.

See it in code

A tiny example

excel

1=COUNTIF(C2:C201, "Daily") / COUNTA(C2:C201)2=AVERAGE(D2:D201)3=COUNTIFS(B2:B201, "Student", C2:C201, "Daily")

Three Google Sheets or Excel formulas for survey results: the share who use something daily, the average rating, and how many students use it daily.

Save yourself the pain

Common mistakes

Oops

Asking two things in one question, like 'Was it fast and easy?'.

Do this instead

Split it into two questions. Otherwise you can't tell which part they were rating.

Oops

Treating '80% said they'd use it' as proof people will.

Do this instead

Stated intent is often higher than real behaviour. Treat it as a signal, not a promise.

Your mission

Build and pilot a survey

~35 min

Quick check

You have no idea yet why people abandon your app. Which method should you start with?

Go deeper

1Draft a 5-question survey for a real or hypothetical feature idea.
2Learn the difference between a Likert scale, multiple choice, and open text — and when to use each.
3Understand why surveys are weak at uncovering the 'why' behind an answer.