What is survey design for startups?
Survey design for startups is writing a short online questionnaire that measures how many people in your target group have a problem, how they handle it now and what they spend on it. Interviews tell you WHAT the problem looks like. A survey tells you HOW COMMON each version of it is.
That order matters. A survey written before any interviews can only ask about what you already believe, so it confirms your guesses with a nice chart. A survey written after 10 interviews uses the words customers actually used, and the multiple-choice options come from real answers you heard.
When should you use it?
- After problem interviews, when you have 2 or 3 competing patterns and want to know which is biggest.
- When you can reach 100 or more people in your segment: a newsletter, a community, a customer list, a partner’s audience.
- For B2C products and services, where audiences are large. B2B surveys work too, but expect fewer answers and lean on interviews more.
Skip it when:
- You haven’t talked to anyone yet. The survey will be built on guesses.
- You want to know if people will buy. A survey records what people say, and buying only shows up once you ask for money.
- Your whole market is 30 companies. Call them instead.
How do you do it, step by step?
-
Write down the one decision the survey will inform. Say: “Should I focus on busy mums or on retired women first?” Every question must help that decision. If a question is “nice to know”, cut it.
-
Start with a screening question. The first question filters out people who are not your customer. Ask about behaviour: “In the last 3 months, have you bought a face moisturiser for yourself?” (Yes / No). In Google Forms, use “Go to section based on answer” on a multiple choice question to send the “No” group straight to a thank-you page (Google’s help page shows where the option sits). Tally’s free plan includes conditional logic for the same thing.
-
Keep it to 10 questions or fewer. A structure that works:
- 1 screener
- 2 or 3 questions on how frequently and when the problem happens
- 2 questions on what they do today and what it costs
- 1 question ranking problems you heard in interviews
- 1 open question: “What is the most frustrating part of [task]?”
- 1 or 2 profile questions (age band, country, role) only if you will use them to compare groups
- 1 optional question asking for an email for a follow-up interview
-
Remove bias from every question. Read each one and ask whether it pushes toward an answer. Fixes:
- Leading: “How much do you hate long checkouts?” becomes “How would you describe the checkout on the sites you use?”
- Two questions in one: “Is the product cheap and easy to use?” becomes two questions.
- Future guesses: “Would you use an app that…?” becomes “What do you use today for…?”
- Missing options: always add “None of these” and “Other (please say)”.
-
Test it on 3 people while you watch. Share your screen or sit next to them. Wherever they pause, re-read or ask “what do you mean by…”, rewrite. This 30 minute check saves the whole survey.
-
Collect 50 to 100 screened answers. Share the link where your segment already is, with one honest sentence: “I’m researching how women over 50 shop for skincare. 3 minutes, 10 questions, no sign-up.” Post in 3 to 5 places and send a reminder after 4 days.
-
Read the results group by group. Open the responses in Google Sheets and build a pivot table that splits answers by your profile question. Read every open answer and tag each one with a 2 to 3 word label, then count the labels. Surprises live in the open answers.
What does a worked example look like?
Say Sofia, a hypothetical founder, wants to sell refillable skincare to women over 50. Her interviews showed 3 frustrations: price, too much packaging, and jars that are hard to open.
She builds a 9 question Tally form and shares it in 4 Facebook groups for women over 50 and in her small newsletter. In 10 days she gets 143 answers, and 96 pass the screener.
The ranking question puts price first. The open question tells a different story: 23 of 96 women mention hands, grip or arthritis, words Sofia never put in her options. Split by age band, that share is highest among women over 65.
Sofia now has a sharper question for her next interviews and a possible first segment. She does not have proof anyone will buy. That comes later with a pre-sale offer.
Which free tools help?
| Tool | What you use it for | Free plan |
|---|---|---|
| Google Forms | Building the survey, section-based skip logic | Free with a Google account |
| Tally | Survey forms with conditional logic and a cleaner look | Free plan with unlimited forms and responses within fair use |
| Google Sheets | Pivot tables, tagging open answers, charts | Free |
What mistakes should you avoid?
- Asking “Would you buy this?” People say yes to be nice and because saying yes costs nothing. Ask about past spending instead.
- Surveying everyone. 500 answers from the wrong people are worth less than 60 from the right ones. Screen hard.
- Averaging a 1 to 5 scale. “3.7 out of 5” hides the fact that half love it and half don’t care. Look at the spread.
- Writing options before interviews. Your options frame the answers. If they come from your head, you only measure your head.
- Asking for an email at the start. It feels like a sales trap and people leave. Make it optional and put it last.
How do you know it worked?
You can answer the decision you wrote in step 1 with numbers. Our rule of thumb for a go signal: at least 50 screened answers, at least 40% of them report the problem in the last month, and at least 20% leave an email for a follow-up. Open answers should repeat the words from your interviews.
If fewer than 20% of screened people have the problem recently, or the open answers point somewhere else, treat that as a no-go for this segment. Go back to interviews with the new pattern before you build anything.
What should you do next?
Book follow-up calls with the people who left their email, then fold everything into interview synthesis and a customer persona built from real data. When you are ready to test prices, the Van Westendorp price survey uses the same tools. And if you want an AI co-founder to review your questions for bias, try the Fe/male Switch startup game.
FAQ
How many survey responses do I need for a startup?
For spotting patterns, our rule of thumb is 50 to 100 screened answers per segment. Fewer than 30 is too shaky to compare groups. More than 200 rarely changes the decision at this stage.
Google Forms or Tally: which is better for a startup survey?
Both are free and fine. Google Forms connects straight to Google Sheets and everyone recognises it. Tally looks more modern, puts one question per page if you want, and its free plan includes conditional logic.
Should I offer a prize for filling in my survey?
A small prize draw raises response numbers, but it also attracts people who only want the prize. If you use one, keep the screener strict and drop answers that rush through in under a minute.
Updated 26 October 2026 · Written by the Fe/male Switch team