Online Surveys Explained: How They Work and How to Run One

Anyone can publish a survey in ten minutes. Fewer people end up with results they can actually use. This online surveys guide is for the second group: students, small business owners, charity staff, community organisers and anyone else who needs honest answers from real people and wants to avoid the common traps. We will go from the first question you should ask yourself to the final summary of findings.

What an online survey is, and what it is not

An online survey is a structured set of questions delivered through a web page or app, with answers collected into a spreadsheet or dashboard. Its strength is reach and speed: you can ask hundreds of people at almost no marginal cost, and responses arrive already digitised.

Its limits are just as important. A survey tells you what people say, not necessarily what they do. It cannot explain a surprising answer unless you ask a follow-up. And it can only describe the people who chose to respond, which may not be the people you care about. Keeping these limits in mind will improve every decision that follows.

Step 1: Decide what you need to know

Before opening any tool, write one or two sentences completing this prompt: “After this survey, I will be able to decide whether to ______.” If you cannot finish it, you are not ready.

Good survey purposes are specific and tied to a decision:

• Should we change our opening hours?
• Which of three workshop topics should we run first?
• What stops members from renewing?

Weak purposes sound like “get feedback” or “understand our audience”. They produce long questionnaires and vague findings. Turn each purpose into a short list of questions you want answered, then write survey questions only for those.

Step 2: Identify who you need to hear from

Your population is the whole group you want to learn about, such as customers from the last year or residents of one village. Your sample is the portion who actually answer.

Probability and convenience samples

A probability sample picks people at random from a complete list, which lets you generalise with confidence, but it requires a list and effort. A convenience sample is whoever you can reach: followers, mailing list subscribers, people at an event. Most small surveys are convenience samples, and that is acceptable if you describe the results honestly. Say “of the people who replied” rather than “people think”.

How many responses do you need?

There is no universal number. It depends on how precise you need to be and how varied the population is. For a simple internal decision, a few dozen thoughtful replies can be plenty. If you want to compare subgroups, you need enough responses in each group. Online sample size calculators exist, but they assume random sampling, so treat their output as a guide. If the stakes are high, consult someone with research training.

Beware non-response bias

People who respond often differ from those who do not. Unhappy customers and enthusiastic fans both tend to reply, while the neutral majority stays silent. Follow-up reminders, a short survey and a clear explanation of why it matters all help reduce this skew.

Step 3: Choose a format and tool

Match the delivery method to your audience.

• Email link works well for customers, members and staff.
• Website pop-up or embedded form catches people in context but can annoy them, so use it sparingly.
• QR code suits events, shops and printed material.
• Messaging apps and social posts are quick but reach a self-selected audience.

For tools, free plans from well-known providers handle most small projects. Our roundup of free survey tools compares options on question types, export formats and limits. Whichever you pick, check that it works well on mobile, because a large share of respondents will answer on a phone.

Step 4: Write the questions

Question writing is where most surveys succeed or fail.

Use plain language

Write as if talking to a friendly stranger. Avoid jargon, abbreviations and double meanings. “How satisfied are you with our service?” is clearer than “Rate your overall experience of our service delivery.”

Ask one thing at a time

“How helpful and friendly was the staff?” asks two questions. Someone could find staff helpful but cold. Split them.

Choose the right question type

Type	Best used for	Caution
Multiple choice	Clear categories	Make options complete and mutually exclusive
Rating scale	Measuring agreement or satisfaction	Label the ends and keep the scale consistent
Ranking	Priorities	Tiring if the list is long
Open text	Reasons and ideas	Harder to analyse, so use sparingly
Yes/No	Simple facts	Misses nuance

Avoid leading and loaded wording

“How much did you enjoy our excellent new menu?” nudges people towards praise. Neutral phrasing gives more honest data. We go into detail in our guide to writing unbiased survey questions.

Include “not applicable” and “prefer not to say”

Forcing an answer produces false data. Offer an exit where a question might not apply or might feel sensitive.

Step 5: Keep it short and well ordered

Every extra question lowers the chance that someone finishes. Remove anything that does not connect to your decision. Start with easy, engaging questions, group related topics together, and place demographic questions near the end unless you need them for routing. Tell people at the start roughly how long the survey will take, and make sure that estimate is honest. Our article on designing a short survey people will finish goes through this step in more depth.

Step 6: Test before you launch

Ask three or four people, ideally some who resemble your real audience, to complete the survey while you watch or ask them to explain their thinking. You will usually catch confusing wording, broken skip logic and missing answer options. Test on a phone and a desktop. Then submit a test response yourself and confirm that it appears in your results the way you expect.

Step 7: Handle data responsibly

If you collect personal data from people in the UK, data protection law applies. For most small projects, a few principles go a long way:

• Collect only what you need. If you do not need names or email addresses, do not ask.
• Say who you are and why you are asking. A short privacy notice at the start is good practice, linking to your fuller policy if you have one.
• Store results securely and restrict who can see raw responses.
• Do not keep data longer than necessary.
• Take extra care with sensitive topics such as health or ethnicity, and with children.

This is general information, not legal advice. For anything beyond a simple survey, check the guidance published by the Information Commissioner’s Office or ask a qualified adviser:.

Offering an anonymous option can improve honesty on delicate topics, but be sure your tool really is anonymous, since some record IP addresses or require sign-in.

Step 8: Launch and promote

Announce your survey where your audience already pays attention. A personal message from a named person tends to work better than a generic broadcast. Explain what the survey is for and what you will do with the results. If you can, share what changed after a previous survey, which shows that responding is worthwhile. Set a clear closing date, and send one or two polite reminders rather than many.

Incentives such as a prize draw can lift response rates but may attract people who care only about the prize. If you offer one, state the terms clearly and make sure it complies with relevant rules for prize promotions:.

Step 9: Clean and analyse the results

Clean first

Export your responses and look for obvious problems: duplicate submissions, rushed answers where someone selected the same option throughout, and incomplete responses. Decide in advance how you will handle these, and apply the rule consistently.

Start simple

For most small surveys, simple counts and percentages are enough. Report how many people answered each question, since not everyone answers everything. Cross-tabulate only when you have a specific question, such as whether newer members answer differently from longer-standing ones, and be careful about drawing conclusions from very small groups.

Read the open-text answers

Group comments into a handful of themes and note how many mention each. Quote a few representative comments, with permission or anonymised, to bring numbers to life.

Be honest about uncertainty

State your sample size, how respondents were recruited and any known limits. Avoid phrases such as “proves” or “most people think” unless your method supports them.

Step 10: Report and act

A short summary beats a long report. Lead with the answer to your original question, then support it with three to five key findings, a note on method and a list of next steps. Share results with respondents where appropriate. A survey that leads to no visible action teaches people not to respond next time.

Common mistakes to avoid

1. Launching without a clear decision to inform.
2. Asking questions you will never analyse.
3. Making every question mandatory.
4. Using long grids of rating scales that bore respondents into straight-lining.
5. Ignoring mobile layout.
6. Reporting small differences as meaningful.
7. Collecting personal data you do not need.

How to Design a Short Survey People Will Finish
Best Free Survey Tools for Small Projects

Frequently asked questions

How long should an online survey be?

As short as your purpose allows. Many practitioners aim for a few minutes at most for general audiences. Test the completion time with real people and trim until it feels light.

Should I make questions compulsory?

Only where an answer is essential to your logic. Too many required questions lead people to quit or to enter random answers.

Can I run a survey without collecting personal data?

Yes. Many surveys are better anonymous. Check that your tool does not quietly record identifiers you do not need.

How do I know whether my results are reliable?

Look at your sample size, who responded, how questions were worded and whether the findings are consistent across groups. If a decision carries serious consequences, seek expert help with design and analysis.

Conclusion

A good online survey starts with a decision, not a tool. Define what you need to know, reach the right people, write plain and neutral questions, keep the experience short, protect personal data and report your findings honestly. Do those things and even a simple free survey can give you answers worth acting on.