How to create a survey

Create a survey using these simple tips. Ask closed questions. Keep it anonymous. Use simple language. Ask one thing per question. Keep it short. And many more.
Get started now

We've helped our customers create tens of thousands of surveys. And so we know that feeling of dread that can come over you when your boss tells you to go off and find a survey builder, or you've decided that you want to create a survey to send out to your customers.

You sit there looking at your computer screen or your phone and think: what do I even ask?

Totally get it. But we're here to guide you through this.

Before creating your survey, ask these questions

In our experience the biggest obstacle to creating a survey is not in choosing a tool that will help you do that, but in deciding exactly what you want to do.

Here's a few questions you should consider before you start writing questions:

  1. What do you want to find out?
  2. Who do you want to ask?
  3. What would a successful survey look like for you?

We're going to give you a list of things to look out for when writing your survey in a bit but before we do that, let's try understand your motives a bit.

Setting clear and focused objectives for your survey will save you a lot of hassle later on. Ensuring that you know what you want to find out from your survey will guide you through the parts of the process that can be tough: writing relevant questions; finding your target audience; presenting usable data in an interesting format.

In fact, twenty seconds spent thinking about who you're going to ask up front will free you up to think about other useful stuff later on. Want to do a customer survey? Is that to a segment of your customers or the whole lot of them? Do you want to know what employees in different job functions think? Or would you prefer to just get a pulse of the organisation as a whole?

Might seem like a lot, but we guarantee it makes the next step easier.

Understanding what success looks like for you in gathering feedback refines the work you've got to do. Successful survey creators know the outcomes they're looking for and so they can make better choices about when to close their surveys, when to report on them and how they need to engage with other stakeholders.

If you don't do this stuff in advance, you might still be successful – but for the sake of a 2 minute brainstorm at your desk, you could boost your results many times over.

Can I create a survey using Word?

Look, we're not saying you can't create a survey using Word but we'd strongly discourage you from doing this.

Here's why:

  1. Every user is using a different operating system – how can you be sure that they'll even be able to open your survey using their computer? It's much better to use a tool that allows users to respond in any browser.
  2. Regardless of whether you decide to print it out or send it to people digitally, you're still going to have to spend hours and hours collating, inputting and formatting responses. That's time you could be spending doing other thing. We routinely save our customers 10+ hours of processing time when they switch to doopoll – and the time they save with our real time results and reporting frees them up to focus on the things that actually matter to them.
  3. Even if you're a total stoic and don't mind a punishing data input regimen each week, you're still going to have to manually update and input data to your reporting tools. Why bother creating charts, graphs and reports by hand when you can have a tool like doopoll do it automatically.

TL;DR Yeh, sure. You can use Word but your productivity will go down and you'll lose value respondents.

How to create a survey online

These tips will help you make sure that you create surveys that are really useful, meet your goals and get you valuable data that you can use in a meaningful way afterwards

Choose an online survey tool

It might seem obvious to you, but choosing a tool for your next survey can be tough when there are just so many survey builders on the market. There's an awful lot to take into consideration but here's the cliff notes:

Make sure you're not just choosing a free tool because it's the easy option, Freemium tools are often a better bet. If you're collecting a lot of data, you should put some budget behind making sure you're going to be able to use that data in a meaningful way.

Most free options are things like Google Forms, or as we discussed above, Word documents (Yuck). But from as little as £9/mo (doopoll's entry package costs the equivalent of 2-3 cups of coffee or one extremely fancy pastry a month), you could get a lot more value.

We compared a whole bunch of survey tools here.

Ask multiple choice questions to start with

Most people make the mistake of jumping in with an open ended question when they design their first survey. But when you're creating your question set, make sure you use multiple choice questions for the first few questions.


Imagine the first question in my survey is and open text answer asking you to 'Describe the service that you had in our restaurant' it's likely you're going to run into a few problems.

Check out how scary that would look in the example below 👇

The first one is as a respondent that's actually quite difficult to answer. Partly because it's so broad that you'd find it hard to start – and may leave right away. But also because as a survey creator it's difficult to manage the data afterwards if the first question is an open text.

Asking multiple choice questions allows you to filter out answers you're less interested in. And it also makes it significantly easier for people to get into the flow of answering your survey!

Use a scale for labels – but make your scale words rather than numbers

One of the easiest survey fails to make is to not set up the scale on your questions properly which leads to biased or confusing options to answer with.

When you ask any kind of question that doesn't require an open ended answer, you'll need to ensure that you choose a scale that people understand. For example:

  • Strongly agree
  • Agree
  • Disagree
  • Strongly disagree

Why didn't we add a 'Neither agree nor disagree' option into that scale? We think that making people express an opinion is a good way to get more meaningful data. However, it's up to you and there are certainly times when you might want to add in a middle ground.

But in any case: be careful not to lead people to an answer by only offering something like:

  • Strongly agree
  • Agree
  • (Because they might want to tell you they disagree)

And also, don't put labels in an order that confuses people's brains. Positive answers should almost always go on the right end of a scale rather than the left, or the top rather than the bottom.

Look how weird this is in practice and yet so many people do it 👇(Flipping the order of the labels on the question below would make it so much easier for my respondents' brains to process)

Using text options instead of numeric options is also helpful for most cases. That's because humans deal primarily in words rather than numbers. Sure, they'll try to tell you how they feel on a scale of 1-5 but what's the difference between 3 and 4?. Whereas it's much easier to decide between 'agree' or 'strongly agree'

An exception to this is some of the more scientific survey types like NPS which requires you to choose a number between 1 and 10. But even those give you a pointer because 1 is usually written '1 – Very Unlikely' and '10 – Very Likely.'

Only ask things that are absolutely necessary

People's time is valuable. Yours and your respondents'. So ensure that you only ask what you absolutely need to know.

This will help you keep your survey focused and lead to better data for you to create change against whatever metric you're trying to influence using the survey.

We recommend only asking about things you'd be willing to change. Don't ask about your pricing if it's not going to change in response to survey results. Equally, you'll create more trust in your employees, customers, or respondents in general if you demonstrate that you're making changes to things they've given you an opinion on.

Don't lead your respondents to a specific option

Even with all the technology that you have at your disposal today, a simple survey is still the best way to get opinions at scale.

But that also means you've got to be a little cautious.

If you're truly interested in doing good work, you shouldn't try to lead your respondents to a specific answer which confirms your suspicions or which gives you fuel for the thing you want to do already.

While it would be nice if everyone agreed with you, they often won't. A survey is the perfect way to test your opinions and thoughts en masse.

Sure, it could be crushing if you don't hear what you want, but imagine the impact that really understanding your customers, employees, or event attendees could have for your company. Satisfaction levels could soar!

FAQs about creating online surveys

Should I use SurveyMonkey to do my survey?

How do I get survey responses?

Ask one thing per question

A double-barreled question is easily avoided but still you often see one in the wild on a survey you're answering. What is it?

It's when you ask more than one thing in a question.

Example: 'What is one thing you would change about the product and how do you feel about the pricing?'

Knowing where to start with that question is tough. Do you go for question 1 or question 2?

Look, survey questions are free. You can make two of them. Just separate double-barreled questions into two questions.

Make it as easy as possible for people to skip questions

You're doing some research into the problems that your employees may have with their boss. It's sometimes tough to answer questions about that topic (and a million others like it) because as a respondent you worry that your responses might come back to bite you.

It's not a bad idea to at least give people the option to not answer any given question – either by skipping it or by offering people 'Prefer not to say' option.

Closed questions are preferable for most things

As much as we love allowing people to express themselves freely, it's good to have some structure in your data when you're using a survey. And so you should use closed question types like sliders, scales or multiple choice for your survey where possible

There's two reasons for this:

  1. It's easier to choose an option from a predefined list – which positively impacts response rates ensuring that you get a better data set
  2. It's easier for you to analyse your data when it's mostly got closed questions. That's because you can filter based on someone selecting an option much more effectively than people who may have talked about a topic in an open ended question.

Leave the personal details until the end of the survey

One last tip for the moment: Anonymous surveys work well. People tell you they don't but we've seen millions of responses to surveys now and I can assure you that anonymity is a positive for lots of uses.

Response rates on anonymous surveys remain high. But there are also other reasons for keeping surveys anonymous.

Firstly, any data that you collect needs to remain safe to comply with legislation like the GDPR or CCPA. When respondents provide a name or an email, they allow you to identify them as individuals. If you don't have personally identifiable information then it's easier to protect your respondents and stay compliant.

Secondly, if you do decide that you want to collect personally identifiable data, only collect what you need and make it optional to do so.

doopoll is anonymous by default and only allows you to link a respondent's responses to their personal data with their consent. We do this by allowing you to turn on email and name capture at the end of the survey as an optional extra for your respondents.

And we make sure that all your data is kept secure and used only by you.

How long should my survey be?

The question of how long your survey is is really tough. A better question to ask is:

How many questions do I need to ask to get the information I need?

It may be that you only need around 3-5 questions to get the answer you're looking to get. But if you need 100 questions, that might be perfectly valid too.

An our data shows that conventional wisdom that the more questions you add, the fewer people respond just isn't true.

For example, we found that on employee engagement surveys, the response rate actually goes up when you add more than 10 questions.

In general, across a sample of over 3000 surveys that have been created on doopoll, we found that there was almost no negative impact on response rate when more questions were added. However, when you add more than 23 questions, it seems that the response rate is more erratic than surveys with fewer than 23 questions.

A chart showing how response rate changes when more questions are added
How long should a survey be?

Where next?

Now that you've got the basics of survey creation in mind here's a few more actions you might like to take:

  1. Create your survey on doopoll – it's free to get started and we've got amazing customer support to help you if you get stuck – here's a link to sign up
  2. Check out our blog for more survey tips and tricks!

Easy to use survey builder.
Create as many surveys and questions as you like.
3-4x more responses.
Responses appear in real time.
Try it free