What can we help you with?








Survey Creation Tips

Use the Right Audience

- Your own customer lists or groups for client service and satisfaction work - not just contact lists, but Twitter followers, Facebook likes and web site visitors - with Toluna QuickSurveys you can harness these people's opinions for free.
- Forums or special-interest groups for niche or job-specific audiences - including services like Quora but often very specialized sites where asking a question on a specific topic will get you an expert answer.
- Online research communities for broad and targeted consumer samples who may know nothing about your company - great for concept testing, market sizing and competitor evaluation.  Why not ask Toluna - we have millions of people ready waiting to answer your questions!

Define Your Objectives

The first step in conducting effective research is to define the parameters of your survey:

- What is the critical question facing your business?
- Who is the target audience?

Some examples:
- Is there interest in my type of product from my target market?
- What will my existing customers think of my new product development?
- Are potential customers dissatisfied with the current market offer?
- What do my potential customers like about my competitors' products?

Create Your Survey

Write clear, precise and short questions. A sure-fire way to lose your respondents is to write expansive questions that require a great deal of reading. You also run the risk of biasing their answer or, at the very least, confusing them.  Make vocabulary precise and unambiguous, and steer clear of double negatives.

Focus the questions so that each asks for just one piece of information.

Beware of bias!  By bias, we mean that the wording of your question points to a preference for a certain answer.  Survey respondents will invariably want to please you, and will be more than happy to follow your lead, effectively squashing the value of your results.  So edit your questions carefully for bias.

Here are a couple of examples:

- Avoid Yes/No questions where possible. For example, if you are looking to talk to people who are looking to purchase a laptop in the next 3 months: DO NOT ask - 'will you buy a laptop in the next 3 months?' DO ask - which of the following products are you considering purchasing in the next 3 months: mobile phone, printer, laptop, scanner, tablet etc.
- Randomize answer lists - if you have a list of answers that are non-sequential (such as a list of brands, names or locations) then make sure your survey tool has a feature that allows you to randomize your responses. This means that the first answer won't always appear at the top, reducing a natural human bias to give more attention to the answers coming higher up the list.  Of course, it does not make sense to randomize sequential answers, such as income or age bands - 16-24, 25-44, 45-54, 55+
- Vary your question types frequently, because an unending string of similar question types will almost always lead to a string of similar answers.
- Beware of open-ended questions - positioned correctly and used sparingly, open text questions are a great source of insight and word clouds can help you make sense of the results easily without examining hundreds of responses. However don't put them at the start of your survey and do limit them to 2-3 per survey maximum.
- Give the respondent a chance to opt out - make sure there is a 'do not know' option where relevant rather than forcing a respondent to select an answer that is not relevant to them.
- Use video and images to stimulate and engage your audience. Simple and effective. Toluna QuickSurveys has an intuitive search function that goes onto the web to look for images on search engines and video on YouTube, as well as allowing you to upload your own files.
- Always test your survey

1. Surveys with incorrect logic. Example seen recently: 'Are you male or female - male', 'Are you married?',' Yes,' and 'How old is your husband?'
2. Spelling mistakes show a lack of care and attention and may pay the same level of respect to your respondent's answers.
3. Bandings - make sure any question with bands includes all the possible options. Example of what not to do: 'How many times do you take a flight a year?' Once, 2-5 times, 10-15 times, 15-20 times.'

There are four errors; the first 2 are not offering an option for 'zero flights' and 'more than 20'. The third is not providing an option for '6-9 flights'. The fourth is having 2 options if you took 15 flights. It's very a common mistake, even for experienced researchers!

Be short:
- Answers - do not give a list of 50 possible answers.  Keep it to less than 10 if possible. Split across multiple questions if necessary.
- Survey length: make sure anyone taking the survey can feasibly complete it within ten minutes.  If it takes longer than ten minutes, you run the risk of respondents losing patience and giving any answer to simply finish the survey.  Experienced market researchers can craft engaging surveys longer than this, but for DIY users it is not advised.