5 tips for a great online survey

Lego is fantastic. You can build entire Lego villages from the little coloured blocks, with masterful castles, soaring spires, and durable little dwellings. Or you can just as easily end up with an unstable, migraine inducing, misshapen ruin.

Online survey tools are not so different.

The popular SurveyMonkey tool and others like it are simply the building blocks to channel knowledge. From their raw materials you can construct an elegant, meaningful survey framework. Or you can construct a misleading, nonsensical mess.

An accurate, carefully conducted survey can minimise risk and inform your business strategy in a way that directly benefits your bottom line. A poor, misleading survey can cause you to take the wrong decisions, costing time and money.

The survey tool simply provides the technical means for you to potentially ask the right questions of the right people at the right time. It doesn't do the thinking for you.

So, how can you ensure the online surveys you build yourself produce the equivalent of stable Lego villages, rather than random piles of plastic rubble? Here are 5 tips to get you started.

Ask a silly question, get a silly answer!

1. Avoid ambiguity

If I ask you "What do you think the prospects are for MICE next year?" how on earth do you answer? Do I mean the furry rodents, the thing you plug into a computer, or maybe the Meetings, Incentive Travel, Conventions and Exhibitions sector? Perhaps I'm talking about Man's Impact on Coastal and Estuarine Ecosystems. Don't assume survey respondents understand what you mean by the terms you use - be specific.

2. Don't influence the answer

It's possible to accidentally influence the answers you get by asking leading or loaded questions.

A loaded question either contains certain "booby-trap" words that generate strong positive or negative reactions, or it contains a presupposition. The classic example of a loaded question is "Have you stopped beating your wife?" - any answer will confirm the false assumption as fact, even if it isn't true.

A leading question, on the other hand, suggests the answer that is being sought. "We are considering removing bottled water from our premises. Would you be prepared to use tap water in order to save the planet?" is a loaded question, as people are effectively emotionally blackmailed into saying yes.

The solution is to quality control your questions to ensure you are seeking the valid answers you need, rather than simply the answer you want.

Don't cause your respondents to expire!

3. Exercise restraint when in comes to length

Don't ask 100 questions. Don't even ask 50, 40 or 30. Your survey respondents only have a limited time on this earth! Respect them by only asking what you really need to know and try not to exceed the 5-minute mark.

When people reach the "will this never end?" pain point, they tend to abandon your survey. Ease their pain by showing them they're nearly done. A sign post to indicate that the finish line is just 2 questions away will encourage a greater proportion of people to stick with it to the end.

4. Don't ask for personal information if you don't need it

Respect people's privacy by only requesting personal information like name and phone number if you absolutely need it. Even then don't make it a compulsory field unless you must and explain to respondents what the data will be used for.

If the survey has a prize draw that requires contact data, give people the choice to join in and ask for personal details at the end, rather than at the beginning. People may be nervous to hand over contact details at the start as they fear you may be tricking them into joining a mailing list.

Right survey, wrong respondents?

5. Plan your targeting in advance

Before you even decide if an online survey is the correct research tool to use, have a very strong picture of whose opinions you are seeking. If you want to find out what engineers think about a certain subject, don't get all your responses from ballet dancers. If you're looking for a representative range of participant ages, don't use the local college newsletter as your only source.

Once you've decided who you're targeting, ensure your survey is set up in such as way that you can see not only the total number of responses that have been completed, but also those from your different target groups.

For example, if you are targeting different types of sports fans and you have 200 responses, it is important to know that these broadly reach across golf, cricket and rugby, rather than come only from supporters of a single football team. If your sample isn't representative of the people you are targeting, this needs to be addressed before you close the survey or the results won't be useful.

What can you do next?

Next time you're using an online survey tool like SurveyMonkey, keep in mind that 90% of the work is down to you. Whether the research is useful or a liability is not down to the tool, but down to what you do with it.

Keep a strong sense of the goal you want the research achieve - what is the business problem you want to solve? Establish a clear picture of who you are targeting in advance of starting work on writing questions - the right survey is meaningless if it gathers responses from the wrong set of people. Finally, apply your quality control filters to iron out ambiguity and bias in your questions and sample.

Just like building with Lego, clear purpose and careful planning is the difference between an unstable muddle and a well-constructed success.

The latest posts from the Tracking Tourism blog: