Online surveys and quizzes are some of the easiest and most effective ways of getting feedback from your clients and potential customers.
If done right, they can help you improve features, tweak your inventory to make it more appealing, and get valuable insights into how people use and engage with your products and services. If done wrong, though, they can offer misleading answers that might steer you into making the wrong decisions.
How do you create a successful, statistically accurate survey? How do you know if the people responding are giving you valid insight into your customers’ needs and wants?
It’s not as hard as you might think — and there are a number of websites and services you can turn to for help.
How do you create a statistically valid survey?
The first step is to think about what your goals are. What are the questions you hope to answer? What do you hope to learn from your survey? How will your decisions be influenced by it?
You also need to think about the population you’re trying to reach. Whose opinions do you want? That can very much affect the questions you ask, the language and tone you use, and how you reach out to them. After all, a survey targeting 40-year-old doctors will be different than one for 13-year-old skateboarders.
Let’s say you run an auto garage and you have a good amount of business from regular clients, but you’re interested in attracting more sports car owners who need specialty auto servicing. That’s a potentially huge group of people with varied interests. So you’ll want to begin to narrow that down, perhaps by focusing on young professionals in their late 20s and early 30s who live within 10 kilometers of your shop.
How do you determine sample size?
You don’t need everyone in the population to take your survey in order to get accurate answers. Instead, you just need responses from a large enough sample group drawn from that population.
How big of a sample group do you need to survey? Survey Monkey, one of the most popular online services, has a calculator you can use.
With this calculator, you first input your desired confidence level and margin of error in order to determine the sample size you’ll need for your survey.
The margin of error reflects any differences between your sample population and the population as a whole. For example, if 80 per cent of people in your sample group answer 'yes' to a question in a survey, and you have a five per cent margin of error, then that means 75 to 85 per cent of people in the general population they’re drawn from would have answered 'yes' to the question. Most surveys have a margin of error of one to 10 per cent, though five per cent is the accepted average.
Confidence level determines how sure you can be that your sample accurately reflects the overall population. Let’s say you have 500 people in your sample group and a 95 per cent confidence level. That means if you choose a different set of 500 people from that same population to survey, there would only be a five per cent difference in their responses. Generally, you want your confidence level to be as high as possible, and 95 per cent is considered the industry standard.
How do you determine response rate?
Once you determine your required sample size, you’ll need to estimate your response rate — out of the people you send your survey to, how many will actually respond? To be on the safe side, assume the number will be low — in the 10 to 15 per cent range.
So how many people do you need to send your survey to in order to get accurate results?
Figuring that out is the easy part — just divide your sample size by your response rate. For example, if you need a sample size of 200 people and expect a 10 per cent response rate, just calculate as follows:
200/20 per cent = 2,000 people
It turns out that even a fairly small sample group can be incredibly reflective of the entire population. Believe it or not, a population of one million can accurately be surveyed with a sample of just 385 people.
How do you find participants?
If you’re not surveying people who are already on your customer list or otherwise in your database, it might seem daunting to try and figure out how to reach people. Again, it’s not as hard as it seems — it just costs a little money to get the job done.
One way is to find a survey or panel company who specializes in creating and maintaining lists of people who’ve expressed interest in participating in surveys (often in return for an incentive of some kind). It’s convenient and they can get your survey sent out quickly. The downside is that the respondents aren’t random and they’re often paid to respond, which can impact the accuracy.
Depending on the number of questions you ask and what populations you want to reach, it can be fairly inexpensive for these companies to distribute your survey to a select population. Of course, the more complicated or sophisticated your request, the more it can cost.
- Exact Data offers opt-in email addresses, postal addresses and phone numbers for over 291 million consumers and 34 million businesses. They offer hundreds of filters and combinations that let you define your perfect population.
- InfoCanada lets you create a free account in order to begin building lists, so even before deploying a survey you can start the process and think about who you want to reach.
- Research Now has a do-it-yourself survey system that lets you reach millions of respondents.
Who can help you design and distribute your survey?
There are also a number of services available to help you create your own survey and distribute it easily online. Most are low cost and easy to use.
SurveyMonkey is one of the most popular survey tools available, and is designed to be easy for small and medium-sized businesses to use. They have a variety of plans, including free options, and provide real-time results that let you track data instantly — even on your mobile device. Since they are geared towards non-experts, they offer a robust ‘how-to’ section that will help you design and implement a successful survey.
Zoho Survey is part of a bigger suite of Zoho business tools, which has a very simple user interface that makes it easy to create and deploy your survey. However, for a more complicated survey, Zoho might be a little too simple since it doesn’t offer as much customization as other options.
SurveyGizmo has a strong mobile app, and makes it easy to track your email campaigns. One handy feature lets you estimate how long it will take people to complete your survey — useful, since an overly long survey might never get completed.
Facebook's survey tool is another great place to get started. With over 1.2 billion people signed up worldwide, it’s undoubtedly one of the largest single sources of data out there. Surveys for Pages offers an easy-to-use interface that lets you target very specific audiences directly from your company Facebook Page. You can even run multiple campaigns to reach different audiences as needed. The only drawback is that you won’t reach anyone who’s opted out of social media.