 |
|
Survey Tips
Creating and executing a successful Web survey is fast and easy, especially when using PeriscopeIQ, but like everything else it helps if you know the rules of the game. So, we've put together this step-by-step guide to help ensure that even your first Web survey is a complete success.
Clearly state the objective or purpose of the survey
It may sound rudimentary, but you should write down the actual purpose for conducting your survey. You should be able to state it in one or two sentences. If you need to, you can include secondary objectives.
Identify the type of people you want to participate in the survey
This will depend on your objective. You may want to simply post the survey on your site for your visitors to complete. This would be appropriate for a web-feedback survey. In other cases, you're going to want to specify who completes the survey. You may also want to control who completes the survey and limit the participants from any particular category or market segment. For example, you may want an equal number of male and female respondents, or you may want representation from four or five industries.
Identifying the type of people you want to participate for some surveys, such as employee satisfaction surveys, may be easy, but for others, it may be more difficult.
Come up with a list of potential survey participants
You may have internal lists, but if you don't there are many list brokers and sources of opt-in lists on the Internet. Opt-in means that the list was developed from people who have agreed to receive unsolicited e-mail. Spamming (the sending of e-mail to people who have not agreed to receive unsolicited e-mail) is heavily frowned upon and should be avoided. We've included some Web-based list sources with this document. Lists are also available from trade magazines and other publications.
Create your survey guide
There
are several approaches to developing a survey guide, but if you're new to this, our advice is to start by constructing the questions that will provide you with the information you need to address the primary objective of the survey. These are the most important questions and you will want to test them for both clarity and function. What do we mean by function? Basically it comes down to this: the answers to the questions you ask should provide the information needed to achieve the objectives of the survey. This isn't always as easy as it sounds. There have been many professionally designed surveys that have omitted or misstated critical questions. The best way to ensure you have asked all the needed questions is to have people actually go through the process of answering the questions and then analyze the results.
After you are satisfied with the core questions you may want to add questions that will be used to evaluate how the response to the survey differs among different groups of respondents. Questions regarding gender, marital status, income level and job title are often used. You may want to reference SurveySolutions for the Web's library which includes a list of demographic questions that can be quickly added to the survey and then edited.
There is no absolute rule as to how long your survey should be, but it should not be any longer than necessary (don't ask questions just to ask them, make sure there is value, and don't ask questions in which you already know the answers). We've deployed successful surveys that took respondents over 30 minutes to complete. In most cases however, the survey should be as concise as possible and be able to be completed within 10 to 15 minutes. Longer surveys require additional time to dowload and often result in a large quantity of incomplete surveys.
Scales are often used to rate respondent preferences and opinions. If you are planning to use a numeric scale (i.e. 1 to 10 or 1 to 5), make sure the lower and upper range of the scale is clearly defined for each question and consistent throughout the survey. Determining the proper scale is very important and should be given a lot of thought. As a general rule of thumb scales should match the ability of the respondent to differentiate. For example, if you are asking someone to measure the performance of a sales representative you would probably want to use a 5-point scale instead of a 10-point scale. The reasoning is that most customers would not be able to differentiate a single point rating difference in this case. What is the difference between a rating of 7 or 8 for a particular sales representative?
Remember, don't get too fancy or make it too complicated. Most survey objectives can be met with the use of easily understood straightforward questions.
Soliciting participation in the survey
Depending on the type of
survey you are doing, you may either choose to solicit respondents or simply post the survey and have visitors to your website complete the survey. If you decide to solicit respondents, there are several alternative methods. The most common is to e-mail people an invitation (see how to write an effective invitation). If you choose this option you can include a link to the survey URL in the body of the e-mail message. URL links are not supported by all e-mail programs so be sure to write the invitation in a way that allows people to either click a link to get to the survey or enter (copy and paste is the best method) the URL in their web browser. This is the most cost-effective method (other than just posting the survey and having it completed by people who visit your site). A well targeted invitation should achieve a 10% to 15% response rate (the response rate is the percentage of people that actually complete the survey). Alternatively, participants can be solicited by telephone, banner advertisement, mail or fax.
We're often asked whether or not it is necessary or helps to include an incentive for people completing a survey. This greatly depends on the existing motivation of the potential participants to complete the survey. Our experience indicates that if it is an area of interest to the participants an incentive is not necessary. If you do decide to provide an incentive, the best type is an incentive that rewards the individual directly. Entries in raffles and contests are less effective than purchase credits, discounts and cash incentives.
Test the survey
In the research community we refer to this as a pilot test. Before going "live" with the survey, you should completely test the entire process. This includes the solicitation, the survey itself and the creation of the results database.
Executing the survey
On average you can
expect to receive approximately 50% of the responses to your survey from a particular solicitation within one day. The vast majority of the responses will be received within four days. Avoid sending the solicitation to respond on a Friday, Saturday or Sunday. The best time to solicit via e-mail is on a Monday or Tuesday evening so that the invitation is there when the participant receives their e-mail Tuesday or Wednesday morning.
That's about all there is to it. You'll be surprised how easy it is and how quickly you can generate meaningful answers to your questions.
|
|
|
