Web Survey Design: Knowing When the 20 Dollar Plan is Not Enough

Being a tech savvy research analyst, I’ve created my fair share of web surveys using all sorts of applications.  The free programs (with the exception of LimeSurvey) are mostly stripped down trial versions of plan based applications.  The next, and most mainstream, line of applications usually fall within the $20 a month price range, a la Survey Monkey (did anyone else notice that they bought out Zoomerang?).
You can create solid surveys at the $20 level but, depending on your circumstances, sometimes you need a little bit more.  Here are nine features that will likely push you to upgrade your service or try something new.

Custom Links

We all hate to come across as an amateur so using the default link can sometimes be out of the question.  Also, having monkey in your link can seem a bit silly.  Of course Survey Monkey knows this, that’s why you have to go up to $65 a month to replace it with their research.net domain.   While you can have your own URL for around 20 bucks a year, if you want to map it to your survey, you’ll need to upgrade/change your survey service.

Encrypted URLS

Concerned about your respondents’ privacy?  Seriously, is this even a question?  Not only should you worry about protecting the data once you have it, you also need to worry about the act of their submission.  There is a reason why every online store uses HTTPS links for their shopping carts, and that reason is encryption.

Respondent Logins and Passwords

Creating unique respondent logins and passwords is the most efficient way to ensure that respondents can stop and continue the survey at a later date.  It also makes it nice and easy to track individual responses and connect to previously collected data.  There are ways to do all of this with the $20 plans but the methods have their quirks and can lead to technical issues.

Preloading Data

With the $20 plans you can usually preload a little bit of data if you use their service to send out invitation emails.  Preloading is most important if you would like to pipe in responses based on previously collected data.  Of course preloading data has limited utility if you can’t pipe.  Which leads us to…

Piping

With basic survey plans, piping (pre-filling responses/questions based on specific information you already know) is often limited to data you collect within the survey.  But if you mix preloading with piping you have the opportunity to completely tailor the survey to individual respondents.

Complex Questions

Most of the time (or all of the time?) you are better off with simple questions.  The problem is, many researchers and evaluators are used to creating massively complex questions built to produce fewer questionnaire pages.  A rule of thumb when designing for the web, reducing the time it takes a respondent to complete a questionnaire is  much more important than reducing the number of pages.  Complex questions just don’t work as well on the web.
With all that said, there may be a time when you need to build a question with multiple response types (i.e. multiple select/open entry/ranking).  Most $20 plans do not let you do this.

Advanced Survey Logic

All logic is not created equal.  The basic plans usually give you simple logic (i.e. if they say yes here jump to page 4).  At the next level you get to create custom rules on each page saying which questions/responses the respondent will see based on previous questions, or even preloaded data.

Question Randomization/Elimination

With a $20 plan you can often randomize responses. If you need to randomize the questions (or randomly eliminate the questions), it’s time to jump up to the next level.

Self Hosting

Depending on the client, or your company, having the ability to let your data reside on an external server is not always an option.  The big companies gloat about their security, but trust is not an easy thing to give. Keeping the data to yourself will often cost you more than $20 a month.