Getting the online survey length just right is one of the hardest things you will ever have to do in market research. There are a number of problems with online surveys that you will come across again and again, some of which have been caused by the early days of questionnaires, for which you are now paying the price.
In terms of online survey length, making surveys too long is a huge problem. People lose interest and either do not complete the questionnaire or they only answer truthfully for the first half or so. On the other hand, making the length too short means that people will not have faith in your questions, as they will feel as if you are steering their answers and not allowing them to really say what they want to say.
Another big problem that directly impacts online survey lengths is that others have used their webpages wrong, asking visitors to complete a survey every time they visited. The effect is that people not only lose faith in that website, but they start to build a dislike to surveys as a whole. This is a significant problem, because research has shown that the number of people who complete online surveys is starting to decline. This is why, besides working on your online survey lengths, you need to work on making sure you reach the right audience.
In terms of actual online survey lengths, it all depends on what sort of questions you ask. Between 10 and 20 questions tends to be more than acceptable. Try to never ask less than 7 questions or more than 35 questions. The first means that people won’t trust your intentions, the second means they will lose interest. Also, be very honest about the number of questions at the start of your questionnaire, so that people know what to expect and allow them to pause and come back to the survey, particularly if it is a longer one.