First time building a survey? Below is a list of questions that I wish I’d had in my back pocket when I built my first survey. Consider these questions as you research online survey partners. Whether you’re polling the general public, current customers, or people who have worn knee-high socks with their Birkenstocks in the past year, you can confidently choose a partner who will help you get the accurate insights you need.

Survey Design

• Will your internal team be building and designing the survey? – If your business doesn’t have the in-house expertise to build a well-designed survey, most survey partners will assume the responsibility of designing and building the survey for an additional fee.

• Does the partner deploy surveys with both quantitative and qualitative questions? Does including them both in the same survey affect price? Quantitative questions have concrete answers, are usually accompanied by a set of answer choices, and are easier for survey respondents to answer. Qualitative questions are open-ended. Because of this, they are harder for survey takers to answer and harder for you to analyze, but can often provide deeper or more specific insight than quantitative questions. Some survey partners let you include both types of questions in a single survey without hiking the survey price, while others charge more for including both types of questions.

 Statistical Validity

• How many people do you need to poll? – It depends. What are you trying to achieve with this survey? Are you trying to test a specific hypothesis or find statistical significance? The survey company will be able to answer this question for you after you’ve described your objectives. Factors, like the number of people in your target population and how representative of the target population your survey results need to be, will affect the answer. The assumption that a larger sample size represents a target population more accurately than a smaller one is not necessarily true.

Survey Fatigue

• How do you avoid survey fatigue? – You’ve put so much time into designing your survey! The last thing you want is survey respondents zoning out halfway through and selecting the first multiple choice answer for all remaining questions because they got bored. This would skew your results. Survey length can be optimized to reduce respondent fatigue or invalid answers. According to SurveyMonkey, 25 – 30 questions or 6 – 7 minutes in length is ideal for their respondent population. These numbers vary depending on the online survey provider, so be sure to ask what length the partner recommends before building your survey.

Let’s Talk Money

• Is pricing guaranteed at any incidence, or does the rate have to be above a certain percentage to guarantee the pricing? – Being picky is going to cost you. If you’re surveying a very specific audience, it’s probably going to be more expensive. Incidence rate is the percentage of potential respondents that qualify for a survey based upon given screening qualifications. For example, if the qualification that a potential respondent needs to meet to be eligible for the survey is being female, the incidence rate will be 50% (assuming that half of the population is female). A lower incidence rate means that a survey partner has to screen a higher number of potential respondents, which may increase your cost.

 Clean Data

• When do you clean your data? – Some providers will proctor surveys while they are live and disqualify respondents who are providing invalid responses (e.g. gibberish answers to free response questions or selecting the first choice for all multiple choice questions) so they are not counted in your final tally of respondents. This is called live proctoring. Other providers will remove these respondents’ answers after the survey has closed. The main difference between these approaches is timeline- if you need your data quickly, you may want to consider a partner who offers live proctoring.