Creating a new survey is simple:
- Visit https://surveys.google.com/your-surveys.
- Click + NEW SURVEY.
- In step 1 (Write questions), select the appropriate question type(s) and write your question(s). Click CONFIRM.
- In step 2 (Pick audience), name your survey and select your target audience. Click CONTINUE.
- In step 3 (Confirm survey), review your survey questions and purchase responses. You also have the option here to adjust the survey frequency. Click BUY NOW or START TEST* if your survey has screening questions.
- Your survey is sent to our team for review and you receive a confirmation email once it has been started.
- You receive another email with a link to your results once your survey is completed.
You can start seeing results once your survey is approved and data is processed, which usually happens in a matter of hours.
Survey creation frequently asked questions
When Google Surveys collects responses from the “general-Internet audience,” it uses published Internet-population data sets for the target-population distribution. For example, when targeting United States, the United States government’s Current Population Survey (CPS) Internet Supplement is the target population distribution. For details, see our methodology overview. Note that our audience incomes are the median income of the areas we surveyed.
You can target questions based on inferred demographics (age, gender, or geography) from the first step of the survey-creation wizard.
For custom audiences such as dog owners or people who play golf, you can use screening questions that let you screen respondents to ensure that they are in your target audience. A screening question can be any non-binary, multiple choice question such as Yes / No / I plan to. As an example, respondents first see your screening question and then those who select “Yes” or “I plan to” answer the following questions from your survey.
Creating surveys in different languages
When targeting surveys to Android-smartphone users in specific countries, you may have the option of choosing languages. When you select a target language, your survey is shown to any user who has opted to take surveys in that language. You are required to write your survey questions in that language.
For example, if you select Spanish in the U.S., your survey is targeted to any user who has elected to receive surveys in Spanish. This might include users opting to take surveys only in Spanish, as well as those taking surveys in both English and Spanish.
Google Surveys does not provide any survey-translation services. All surveys must be written and submitted in the language that the survey is targeting. Surveys are not automatically translated.
Additionally, in some countries, you may choose to target a subset of users who speak multiple languages. When targeting these multilingual users, you are prompted to select the language of your survey questions. This lets us ensure that the question mechanics (e.g., the “Submit” button) are in the appropriate language.
Language and subpopulation validation
We run periodic validation surveys to help us verify the chosen language capabilities of Google Opinion Rewards users. These tests evaluate language fluency and proficiency. One of the primary use cases for language filtering is to target specific subpopulations. For example, in the United States, researchers may use screening questions to target bilingual respondents that identify as Hispanic or Latino. We evaluate survey results from these subpopulations in addition to language.
How and when to use language targeting
Some example use cases for language targeting include:
- Evaluating language-specific creatives, promotions, or other marketing materials
- Comparing behavioral characteristics or preferences between language populations
- Amplifying general population research with alternate languages (e.g., French speakers in Quebec)*
- Filtering to a specific audience to increase incidence of a subpopulation*
*Any surveys targeting sensitive demographics must comply with program policies.
Google Surveys does not support matrix questions, or grids with response categories along the top and a list of questions down the side, which often prompt participants to abandon surveys (Brecko, Carstens; 2006). Instead, we suggest that you break out each row of a matrix question into its own separate multiple choice or rating question. This way, respondents can focus on each question and provide more accurate answers.