Follow this set of instructions if you have your own third party sample provider.

1. In the Surveys app, author the survey.

   In the survey, include hidden questions to capture identifying fields in case you need to verify participant status later.

   The hidden question captures identifying information passed from the panel company to the Alida platform. If you are capturing text, number, or ID strings, create a Short Answer or Long Answer question.

   
2. Go to Distribute. Create a new activity link and get the live link.
   
3. Append the hidden question name to the live link.
   1. Copy and paste the link into a text editor.
   2. At the end of the link, type a question mark (?).
   3. Type the hidden question name (for example, RID).
   4. Type an equal sign (=).

   Example
   https://doctest03.van.sit1.alidalabs.com/c/a/5wWSPPO2Zq47gTPWniWGYM?RID=

4. In the panel company's project, add the appended live link.

   Depending on your situation, you may complete this step yourself or get a project manager from the panel company to help you. If a project manager is helping you, send them the appended live link from step 3. The panel company must append the identifying value to the end to pass back to the survey. The panel company also prepares a sample that meets your needs, as well as customized links for tracking different participant statuses:

   • Complete
   • Over Quota
   • Disqualified

   
5. In the Surveys app, in your survey, add End Survey actions for each participant status.
   • Set the End survey behavior to one of the options that includes a redirect URL.
   • Copy and paste the customized links in the Redirect to URL fields.
   • Add pipes at the end of the redirect URLs to insert the identifying field captured by the hidden question.
   • Test your survey and ensure the survey logic redirects participants to the appropriate End Survey action.
   • When you are satisfied, publish your changes.

   

Follow this set of instructions if you are using Distribute > External Audiences.

1. In the Surveys app, author the survey.

   These authoring steps are required:

   • Include screener questions for any audience criteria not covered by the third party sample provider.

     For example, if you only want participants who shopped at a specific store in the last month, you need to include a question that asks for this information and disqualify participants who answer "No."

   • Include questions for audience criteria that are covered by the third party sample provider, and that you'll you need to segment or filter data later.

     You cannot reference the third party sample provider's audience criteria in reporting, so you'll need to capture this information in your survey if you want it for analysis. For example, you can include questions about gender, age, region, household income, or any other demographic information.

   • Add survey logic and End Survey actions for Complete, Disqualified, and Over Quota participant statuses.
   • Test your survey thoroughly.

   These authoring steps are not required, but are recommended as best practice. They can help you identify bad data later:

   • Add a speed check.

     Speed checks help you identify bad actors who are speeding through the survey instead of taking the time to answer questions thoughtfully. Generally, participants who complete the survey in less than 30% of the expected survey length can be considered speeders.

     For modern surveys, you can time how long it takes participants to complete a question using scripting. For more information, see "How do I time how long it takes a participant to complete a question? (Modern Surveys)" in the Alida Help Center.

   • Add trap questions.

     Trap questions highlight inconsistencies in responses. One example of using trap questions is asking for the participant's Date of Birth and age in separate questions on separate pages. If the math is inconsistent, this could indicate bad data.

   • Add red herring questions.

     Red herring questions help identify participants who aren't paying attention. One example of using a red herring question is asking the participant to choose the blue option from several color options. If they do not choose the blue option, this could indicate bad data.

2. Go to Distribute > External Audiences and distribute your survey.

   For more information, see Request a third party sample from Alida.

   Note:

   • Quotas

     The current release of this feature does not support quotas. To work around this, create separate external audience requests for each quota segment.

   • Soft launch vs. full launch

     The current release of this feature does not allow you to delineate soft launches vs. full launches. To work around this, create separate external audience requests: one for 10% of the full sample size you require, and then another request for the remaining 90%.

3. Monitor incoming responses in reporting.
   Use the speed checks, trap questions, and red herring questions to identify bad responses. If this is your soft launch, you can calculate the percentage of bad responses you have received and use it to predict how many bad responses you will receive overall. Then, you can adjust your full launch projections and increase the sample size to account for the extra number of responses you need to collect.
4. Flag bad responses so you can exclude them from reporting and the sample you're paying for:
   1. Identify which responses are bad.
      Export your report data and identify which member IDs are associated with bad responses.
   2. In Community > Profile Variables, add a profile variable to use as a bad response flag for the survey.
      For example, create a Single Choice profile variable called SurveyName_BadData with the values Yes and No.
   3. Import the profile variable values for SurveyName_BadData. Ensure your CSV file meets CSV file requirements.
      For example, the CSV file can have 2 columns:

      • MemberId
      • SurveyName_BadData
   4. Use the SurveyName_BadData profile variable to filter your report.
   5. Contact Alida Technical Support with your information regarding bad responses.
      Your sample size and incurred cost will be adjusted accordingly.