Creating a Targeted Survey: Panels to Reach Your Audience

Definition: A targeted survey is used to research a specific audience, frequently utilizing a survey panel provider. A paneling service generally has information about its respondents, making it easy to screen for targeted attributes like region or job category.

This article will cover when it is best to use a targeted survey, how to create a targeted survey, screening options, and the cost of a targeted survey, along with the best practices and tips.

When to Use a Targeted Survey

The most common reason to use a targeted survey is to conduct market research. For example, when you want to launch a new product or service, gathering the opinions and attitudes of your audience will help guide the decision-making process.

Consider a company that wants to launch a new streaming service. They would like to research what show genres are most interesting, current shows they watch, and the monthly price someone is willing to pay. This information could be segmented by age, gender, or job role. This data can all be collected using a targeted survey.

In that example, the company could use a generic targeted panel like the USA. However, suppose the company identifies a specific trend, such as a male preference for a particular show. In that case, the company could do a follow-up survey with a more specific audience target, males.

A specific target audience would also be ideal if a company wanted to expand current offerings to a new customer segment. For example, a customer relationship management (CRM) software provider might want to add new features to the logistics industry. In that case, they might send a targeted survey to salespeople who work for logistics companies.

Academic projects are another everyday use of targeted surveys—anything from attitudes and opinions in public policy to general consumer research. Often an academic project will only have the ability to collect a certain number of responses from classmates or faculty; a targeted survey using a panel provider can help bridge the gap.

How to Create a Targeted Survey

The first step to creating a targeted survey is to define the specific audience you need. Having audience criteria explicitly laid out will make creating the study easier.

Once the specific audience is known, you should draft your screening questions. Regardless of your panel provider, you always want the first page of your survey to have screening questions. The screening questions serve as a double-check to ensure the people taking your survey are qualified.

Depending on your panel provider, you may need to set up direct links (skip logic) for the screening questions. Most often, when respondents do not meet specific criteria, that would be disqualified and sent to a particular URL.

After the screening questions are created, you can draft your survey. You want to keep your survey as short as possible. Here are some tips for making the study:

  • Use MaxDiff when measuring what is most important, such as asking about the most valued product features. MaxDiff is easier for respondents than a ranking question and provides better data.
  • If you have different product concepts ready, break them down into attributes and levels and use conjoint analysis. This question type will use a statistical model on the backend to quantity what is most important.
  • When conducting pricing research, don’t use an input textbox asking, “what would you pay for this.” Use either conjoint analysis with a price attribute, Gabor Granger (preferred in 90% of projects; helps determine the demand and the revenue-maximizing price), or Van Westendorp (for new products where you’re unsure what the market will accept).
  • Use a Likert scale to group related sub-questions together. This will make your survey more concise. The SurveyKing Likert scale is mobile-ready, making for an excellent user experience.
  • If you want to ask specific questions to different target audiences, such as gender, use skip logic to show only relevant questions to a particular group.

It’s a good practice to always include some open-ended text questions in your survey. Usually just asking for additional comments. This will help identify insights you may have yet to think of and help screen out low-quality responses.

Generally, a panel provider will want respondents who fully complete the survey to be sent to a specific URL. That way, the panel can log it in their system as complete. So, first, you would set this up using skip logic. Then, you would build a rule that when the last page is complete, direct to the URL as needed.

When doing studies that require job roles or company size, it might be a good idea to include questions so users can verify their information by having their LinkedIn profile or a company email address.

When sending a targeted survey using the SurveyKing audience feature, our support staff will help craft the screening questions and set up the proper redirect links for you.

Cleaning Your Data

Once you collect responses from a targeted survey, you will want to clean your data. This will ensure you are making decisions with the best responses only.

Removing speeders from your data set is the first step in the cleaning process. Generally, you would want to remove respondents over two standard deviations from the average response time. Many times for simple surveys, speeders will often rush through and complete the survey in under one minute, making it easy to screen those out using a simple spreadsheet filter.

Respondents who mark the first answer for each question should also be removed. This is called straight-lining. Removing straight-lining responses can also be accomplished using a spreadsheet formula.

If you include an open-ended text question in the survey, those answers can also be used to clean your data. Generally, machine learning is used to detect gibberish and other nonrelevant answers. But, again, just using a simple spreadsheet formula (for charter length) and filter can help quickly remove low-quality responses.

An additional quality control check could be adding trap questions to your survey. Then, people who did not mark the correct answer could easily be removed from the data set. 

When sending a targeted survey using the SurveyKing audience feature, our support staff will help clean your data and identify low-quality responses.

Screening Options

The options to screen respondents will help you reach your target audience. The technical term for screening options is called profiling. Generally, location, age, gender, and job category are standard profiling options.

Some profiling options may only have a limited number of respondents in the pool. Therefore, a specific target may only be feasible if the number of respondents in the paneling pool is substantial.

Here is a list of the SurveyKing profiling options you can currently select from:

  • Automotive – Access to a car, car brand, car type/model, model year
  • Business and occupation – Job industry, job department, job role, annual revenue B2B type, if any, revenue stream, payroll processor.
  • Education – Major, degree type, year of graduation.
  • Electronics – Products owned, internet connection types, primary internet browser.
  • Ethnicity – County born in, race, length of US residency.
  • Finance – Personal income, household income, type of financial products owned.
  • Food and Beverage – Beverages consumed, alcoholic drinks per week, fast food frequency and brands consumed, dietary preferences.
  • Gaming – Platform used, hours spent playing, types played, number of purchases.
  • Healthcare – glasses/contacts, diagnosed conditions, surgery types, insurance types.
  • Hobbies – Gambling interest, general hobbies, music, movies, sports, exercise level.
  • Household – Children, pets, marital status, persons in the household, political views, general purchasing habits.
  • Media – Amount of television watched, publications read, social media usage, streaming subscriptions.
  • Mobile – Brand owned (business and personal), the carrier used.
  • Mother/Baby – Expecting, products used, breastfeeding habits.
  • Region – General size of the city.
  • Travel – Travel habits, vacation types, recent countries/regions traveled to, airlines used.

The Cost of a Targeted Survey

Depending on the platform used, a target survey may have varying costs. For example, on the SurveyKing platform, each completed response costs around $3 for a generic panel, meaning no specific screening criteria.

The more specific the profiling criteria are (such as household income), the higher the cost for each completed response. This is because the more specific your audience is, the fewer possible respondents to take it. Thus, the reward for respondents needs to be increased to ensure high completion rates. Therefore, an important term when calculating the cost of a targeted survey is the incidence rate, defined as the percentage of people who will meet your targeted criteria.

If you want to collect personal information, such as an email address or a link to the respondent’s LinkedIn profile, that would usually increase the cost of each response by $2. This information is referred to as compulsory personal identifying information.

You might notice advertisements for targeted responses of less than $1 USD with various panel providers. Generally, low costs like this are from outsourced leads from mobile apps or games where respondents are given upgrades in return for taking a survey. While this data can be high quality, it is essential to understand where some panels’ source leads from.

How Many Response Are Needed

You can use a sample size calculator to determine the needed responses. For example, for an infinitely large population, you would need 400 responses to get 5% margin of error or about 1,000 for a 3% margin of error.

For studies that use a MaxDiff or Conjoint question, you want to collect at least 200 responses to ensure the statistical model has enough data.

If your project has a smaller budget, it may not be possible to collect 400 responses from a paid survey panel. However, even 200 responses may provide enough quality data for preliminary research in those cases. Furthermore, a sample size of 200 will still give an 85% confidence interval for an infinite population.

To boost the number of responses for a targeted survey, you could look at free options such as emailing a small segment of your current customer list. You could also ask colleagues/classmates to take the survey, post the survey on social media, or even post QR codes in your workplace or throughout campus.

Get Started Now

We have you covered on anything from customer surveys, employee surveys, to market research. Get started and create your first survey for free.