SurveyMonkey Logo
Plans & Pricing

How to do market research: the ultimate guide

author_morgan molnar

Morgan Molnar

Learn how to conduct your own market research with survey design, sampling, and analysis best practices from our survey experts.

  1. Intro to market research: Learn what market research is, why it’s important, and the difference between various types of market research so you can choose the best method for your project.
  2. Planning: Use our frameworks to define your business questions and research goals, outline your project scope, and document your plan in a research brief.
  3. Survey design: Get expert tips for how to structure your market research survey, write good questions, and optimize for data quality—plus survey templates and a pre-launch checklist.
  4. Fielding: Read our recommendations for how to reach your ideal survey respondents, determine the right number of people to survey, and time your launch for optimal response rates.
  5. Analysis: Get best practices for preparing your dataset, interpreting and visualizing your results, analyzing trends over time, and understanding statistical significance.
  6. Taking action: Learn how to find the stories in your data, turn insights into strategic business recommendations, and inspire action from your stakeholders.

01 Intro to market research

Everything you need to know before getting started

Market research is essential to every organization.

According to the latest global market research report by ESOMAR, a global community of market researchers, roughly $45 billion is spent on market research annually around the globe. The majority of these resources are spent contracting full-service vendors and agencies that do your market research for you. But what if you could cut out the middleman and do it yourself? Doing your own market research is much cheaper, faster, and not nearly as difficult as you might think, as long as you know what you’re doing.

In this guide, you’ll learn everything you need to know to run your own market research programs and get insights to help everyone at your company make business decisions—without costly middlemen.

What is market research?

Market research is the process of collecting information on consumers’ behaviors and preferences, category trends, and/or competitive intelligence. Market research is typically conducted by organizations to inform product development and go-to-market strategy to ultimately drive business growth.

Market research can help companies answer questions like:

  • How large is the market opportunity for my product/service?
  • How does my brand stack up against the competition?
  • Which demographics are most likely to buy my product/service?
  • Which advertising campaign will resonate best with my target market?

Why is market research important?

Sure, business decisions can be made based on gut instincts alone, but doing so comes with high risk. After all, not all of us can be brilliant visionaries like Steve Jobs, who famously said, “It isn’t the consumers’ job to know what they want.” We beg to differ. Market research provides the necessary data-backed evidence to help you make those decisions with confidence. Here’s why market research is so important:

  • Market research steers your business strategy. When it comes to deciding on the company direction, heavy investment (dollars or personnel) is on the line. Market research is the essential validation that assures you that you can move forward or the warning bell that signals you need to go in a different marketing or product direction.
  • Market research helps you avoid costly mistakes. According to Harvard Business School professor Clay Christensen, 95% of new product launches fail. And launching a new product costs serious money. If you’re launching a new product, market research can give you the data you need to ensure your product is in the winning 5%.
95% of new product launches fail
  • Market research builds credibility. Whether you’re trying to create persuasive marketing collateral, become a thought leader in your industry, or impress your C-Suite, market research arms you with the facts to back up your arguments and claims.
  • Market research shows you where to go next. We’ve all heard the phrase “innovate or die.” If your business isn’t adapting to market trends and evolving consumer needs, you’ll likely get left in the dust by your competitors. Market research can diagnose when your brand is starting to get stagnant, help you know which new products/campaigns to launch, and guide you on which markets to enter next.

Types of market research

Market research is a broad term that encompasses several different types of information gathering. It can mean different things to different people and be viewed through a number of different lenses. In order to give you a comprehensive understanding of it, we’ll walk through each of common types of market research, their pros and cons, and how they’re most commonly used. We’ll talk about the differences between:

  • Primary research versus secondary research
  • Quantitative research versus qualitative research
  • DIY market research versus full-service market research

Primary research versus secondary research

Fundamentally, market research can be broken down into two major categories: primary research and secondary research.

  • Primary research is when you collect new information for the first time. Primary research can take the form of either qualitative or quantitative research (more to come on that), but the main distinguishing factor is that the research hasn’t been conducted before.
  • Secondary research is the gathering, consolidation, and summarization of data and research that already exists. For example, if you were to go digging into the US Census for population stats, you would be doing secondary research.

For the purpose of this guide, we will be focusing on primary research. Once you’re done reading, you’ll be a pro at conducting your own primary market research.

Quantitative research versus qualitative research

Quantitative and qualitative research can be done individually or in combination to get broader and deeper insights.

  • Quantitative research is the process of gathering structured, numerical data and using statistical analysis (anything from simple averages to predictive analytics) to make sense of it. Quantitative research can be done at scale and with a much larger sample sizes, because numerical results are easy to aggregate and summarize.
  • Qualitative research typically dives into a particular topic to understand the subjects’ experiences, thoughts, and opinions in detail. Qualitative research is the process of exploring and collecting unstructured information—like text, images, audio, and video—and summarizing the findings into themes. Because of this, qualitative research is done with a smaller sample size—usually anywhere from 5 to 20 research subjects.

The main differences between quantitative and qualitative research are summarized below:

Quantitative researchQualitative research
Research design:
-Large sample size
Research design:
-Smaller sample sizes
-Structured data
-Aggregates and averages
-Presented in numbers
-Unstructured data
-In-depth details
-Presented in themes
-Faster and easier to analyze results with spreadsheets and statistical software
-Larger sample sizes mean the results are more likely to be statistically significant
-Deep dive into the “how” and “why”
-Exploratory research design provides the opportunity to capture concepts or ideas you hadn’t thought of before
-Less detail around “how” and “why”
-Structured research design means it’s possible the research isn’t capturing all possible concepts or ideas
-Time consuming and difficult to analyze the findings
-Smaller sample sizes mean the results are less likely to be statistically significant

Quantitative research examples

  • Surveys, especially market research surveys, can be used to collect feedback, perceptions, and opinions from a sample of your target market. Survey data is most often quantitative (e.g.: 90% of people wouldn’t feel safe as a passenger in a self-driving car), but can also be qualitative if open-ended questions are used.
  • In-market experiments involve launching a product/campaign/website/etc. in a single geographic or demographic market (the test group) and measuring the business impact compared to a similar market that wasn’t exposed to the launch (the control group). In-market experiments are great for testing things like digital ad effectiveness or whether a product extension will win market share from a competitor instead of cannibalizing your existing offerings.
  • Implicit data collection is always-on, passive data gathering of actions or transactions. Examples of this could be capturing barcode scanning data at a store checkout to know what items were purchased at what price, or measuring website traffic, or mobile location tracking. Implicit data collection differs from explicit data collection where the information is directly provided as part of the research study (e.g.: in a survey).

Qualitative research examples

  • In-depth Interviews (IDIs) are typically conducted one-on-one and go deep into a particular topic. Qualitative interviews can be in-person, over the phone, or online. Interviews are great for exploring perspectives and feelings on a topic in participants’ own words.
  • Focus groups gather a small group of people together (typically in-person) to discuss a topic. Focus groups are ideal for getting feedback when physical interaction with a product concept is necessary (e.g.: when testing new flavors or scents, or watching a screening of a TV show pilot).
  • Observational research is when the researcher simply observes subjects going about their normal routine and documents what they see. For example, researchers could observe a shopper’s path through a retail or grocery store to better understand how to lay out the store experience. The “results” in the case of observational research usually come in the form of written researcher notes.
  • Yes, even surveys. It might seem like surveys are meant for quantitative research, but they can also be used for qualitative research. Open-ended survey questions provide a flexible way for survey takers to provide more detail around their answers, making the results qualitative in nature.

DIY market research versus full-service market research

When your organization needs to conduct market research, there are a couple ways to go about it. The two main approaches are do-it-yourself (DIY) market research or to use a full-service market research firm.

  • DIY market research is when organizations do their own market research using self-service tools like survey platforms and online survey panels like SurveyMonkey Audience. Typically, DIY market research is done opportunistically by the team that needs the results, but the research could also be done by a centralized internal insights team using similar tools. DIY market research can be just as robust and reliable as full-service market research. While it but does require some expertise this ultimate guide will teach you everything you need to know to run your own successful DIY market research project.
  • Full-service market research is when an organization pays a vendor to conduct market research for them. The vendor could be a market research firm, marketing agency, or consulting firm. Things like research design, data collection, analysis, and business recommendations can all be handled by a vendor, but it can take time and is more expensive.

Here is a summary of the pros and cons of DIY market research and full-service market research for you to consider when deciding how you’ll conduct your research:

DIY market researchFull-service market research
-Instant results
-Complete control over research design and timeline
-All of the work is done for you by the vendor
-No expertise needed

-Requires work internally to set up the research and analyze results

-Takes a long time
-Requires vendor on-boarding to ensure relevant recommendations

Now that you’re well versed in the different types of market research, it’s time to get you off and running! In this guide, you’ll learn everything you need to know to conduct your own market research from start to finish—from creating a project plan to designing your survey to collecting representative responses to turning results into action.

The best practices in this guide span any market research use case, but we’ll focus on the three example surveys you see below: consumer behavior, ad testing, and brand tracking. By the time you’ve finished reading this guide, you’ll be a pro at DIY market research!

Consumer Behavior

A survey measuring millennials’ usage of video streaming services.

View example

Ad Testing

A survey assessing print ad concepts for a fictional dog food company.

View example

Brand Tracking

A survey tracking awareness and perception of sparkling water brands.

View example

02 Planning

How to create a market research plan

Before you get knee-deep in survey design and data collection, it’s important to have a clear plan in place for your market research. Knowing when you need market research, understanding what type of market research is important to your business, aligning with your stakeholders, and scoping out your project ahead of time will ensure you stay on track and deliver actionable results.

When to do market research

So, when’s the best time to do market research? Trick question! Continuously. No matter where you work, you can always tap into the market’s opinions and preferences to be better at your job.

Let’s look at this in the context of a product’s lifecycle:

How market research fits into the product life cycle

How market research fits into the product life cycle
  • Development: When you’re planning to launch a new product, you need to fundamentally understand the market opportunity, consumers’ needs and preferences, how they buy, what competitors are already doing, and what general trends are showing up in the industry. When you get to product ideation, deciding early things like the product name and logo can also benefit from consumer feedback.
  • Introduction: When you get closer to product launch, testing various iterations of your product to make sure you’ve nailed product-market fit is critical. Beyond the product itself, you’ll want to test things like packaging, claims, website messaging, and other early marketing materials. When you get your first customers, they’ll be a gold mine for early-stage feedback so you can iterate.
  • Growth: As you start to grow your product sales and build your brand, market and customer feedback becomes even more important. In the growth stage when customer acquisition is extremely important, doing market research allows you to test and optimize your messaging, keep tabs on your brand perception, and identify ways to improve your product so you keep growing your customer base.
  • Maturity: Product maturity is not a time to rest on your laurels. Keeping a continuous pulse on your brand, competitors, and industry trends when your product is in the mature stage can help you decide where to go next. Maybe it’s time to expand to a new market, refresh your brand, or develop a product line extension.
  • Decline / Extension: When your product sales start to drop, it’s critical to diagnose why. Is it merely seasonality? A new product in the category that’s stealing market share? Market research can help you detect declines early. And if you’re looking to expand your product offerings to offset the decline, the need for market research reappears as your re-enter the development stage.

If the answer to “when should I do market research” is “all the time,” how do companies manage and plan for that? One way is budgeting for large annual studies—like brand tracking and competitive research every year or more frequently. In addition to that, more and more companies have taken a page from the world’s most innovative companies to make their market research more agile.

What is agile market research?

Agile market research is an approach to conducting market research in which projects are structured in small, frequent “sprints” so you can adapt to challenges on the fly.

Rooted in the Agile Methodology first introduced in the software development space, agile market research takes on a lot of the same characteristics you’d find in startup culture. Agile market research goes beyond just being faster. As SurveyMonkey president Tom Hale described in his piece in Quirk’s, “The era of doing a couple of large, set-in-stone projects a year is gone. Today your research goals adapt to the ever-changing needs of the business, which translates into frequent projects that validate your strategy along the way.”

“The era of doing a couple of large, set-in-stone projects a year is gone. Today your research goals adapt to the ever-changing needs of the business, which translates into frequent projects that validate your strategy along the way.”

The movement towards agile market research

Agile market research isn’t a brand new concept, but it has spread like wildfire. Even a couple years ago, 78% of researchers were planning to adopt agile market research methodologies, according to a study by GutCheck. With DIY research tools like SurveyMonkey and SurveyMonkey Audience, more people and teams within an organization are able to do their own market research without having to rely on centralized insights teams or full-service agencies, and this allows them to always have the data they need to make decisions faster.

Below is SurveyMonkey’s agile market research framework. It visualizes a cyclical  approach to exploring, testing, validating, and optimizing strategies—whether they’re ideas, concepts,  campaigns, you name it—all the while continuously tracking progress. It takes what otherwise might be a long, drawn out research timeline and breaks it into sprints of smaller, more manageable projects. The beauty of this framework is its versatility. It shows you how agile market research can help you for nearly any challenge or project you might encounter at your job.

Agile market research framework: explore, test, validate, optimize, and track

Each research phase fundamentally asks the following questions:

  • Explore: What are the needs of the market?
  • Test: Which idea does my market prefer?
  • Validate: How successful will my idea be?
  • Optimize: Which improvements will make my idea even better?
  • Track: How am I doing?

The key takeaway here is that with the right tools and processes, research can be an ongoing activity that any team in your organization can own. Now, that doesn’t mean that you should just start conducting research willy-nilly. Having a clear direction and plan will make frequent research that much more strategic and actionable. The first step is defining the business question your research aims to answer.

Defining the business question and research goal

Starting a market research project from scratch can feel intimidating. If you break it into chunks and have a clear vision of what you’re looking for, things become easier. In order to focus your research, you should lay out the business question and research goals you’re looking to achieve.

The business question is a short summary of the problem you’re solving for and the context of how it fits in to your business. We’re not talking about survey questions here—business questions are high-level goals or challenges that tie directly to your business objectives that can help you make better decisions. A business question might involve:

  • Knowledge gaps: things you don’t know about your industry, competitors, or target buyers
  • Business phenomena: trends you’re seeing in the business (e.g.: dips in sales or increases in churn) that need explaining
  • Predictions: you might just be looking to be one step ahead of your competitors
  • And more…

The research goal is an outline of the specific facts or metrics you hope to learn with your research. In other words, your research goals are what helps you answer your business question, you can map research goals to that question. Writing strong, relevant research goals is important because they will later  translate to specific survey questions later on.

A good rule of thumb is to have no more than 3 research goals so you ensure your survey is focused and not overwhelming for respondents.

To bring this to life, here are some hypothetical business questions and research goals:

Business questionResearch goals
Consumer behavior:
We’re considering investing in a couple video streaming services companies, and we need to understand the existing landscape and perceptions so we invest wisely.

1. Learn what tech brands and apps are most popular among millenials

2. Gather proof points around quantity/satisfaction of apps used

3. Understand millennials’ usage and attitudes towards streaming services
Ad testing:
We’re very close to going to market with our new dog food, and our designers have come up with several great designs for print ads. How do we choose which design to go with?
1. Compare consumer appeal and preference for each ad design

2. Identify which design consumers would be willing to pay more for

3. Assess any differences by consumer demographics
Brand tracking:
We’re an established brand in the sparkling water category, but a lot of new brands have launched in the last year. What does that mean for us?
1. Measure brand awareness for all major brands in the category

2. Assess each brand’s perception and associations

3. Understand brand adoption for our brand and the new entrants

As you go about connecting your business question to your research goals, you should begin thinking about summarizing how your results will be used, what format they’ll be delivered in (a summary in a spreadsheet? a formal presentation to executives?), and who the key stakeholders are for the research (typically those responsible for business decisions). All of this information will be helpful later when you create a brief to help you summarize (or defend) your market research project.

Outlining your project scope

When your objectives are clear, the next step is to outline your project scope. This includes your project budget, who you’ll be surveying and how, and the estimated project timeline.

Know your budget

Okay, so this is a bit of a chicken-or-the-egg scenario. Should the project scope determine the budget or vice versa? The reality is that most teams are given a set budget for the year and need to figure out how to work within it. Earlier in the year (or planning cycle), you might have more unallocated funds to work with, but as the year goes on, things might get tighter. Where you stand with budget will likely determine how you move forward.

Things to consider when figuring out your budget:

  • What is the estimated price of the sample? DIY solutions like SurveyMonkey Audience let you calculate the cost of your responses before creating a survey. Things that affect sample price are: sample size, targeting criteria, and the number of people you expect to qualify for your survey.
  • Will you need any services, or will your team have capacity to do things like survey translations for international studies and data analysis?
  • How frequently will you be running the study? Is this a one-off project or a recurring tracking study? The number of waves will affect your project cost.
  • If you’re using a full-service research vendor, are there any additional agency fees?

Figure out how you’re going to reach your target audience

Knowing who you need to target with your survey and how you plan to do it will have implications for how you resource your project (both dollars and team members working on it). Try to answer the following questions before getting started:

  • Who am I surveying?
  • How many people do I need to survey?
  • Do I already have access to those people?
  • Will I be able to get responses from the people I have access to fast enough?
  • How will I send my survey to these people?
  • Alternatively, would an online panel have the people I need to reach?

Not to worry, we’ll go into much more depth on how to answer these questions in the fielding section of this guide.

Create a project timeline

Understanding how long each step in a market research project takes will ensure you start the work early enough to have actionable insights before you need to make your decisions. Here’s the timeline you can expect for a large scale DIY market research project.

Market research timeline

market research project timeline

Timing considerations for each step:

  • Planning: Depending on the number of stakeholders for your project, it could take a couple meetings to make sure  everyone agrees on the goals of the research.
  • Survey design: This step could be pretty quick if you start with one of our market research survey templates, but might take longer if each stakeholder wants to have input into the survey design.
  • Fielding: The time it takes to collect all of your responses will depend on your target (a broader audience will take less time to field than a more specific audience) and your methodology (an online panel like SurveyMonkey Audience could take as little as a couple hours, but sending a survey via social media or email could take longer as it requires reminders).
  • Analysis: You’ll want to consider how exhaustive you think your analysis will be. If you’re looking to describe the opinions or perceptions of people in your sample, it might mean a few days to create simple charts and graphs, but if you need to run statistical analyses or models, this portion might take longer.
  • Roadshow: Once you have your insights consolidated, showing them to your stakeholders could just be a meeting or two (even an email if you’re moving quickly), or it might require longer discussions if closely tied to a large strategic initiative.
  • Action: The most important step! While there might be individual decision points, it’s tough to assign a timeframe for the action step as it should be ongoing once your results have been shared.

Now, these are timeline estimates for a larger-scale research project with several stakeholders. The beauty of agile market research is that it’s frequent and built into your business process. You might follow the above timeline for the first run, but future projects could be done on the fly in a much more accelerated timeline (perhaps even just days from start to finish).

Documenting your plan in a research brief

Once you’ve worked through everything above, you should document everything in a research brief. Research briefs are a great way to keep everything in one place. A well-written research brief will be able to tell anyone who’s interested exactly what you will be studying (and what you won’t be) to make sure everyone is on the same page. And if anyone wants to suggest something that’s not part of your experiment, your research brief will help you show why you need to stick to the plan.

Here’s an example research brief you can follow:

And with your plan all set, you’re ready to start drafting your survey!

03 Survey design

How to design a successful market research survey

In this section, we’ll teach you everything you need to know about writing an effective market research survey. We’ll start with how to structure your survey, discuss question writing tips, tell you ways to optimize your survey for data quality, provide market research survey templates to get you started, and even share a pre-launch checklist. Let’s get to it!

To start, take a look at the below diagram. It breaks down the anatomy of a good market research survey. We’ll dive into all of these recommendations in more detail in this chapter.

Anatomy of a good market research survey:

market research design

How to structure your market research survey

Before you begin writing survey questions, it’s important to know how to structure a good market research survey. While there isn’t a perfect formula, there are several fundamental things to take note of before you get writing:

  • Go from general to specific. Treat your survey like a conversation with someone you just met. You wouldn’t start a conversation by asking what they think of the packaging on their favorite brand of pet food. You’d warm them up by asking first if they had a dog, what kind it was, what kind of dog food they buy, etc. That’s what we mean by going from general to specific. Start your survey with higher level questions about the category you’re interested in, and then get into specific preferences and opinions.
  • Start simple. Open-ended questions require respondents to think harder. Our research shows that seeing an open-ended question at the very beginning of a survey may make some respondents drop out, lowering your survey’s completion rate. The same goes for lengthy intros and a large matrix/rating scale questions. An easy fix? Begin your survey with a simple multiple choice question to draw respondents into it.

Average completion rate by type of opening question

average survey completion rate by type of opening question
  • Keep screening questions at the beginning. If you are trying to target a specific group of people using screening questions in your survey, it’s best to use them in the beginning of your survey. That way, you’ll disqualify respondents who aren’t relevant for your survey before they get too far.
  • Include your demographic questions last. Things like age and zip code are easy and fast to answer, but there are demographic questions like household income and sexual orientation that are more sensitive. To avoid drops in response rate, build up some trust and engagement with the respondent first (with your survey) before asking them personal demographic questions. Also, this allows you to ask the questions that answer your research goals first, while respondents are still fresh (even a fatigued respondent will still answer their age correctly).
    • Note: if you’re using a demographic question as a screener, it’s still best to include it at the beginning of the survey.
  • Limit the use of page breaks. Page breaks mean extra clicks for your respondents and too many could annoy them. However, there are times when page breaks are necessary (e.g.: when using skip logic or after your screening questions). You can also use page breaks to separate different themes of your survey. In general, you can think about a Goldilocks rule for page breaks: one question per page is way too few; all questions on one page is way too many; 5-10 questions per page is juuuuust right.
  • Hide the “previous” button. In market research surveys, which so often evaluate things like first impressions or awareness of brands, it’s important that people can’t go back and change their first answers.  In the brand awareness survey example, you might be looking for the first brands that come to the respondents’ mind. If on the next page of the survey you have included brand names or imagery, you don’t want that respondent to be able to return back to the awareness question to change their answer. Hiding the “previous” button in your survey can avoid this.
    • Note that this is done automatically when you send your survey using our Audience panel.
  • Know your limitations. Make sure you’re adhering to any structural requirements of the platform or panel you’ll be using. With SurveyMonkey Audience, there is a maximum 50 question limit among a couple other survey design guidelines for ensuring data quality.

Now that we’ve reviewed how to structure your survey,  let’s look at strategies for writing your survey questions in a way that ensures you don’t bias your respondents.

How to write market research survey questions

A lot of things can inadvertently bias your survey respondents—the wording of your question, the sentence structure, how you choose your answer options, and more. Luckily it’s pretty easy to avoid if you adhere to these tips.

1. Follow the golden rules of good survey questions

No matter the question, the golden rules for writing a good survey question are:

  • Mutually exclusive answer options, meaning that there are no overlaps in answer options (e.g. age ranges 18-34 and 34-45 both have 34 in them).
  • Collectively exhaustive answer options, meaning that the answer options cover the full range of possibilities.
Bad question:
How much time do you spend watching TV in a typical week?

• 1-3 hours
• 3-6 hours
• 6-9 hours
• 10-12 hours

This question is bad because the answer options overlap and it doesn’t account for people who watch fewer than 1 hour of TV or more than 12 hours or TV.
Good question:
How much time do you spend watching TV in a typical week?

• I don’t watch TV
• 1-3 hours
• 4-6 hours
• 7-9 hours
• 10 hours or more

This question is good because the answer options don’t overlap and it accounts for people who watch any amount of TV.

2. Balance your answer options

Ordinal scale questions like Likert scale questions are extremely common in market research. They typically come in the form of a 5-, or 7-point scale, which means they include 5 or 7  answer options. The key with these types of questions is balancing the number of positive, neutral, and negative answer options—there should be equal parts positive and negative.

Bad question:
How likely are you to purchase this product?

• Extremely likely
• Likely
• Not likely

This question is bad because two out of the three answer options are positive.
Good question:
How likely are you to purchase this product?

• Extremely likely
• Very likely
• Somewhat likely
• Not so likely
• Not at all likely

This question is good because it uses 5 answer options that are balanced with 2 positive, 1 neutral, and 2 negative answer options.

Also, avoid agree/disagree questions. Survey scientists have found again and again that they cause desirability bias, the tendency for respondents to agree more often than they normally would. We recommend using Likert Scale questions instead.

If you’re having trouble coming up with the right answer options, SurveyMonkey’s Answer Genius can help by using artificial intelligence to predict the answer options you should use for a given question.

SurveyMonkey Answer Genius
SurveyMonkey Answer Genius
SurveyMonkey Answer Genius
Caption pro tip
Pro tip: Don’t trip up your respondents by switching up the question structure on every question. Whether you use 5-point scales or 7-point scales, go positive to negative or negative to positive, choose one question structure and stick to it.

3. Give your respondents a way out

Make sure every respondent who enters your survey is able to answer the questions in your survey—even if the survey question isn’t relevant to them. Including an “other” or “none of the above” or “never” answer option can help you achieve this.

Bad question:
How often do you watch video using streaming services (like Netflix or Hulu)?

• Every day
• Several times a week
• About once a week
• Several times a month
• About once a month
• Several times a year or less frequently

This question is bad because it doesn’t account for people who have never watched video using a streaming service.
Good question:
How often do you watch video using streaming services (like Netflix or Hulu)?

• Every day
• Several times a week
• About once a week
• Several times a month
• About once a month
• Several times a year or less frequently
• I have never watched video

This question is good because it gives respondents who have never watched video using a streaming service an answer option that’s relevant to their experience.

4. Stay away from jargon

Jargon, industry-specific terms, acronyms, and slang have no place in a survey. If your respondents don’t understand what you’re asking you may end up with bad data, so don’t assume they know specific terms. Make your question wording as clear and simple as possible, and if you must, explain what technical terms are with examples (either within the question itself or in a survey introduction).

Bad question:
How often do you use streaming services?

• Every day
• Several times a week
• About once a week
• Several times a month
• About once a month
• Several times a year or less frequently
• I have never used a streaming service

This question is bad because the respondent may not know what a “streaming service” is, and they might interpret it as audio or video or both.
Good question:
How often do you watch video using streaming services (like Netflix or Hulu)?

• Every day
• Several times a week
• About once a week
• Several times a month
• About once a month
• Several times a year or less frequently
• I have never watched video using a streaming service

This question is good because the wording clearly specifies what type of streaming service (video) and gives examples.

5. Avoid double-barreled questions

Be careful of words like “and” and “or” in your survey question. It may mean you’re asking two questions in one (double-barreled questions), which could make the results unusable. Make sure you’re sticking to one topic in your question, or split your question into two parts.

Bad question:
How useful would this app be for getting your daily news and accessing popular media?

• Extremely useful
• Very useful
• Somewhat useful
• Not so useful
• Not at all useful

This question is bad because it asks about BOTH daily news and popular media. Does the respondent’s answer refer to daily news or popular media? Both? You don’t know.
Good question:
How useful would this app be for getting your daily news?

• Extremely useful
• Very useful
• Somewhat useful
• Not so useful
• Not at all useful

How useful would this app be for accessing popular media?

• Extremely useful
• Very useful
• Somewhat useful
• Not so useful
• Not at all useful

These questions are good because they split the two ideas (getting daily news and accessing popular media) into separate questions.

6. Randomize your answer options

You don’t want the way you worded a question or the order of your answer options to influence respondents’ answers. Order bias can be reduced by randomizing your answer options. Note: Don’t randomize Likert scale questions—just make sure the answer options show up in a consistent order (positive to negative or negative to positive). Also, don’t bother randomizing if it makes more sense to arrange answer options alphabetically (e.g. US states and territories).

Bad question:
Which of the following product categories have you purchased in the last 30 days? Select all that apply.

❑ Pet food
❑ Potato chips
❑ Sparkling water
❑ Packaged cookies
❑ Baby food
❑ Ice cream
❑ None of these

This question is bad if you intend to understand usage of pet food because that answer option is at the beginning and subject to order bias and will probably get more responses than the rest..
Good question:
Which of the following product categories have you purchased in the last 30 days? Select all that apply.

❑ Ice cream
❑ Packaged cookies
❑ Pet food
❑ Sparkling water
❑ Baby food
❑ Potato chips
❑ None of these

This question is good because the answer options are randomized for each respondent, reducing order bias and increasing the validity of results.
Caption pro tip
Pro tip: When randomizing your answer options, you can select “Do not randomize the last choice” so your “none” or “other” answer option stays at the bottom. Note: If you have checked the box to include an “Other (please specify)” answer option, that option will automatically be locked at the bottom when randomizing your answer options.

7. Be careful how you phrase your screening questions

Screening questions ensure your respondents actually meet your survey criteria. But savvy online panelists have caught on to the fact that failing a screening question means they get kicked out of the survey (and therefore, miss out on the incentive). You’re more likely to get higher data quality when respondents are unaware you’re using a screening question at the beginning of your survey. Avoid obvious screening questions like yes/no questions because acquiescence bias (tendency to answer “yes” to appear more agreeable) may influence respondents to answer dishonestly.

Bad question:
Have you purchased pet treats in the last 30 days?

• Yes
• No [Disqualified]

This question is bad because it’s clear to respondents that this might be a screening question and influence them to answer “yes”.
Good question:
How often do you purchase pet treats?

• Once a week or more
• Several times a month
• Once a month
• A few times a year or less [Disqualified]
• Never [Disqualified]

This question is good because it’s not immediately clear that this is a screening question.
Caption pro tip
Pro tip: make sure your screening question logic is sound before launching your survey. There are nuances (especially with checkbox question types) that you’ll want to preview and test to validate that your logic is working correctly.

Tips for optimizing your survey for data quality

Data quality can be attributed to many things, but optimizing your survey design is one of the most important (and controllable!) ways to ensure reliable results. Here’s what you should and shouldn’t do:


DO optimize your survey for mobile devices. On the SurveyMonkey platform, roughly 40% of surveys are being taken on a mobile device in the US, and that percentage is even higher internationally. Simplify your survey questions. Keep your survey short. Remove logos and progress bars that take up valuable screen space.

DO use survey logic to make surveys more relevant for respondents. This means using skip logic to make sure they aren’t answering questions they can’t answer and disqualification logic to make sure the right people are taking your survey. Double checking your survey logic can be a challenge, so we built a logic preview tool to make it easy for you.

DO make your questions required. If your questions aren’t mandatory, you might end up with an incomplete data set, which not ideal when you’re paying for responses. We recommend requiring answers to all of your questions in most cases. Note: if you’re using a matrix question, make sure you require all rows in the question.

DO turn off survey & page titles before launching your survey, especially if you think they will distract or bias your respondents. For example, if your brand name is in the survey title, you probably don’t want people to read it before they answer a brand awareness survey.

DO keep it interesting. Remember that there are humans on the other end of your survey. Be engaging! And if your survey is about a dry topic, try using colloquial language (but not slang!) as much as possible. You’ll get higher response rates and better data quality when respondents find the survey fun to take.


DON’T ask too many questions. Not surprisingly, our research has shown that longer surveys mean lower survey completion rates. We recommend sticking to 30 questions or less.

DON’T include too many answer options. 8-10 answer options is a good maximum to keep in mind for multiple choice or checkbox questions. Any more and people stop reading and are less likely to give a thoughtful answer. Try to stay under 6 for ranking questions. And stick to a maximum of 5 rows by 5 columns maximum for matrix questions.

DON’T include video content longer than 90 seconds. We’ve seen that multiple videos and videos longer than 60-90 seconds lead to lower completion rates.

Pro tip: turn off our One Question At A Time feature when using media like images or video in your market research survey. This allows easier scrolling up and down so respondents can revisit the media when answering follow-up questions.

DON’T edit or change your survey when you’re in the middle of collecting responses. Responses collected before and after you make your edits might not be comparable.

If you’ve read this far and you still don’t know where to begin, we’ve got 20+ market research survey templates waiting for you!

Market research survey templates

We have several market research survey templates to help you get started. All you have to do is customize the survey template for your product/service category and you’re good to go!


Market sizing

Quantify the demand for your product, and pinpoint which market to enter next.

See template

Competitive intelligence

Measure awareness, usage, and NPS® for the key players in your category.

See template

Category awareness

Gauge the levels of awareness and adoption for emerging product categories.

See template

Consumer behavior

Understand your target market’s behaviors, perceptions, and needs.

See template

Shopper insights

Learn who, where, and how people shop for your product/service category.

See template

Path to purchase

Deep dive into the purchase decision making process of your target buyers.

See template
Product development

Product testing

Get feedback on your product concepts to learn which one will win in-market.

See template

Package testing

Test to see which packaging designs and messaging will stand out most with buyers.

See template

Pricing research

Know what consumers are willing to pay for your product and what’s too expensive.

See template

Name testing

Test company, brand, or product name ideas before launching.

See template

Marketing effectiveness

Ad testing

Test your advertising campaigns to know which will drive higher purchase intent.

See template

Logo testing

Get feedback on your logo ideas to find out which is most appealing to buyers.

See template

Message testing

See which marketing messages or claims resonate most with your target buyer.

See template

Brand awareness

Learn how many people are aware of your brand plus how and when they heard about it.

See template

Brand conversion

Capture the entire brand funnel from awareness to consideration to loyalty.

See template

Brand performance

See what’s most important when making a purchase and how your brand performs.

See template

Can’t find what you’re looking for? See all market research survey templates, or check out our question bank when you create your next survey.

And if you’re planning to do a lot of research, you can create your own shared library of internal survey templates and a custom question bank so the research across your organization is consistent over time.

Things to check before you launch

Here’s a pre-launch checklist for you to follow so you can make sure your market research survey is ready to send!

❑ Review your market research brief and make sure your survey appropriately addresses your business question and research goals.

Preview and test your survey so you can experience it the way a respondent would.

❑ Make sure your survey works well for different device types, especially mobile.

❑ Double check your survey logic and question validation using our logic preview feature.

❑ Put yourself in the shoes of someone you’re NOT targeting to make sure all questions are answerable and/or you have the appropriate disqualification logic.

❑ Double check the structural components of your survey—navigation buttons, survey and page titles, required questions, etc.

❑ Make sure the question text matches the question type. For example, if you want respondents to select all that apply, make sure you’re using the checkboxes question type.

❑ Double check that your media (e.g.: videos) are accessible when not logged into your media account.

❑ Give your survey a spelling/grammar once-over.

❑ If you’re unfamiliar with how the results will show up, try creating a weblink and take your own survey a couple times to make sure the results show up the way you expect. Don’t forget to delete your test responses before you analyze your data.

❑ Get a second pair of eyes—sometimes that’ll uncover anything you’ve missed!

Once you’re satisfied with your survey design, the next step is to collect survey responses!

04 Fielding

How to collect survey responses from the right people

Now that your survey is ready to launch (or as we say, get into “the field”), it’s time to figure out who you’ll target, how you’ll reach them, how many responses you’ll need, and when to send your survey.

Determining who you should target

When you’re sending out a market research survey, your ultimate goal is to understand the behaviors and perceptions of the target population you’re interested in. In order to do that, you need to know who your target audience is to begin with.

Here are some example characteristics that you could use to define your target audience:

  • Demographics: location, age, gender, education, household income, race, marital status, parental status
  • Employment & firmographics: employment status, job function, job level, industry, company size
  • Shopping habits: stores shopped at recently, online shopping frequency, restaurants visited recently, likelihood to make a ___ purchase in the next 12 months
  • Behavioral attributes: mobile device/app usage, pet ownership, exercise frequency, dietary restrictions, hobbies
SurveyMonkey Audience: SurveyMonkey's global panel for DIY market research

Now, if you’re a well-established company, you probably already have a good sense of who your target market is and the attributes that define it. But if you don’t that’s okay too. With market research there are times when you want to target a broad audience and times when you want to target a narrower group of people.

Examples of when to use broad survey targeting:

  • When you’re not sure who your target market is and are looking to define and segment the population who is interested in your product/service.
  • When your product/service category appeals to the general population. For example, some household products, foods, or mobile phones could be purchased by any gender, of any age, in any geography.
  • When you are measuring something at the top of the marketing funnel, like brand awareness and favorability. More people know about brands than purchase them, so targeting a broader audience to understand the spectrum of awareness across demographics could help you uncover deeper insights.

Examples of when to use narrow survey targeting:

  • When your product/service category is only relevant to a subset of the population. For example, if you’re assessing the appeal of a dog food package design, you’ll likely want to target dog owners. If you want to know what features to prioritize in a marketing automation software, best to survey marketers.
  • When you want to ask location-specific questions. For example, you might want to assess the impact of an ad in a specific geography or DMA (designated market area). If you’re only running the ad on the east coast of the U.S., no sense surveying the entire country to measure ad awareness and recall.
  • When you’re measuring something at the bottom of the marketing funnel, like purchase intent, satisfaction, or loyalty. In this case, you probably want to target category purchasers, or people who have already used your product (or perhaps even your competitor’s product).

How to reach your ideal survey respondents

How you’re going to reach the people you need to survey depends on who you’re surveying. Rest assured there are a variety of options for getting your survey in front of the right people.

Existing contacts

Existing contacts could be your friends, family, coworkers, customers, website visitors, social followers - anyone you already have access to. In most cases, you have some way of contacting them to invite them to take your survey.

Common ways of surveying existing contacts:

A sample of your target market

With most market research surveys, what you need to do is obtain a representative sample of your target market, meaning a group of people that accurately mirrors the population you’re interested in.

Market research panels, like SurveyMonkey Audience, offer a way to access people you wouldn’t otherwise be able to reach.

Survey panels will also do a lot of the work for you, like recruiting a diverse population to join the panel, inviting the right people to take your survey, incentivizing respondents, and monitoring your completion rate.

Pros and cons of different respondent sources

Existing contacts (customers, followers, etc.)Representative sample of your target market
-It’s low cost since all you have to pay for is the survey incentive, if applicable.
-Respondents are easy to reach since you’ve already got their info or a way to contact them.
-It’s easy to get a balanced, representative sample that reduces bias in results.
-You can get more granular and specific with your survey targeting.
-Panels make inaccessible people accessible.
-It’s fast. You can collect 1,000 responses in a couple days versus the weeks it might take to get responses on your own.
-Existing contacts make an inherently biased sample because they usually already know and like you, so it’s less likely they’ll answer your survey objectively.
-You need to have a lot of contacts to survey in order receive enough survey responses.
-It can take a long time to reach a high sample size. A couple reminders are usually necessary.
-It takes more work on your part to monitor the survey progress, manage reminders, and send incentives.
-Buying responses costs money. DIY options like SurveyMonkey Audience are usually cheaper.
-Finding a niche sample can be challenging for some targets. For example, cardiologists may be hard to reach with an online panel.

Ensuring representation

If you’re surveying the general population using a survey panel, you’ll still want to make sure that your sample is representative of the population in the location you’re surveying. For example, if you want to survey adults in the U.S., the breakdown of certain demographics like age and gender in your survey sample should match the breakdown of the U.S. population. The U.S. Census is a great source of this information—see the table below.

Once you know what the breakdown of your sample should be, you can specify the balancing criteria for your sample. Balancing refers to the distribution of demographic buckets (e.g.: age, gender) in your survey sample and is specified when you target your audience. SurveyMonkey Audience allows Census balancing as the default, but you can also use our custom balancing capabilities to set your own demographic distribution. This way, your sample won’t skew to one age or gender and throw off your results.

Targeting a specific group of people

Now, let’s say you don’t want to target the general population, but you have a more specific target in mind. In our dog food example, let’s say the product is specifically for small dogs. If you think about the U.S., pet owners are a subset of the adult population, dog owners are a subset of pet owners, and small dog owners are a subset of all dog owners. If you want to understand the packaging appeal of your product, it makes sense to target this relatively small population.

small dog owners

There are two ways of targeting a narrow population like small dog owners.

1. Choose from pre-profiled targeting options. With SurveyMonkey Audience, we have already asked our panelists questions typically used for targeting a survey sample:

  • Demographics: age, gender, household income, etc.
  • Firmographics: industry, job title, job level, etc.
  • Behavioral attributes: devices used, apps downloaded, etc.
  • Shopping habits: primary grocery shoppers, stores shopped at, etc.

Check out the targeting options we have available for your next market research project.

Note: Don’t forget to double check the balancing options. Even when you’re targeting a narrower group of people, you want to ensure the proper demographic balancing. That way, you won’t end up with only female small dog owners instead of a balance of male and female small dog owners.

2. Using custom screening questions. When the available targeting options don’t get you the exact target audience you’re looking for, you can write your own screening questions to disqualify people you don’t need. For example, if we only had a targeting option for dog owners, but you wanted to survey small dog owners, you could use a screening question to disqualify dog owners whose dog is over a certain weight.

survey screening questions

When you use screening questions in your survey, you’ll need to estimate the associated qualification rate (also sometimes called incidence rate), which is the percentage of respondents who make it past your screening questions and go on to take the rest of your survey. This will help us calculate the total number of survey invites to send so you get the total number of completed responses you ordered.

It’s important to understand how targeting options and screening questions work together. For example, you could use targeting options to specify dog owners, and then use a screening question to weed out large dog owners. That way, only small dog owners take your survey. When estimating your qualification rate in this case, it would be the percentage of dog owners (who you targeted with the targeting options) that you think own small dogs (who will qualify past your screening question).

Find a targeted group of people to survey
When surveying your customers or followers isn't right for your survey topic, use SurveyMonkey Audience to find the people you need.
Learn more

How many people to survey

Once you know who to send your survey to, the next question is how many people should you be surveying? The sample size of your survey will have an effect on your ability to analyze and draw conclusions from your results, so getting enough responses is key to doing effective market research.

When choosing a sample size for your market research survey, here’s what you need to consider:

Use case

Are you doing a quick gut-check, making a strategic business decision, or trying to get your results published? How you’ll be using your results can help determine your sample size. Here are some sample size recommendations:

survey sample size recommendations

Margin of error

Margin of Error is an indicator of how closely the  survey results from your sample represent the entire population you’re targeting. Essentially, it tells you how confident you can be in your survey results. Here’s how to interpret margin of error: Let’s say you surveyed 400 people, and 50% are aware of your brand. With a margin of error of +/- 5%, you can be confident that between 45% and 55% of the target population is aware of your brand.

If you want a tighter margin of error, you’ll want to survey a larger sample size.

Factors that Influence Margin of Error:

  • Population size: the size of the entire group you’re interested in (e.g. target market).
  • Confidence level: tells you how reliable a measure is. The market research standard is 95%.
  • Sample size: the number of respondents that you surveyed

Margin of error example

margin of error by survey sample size
Margin of Error calculations are based on a population size of 100,000+ and a 95% confidence level.
Dashed line
For more information, see our Margin of Error calculator


If you’re looking for an extremely specific type of respondent, the panel you’re working with might only have so many in their inventory. The feasibility of your sample can determine the max number of people you can survey at a given time. To get around this, you could loosen your targeting criteria and ask extra questions that can help you identify the respondents you’re looking for. Then, once you’ve collected your responses, you can use  filters in your data analysis to see how those respondents answered.. That way, you’ll be working with more data at the total level, but still have a way to see how that specific group answered your questions.


Let’s be realistic: We all have to work within a budget. The good news about DIY market research and online panels is that they are much less expensive than traditional full-service vendors. Even still, your budget will play a role in the number of responses you’re able to collect. But remember, even if you can’t reach the sample size you were hoping for, some data is always better than no data at all.

Survey logic

Sometimes your survey logic will influence the number of people you need to survey. Let’s say you’re testing 2 different ad campaigns and you used our A/B test functionality (or a monadic survey design) so that 50% of your respondents see one campaign and 50% of respondents see the other. If you want 200 responses for each ad campaign, you’ll need 400 total responses.

When to launch your survey

Is there an ideal time to launch your market research project? Not really. Neither time of day, day of the week, seasons, nor holidays really make a difference for data quality. It’s best to time your project based mainly on your project goals and research plan. However, there are some implications for response volume and data consistency (if you’re running multiple projects or a tracking study) that you should keep in mind.

Response volume considerations:

  • More people take surveys during the day than at night.
  • More people take surveys on weekdays than on weekends.
  • On weekdays, the response volume on desktops/laptops is higher than mobile devices during the day. At night, it switches, and you see more responses on mobile devices than desktops/laptops.
survey response volume by time of day for weekdays and weekends

What does this mean for you? One thing you might want to consider sending dense or image-heavy surveys on a Friday night because it’s more likely people will be taking the survey on their mobile devices.

Data consistency considerations:

  • When running a tracking study, data consistency is key. Since the main thing you’re trying to measure is how results change over time, you want to avoid any other confounding factors that might impact the difference in results between waves. Try to launch your survey at the same time and day of the week if possible (e.g. Tuesday mornings). And if you’re doing quarterly or annual surveys, keep the months consistent (e.g.: first month of the quarter), and take note of national holidays.
  • Seasonality can affect some brands more than others, and you might want to take that into consideration when running market research studies. Let’s say you work for a mountain resort that has winter ski/snowboard clientele and summer camping clientele. Customer and market surveys should be suited for the season you’re in. And comparing or aggregating results from summer and winter surveys should be done with caution.

Monitoring your survey in the field

You did it! You launched your survey! Now, what? If you’re using our Audience panel, all you have to do is sit back and watch the results roll in automatically. If you are surveying your own contacts, there’s a little more to do to make sure you get all of the responses you need.

Keep an eye on the results coming in

You’ll want to monitor your survey as it’s gathering responses so you can (a) make sure you get results on time and (b) detect any potential issues with your survey. There are three metrics you should be aware of:

  • Response rate: the percentage of people who respond to your survey. This is calculated by dividing the number of people who complete your survey by the number of people you invited. Response rate can be affected by things like email open & click-through rates, as well as survey engagement.
  • Completion rate: the percentage of people who complete your survey of those who started it.

A low completion rate indicates something may be wrong with your survey because few people who begin the survey actually finish it.

  • Qualification rate (a.k.a.: incidence rate): the percentage of respondents who make it through your screening questions and go on to take your survey. When using an online panel, if you overestimate your qualification rate (meaning more respondents get disqualified than you anticipated), you may receive fewer responses than you ordered.

Note: when you use SurveyMonkey Audience to collect targeted responses, our system will monitor your completion rate. If it gets too low, we may pause your project and have one of our survey experts help you get it back on track.

It may be DIY, but you're not alone!
Our team of Audience experts are here to help you get your project off the ground.
Contact sales

Remember to send reminders

When surveying your own contacts, it’s rare to get all of the responses you need with just one invite, but that first invitation is extremely important. Our research shows that about ¾ of your responses will come in the first 24 hours of sending your survey via email. Reminders are a good way to boost your response rates. Our recommendation:

  • How many? 1-3 reminders should do it.
  • When? 48 to 72 hours after your initial survey invite is the best time to send reminders

Offering incentives

Offering a survey incentive is another way to improve your response rates. Whether you offer an incentive depends on your budget, but when you use a survey panel, the respondent incentives are included in the price of the responses. Check out SurveyMonkey’s integrations with digital gift companies—an easy way to offer incentives for your survey.

When to close your survey

When you’ve collected all of the responses you need, it’s a good idea to close your survey. That way, once you start your analysis, you’ll have a static dataset without responses trickling in here and there. If you’ve got an always-on survey, like a survey embedded in your website, then you don’t ever want to close your survey.

Note: If you’re using our Audience panel, we’ll close your survey after you’ve received all of the results you ordered.

05 Analysis

How to make sense of your market research results

Hooray! The results are in! Now it’s time to figure out what the results mean so you can turn your in-depth data into actionable insights.


How to prepare for data analysis

Before you start diving into your results, you’ll want to make sure that you’re working with a complete, clean dataset. If you do these things first, you’ll have even more confidence in the findings you deliver to your stakeholders.

Wait for all of your responses to come in

Trust us, we know it’s tempting! When results start flowing into your survey, your first instinct is to start combing through the early data, but be careful! Even if you’ve ordered a balanced sample from a panel like SurveyMonkey Audience, the results aren’t necessarily arriving in a balanced order.

For example, if you launched the survey early in the morning on the east coast, you’ll likely get a lot of responses from the east coast before the west coast. We also know that older female demographic buckets tend to fill up before younger male demographic buckets. So, the conclusion here is to wait until you’ve received all of the responses you ordered and your survey is closed to begin your analysis.

Clean up your dataset

Once all of your results are in, it’s time for a quick cleanup to make sure that you’re only analyzing high-quality survey responses. Essentially, what this translates to is deleting or filtering out responses from people who don’t meet your criteria or haven’t thoughtfully answered your survey.

We get this question a lot: “But aren’t I skewing my data by removing responses?” The answer is that if the responses are low quality (i.e. if you’re certain the respondent didn’t take the time to answer your questions carefully), then removing them will improve your data’s reliability. Even full-service research firms do this before delivering results to clients, so if you’re doing your own market research, you should expect to do a little data cleaning.

A few of the top things to look for:

  • Incomplete responses.
  • Responses that fall outside of your targeting criteria.
  • Speeders: respondents who completed your survey way too fast.
  • Unrealistic answers. 5,000 hours of TV per week? Yeah, right.
  • Gibberish in your open-ended responses.

Note: If you made your open-ended questions required, it’s possible that the respondent didn’t have an answer for the question. Write-ins like “none”, “n/a” or misspellings aren’t a sign of poor data quality.

Make sure your sample represents your target population

Now, if you specified census or custom balancing when purchasing responses, representation probably isn’t an issue since major demographic buckets like age and gender will have the distribution you ordered.

If you sent your survey to your own contacts via email or social media, it’s possible your data isn’t balanced. Check your sample’s demographic breakdown compared to the population you’re interested in. If they’re similar, you’re good to go.

If your sample’s demographics do not match the population’s you might consider weighting your results. Weighting is an adjustment technique used after data has been collected to make sure the demographic profile of your respondents matches your population. Weighting involves custom calculations with a statistical software to fix any lopsidedness in your sample by giving more weight to underrepresented groups and less weight to overrepresented groups.

Have a plan of attack

If you’re a data nerd like us, it’s easy to get sucked into the data abyss and lose all sense of time and space. Having an analysis plan helps you avoid this. Now is the time to revisit your business questions and research goals. At this point, you should know exactly how you’re going to answer your main questions, so your analysis can be focused and efficient.

And now, with your analysis plan at the ready, data spick-and-span, you’re ready to dive into your results!

Tips for making sense of your results

How you summarize and visualize your results is important for understanding and communicating your findings to stakeholders. Below are some guidelines for how to make it easy to spot those valuable insights.

Visualizing your results

We did a study on research-backed content consumption and saw that 42% of people found data visualized with charts, graphs or infographics more enjoyable to read than data written out in a sentence or presented in a table. Here are the most common chart types—from bar graphs to word clouds—and when to use them:

choosing the appropriate chart type for your survey data analysis

Knowing what chart type to use is the first step. Next, surround your chart with the right details and context so people can understand what they’re looking at. Here’s what to include in any chart you’re presenting:

  • Chart title, or survey question text
  • Axis labels, or answer option text
  • Data labels, as either values or percentages
  • Sample size
  • Dates where applicable, especially for trended charts

This is an example of a chart that’s presentation-ready from a brand awareness survey:

Aided brand awareness: sparkling water category

Summarizing Likert scale question results with Top 2 Box Scores

If you have any Likert scale questions in your market research survey, Top 2 Box Scores combine the highest 2 responses of the scale into one easy-to-interpret metric. Consolidating your results in this way makes it easier to:

  • Report overall positive sentiment on a question
  • Compare the results of multiple Likert scale questions
  • Monitor trends over time

Top 2 Box Score

Top-2-box scores
Caption pro tip
Pro tip: if you’re testing concepts like ads, logos, products, or packaging, Top 2 Box Scores are a great way to summarize your results across concepts into a scorecard. Read about how to do this in our ultimate guide to concept testing.
You might also like...
How to use a Top 2 Box Score in your survey analysis.
Read on

Segmenting your results for deeper insights

When you first get your survey results, you’ll likely looking at aggregated answers for the entire sample you collected. Looking at how individual segments of your population answer your survey is one way to uncover insights that could be critical to your analysis.

Here are a few segments of your sample you could look into:

  • Demographic segments, like gender and age groups
  • Geographic segments—countries, regions, states, etc.
  • Behavioral segments like frequent category purchasers

Once you know what segments could be interesting to dive into, there are two ways to approach segmenting your results for deeper insights:

  • Filtering your results. Let’s say you surveyed a broad population, but just wanted to understand how young females responded. Filtering your data allows you to do this. You might find that the answers from young females differ from the total population, which  might make you choose to go in a different business direction for that target.
  • Making comparisons. The benefits of looking at individual segments within your results multiply when you compare them to other segments or to your data as a whole. Take this example from our sparkling water brand awareness study. Across brands, awareness was higher for women than men. Brands like La Croix and Bubly saw larger differences by gender than brands like Schweppes.

Important note: When segmenting your data, be aware of the base sizes you end up with. If you only surveyed 300 people total, segmenting your data by US state would leave you with a handful of responses from each state (at best). This would make it much harder to draw significant conclusions from those responses.

How to analyze data trends over time

Collecting data over time can be immensely useful to a business. Tracking things like market trends, consumer perceptions, brand awareness, and competitive intelligence can provide valuable indicators that companies need to adapt to.

Start with a baseline

A baseline is your first set of results. Whether that’s your first wave of a tracking survey, or the first month of an always-on survey, the baseline is what you’ll be benchmarking against as you collect more data. Once you have a baseline, the next time you run your survey, you’ll be able to understand if things have changed or stayed the same.

Collect data over time

The key to understanding market trends is to consistently collect data over time. But the types of projects you’re running can make a big difference in how you collect that data:

  • Set up a survey in a place that gets consistent traffic, like on your website or in an always-on triggered email, to continuously survey the market.
  • Repeatedly send surveys to a sample population at specific intervals (we call those surveys “waves”). Trackers are commonly run in waves. The frequency of your survey waves depends on several factors: how fast your category is moving, how easy it is to reach your target market, and your budget.

In our sparkling water brand tracking study, we decided on quarterly waves. Here’s what the awareness results look like from the first two waves.


A couple things to consider when doing a trended analysis:

  • Make sure everything stays consistent in each wave of your study. That means the targeting criteria, balancing and other sample specifications should be the same in each wave. Essentially, the only thing you do want to be different is the date the samples were collected.
  • If you’re working with an always-on survey, you’ll want to make sure the way you split up your time periods consistently, i.e. months, quarters, years.
  • It’s natural to see differences in results across waves in a tracker and draw immediate conclusions, but it’s critical to understand whether those differences are statistically significant before making strategic decisions. It’s critical to understand what statistical significance is, how to calculate it, and why it plays such an important role in analyzing data trends.

Statistical significance, explained

What is statistical significance

Statistical significance shows whether one group's answers are substantially different from another group's answers by using statistical testing. When a difference between sample groups is statistically significant, it means you can be confident your results represent a real characteristic of the population instead of random variation in your sample. Statistical significance is useful in market research when applied to differences between respondent groups (e.g.: male vs female) or differences between waves of a tracking study.

How to tell if differences are significant?

A practical rule of thumb is to check if the change between one group and the other is outside the margin of error based on the sample size. If the change is greater than the margin of error, it’s likely a statistically significant change.

You can check your margin of error using SurveyMonkey’s margin of error calculator, or use the table below as a guide. As a general rule, the margin of error gets smaller as the sample gets bigger.

Margin of error by survey sample size
Margin of Error calculations are based on a population size of 100,000+ and a 95% confidence level.

Here’s an example: Let’s say you surveyed 400 people, and find that your brand awareness is 50% in Wave 1 and 53% in Wave 2. This change of +3% is within the margin of error of +/- 5%, which means we cannot conclude that the brand awareness has grown between the two waves.

Margin of error example

example margin of error

One thing to remember—while Statistical Significance can tell you what is statistically different, it does not necessarily mean that the difference is meaningful for your business. It’s up to you as the interpreter of the results to determine if that statistical significance is important.

Statistical testing

Beyond using the sample size and margin of error to estimate whether results are significantly different, you can use statistical testing (a.k.a.: stat testing) to know for sure. Statistical testing is commonly built right into survey software. In SurveyMonkey, you can use stat testing with comparisons and cross-tab reports.

Alternatively, you can use calculators to do your own stat testing on groups of results.

Citing your research

Once you complete your analysis and are ready to publish your research, it’s a good idea to  write a methodology summary—whether it’s in a presentation or a research report. A methodology summary explains where the results came from so anyone reading your analysis has the right context.

A survey methodology summary should contain:

  • The data collection method (i.e. SurveyMonkey Audience, customers, etc.)
  • When you fielded your survey (date of launch - date of completion)
  • Sample size
  • Country/location
  • Age range
  • Whether the sample was balanced or weighted (and if it was, based on what population metrics)

Here’s an example:

This study was fielded using SurveyMonkey Audience from June 20-June 22, 2019 with a U.S. sample of 1,013 adults age 18+. The sample was balanced on age and gender according to the 2010 US Census.

06 Taking action

Transform insights into strategic business recommendations

The results are in, you have your findings, and now you need to figure out what to do next. Or, maybe it’s clear what you should do, but you need to convince the right people. We’ve got a few suggestions for how you can go beyond the insights to craft a story, develop clear recommendations, and get buy-in from leadership.

Finding the story in your data

Think about some of the best presentations you’ve seen. What makes them stand out? Usually the answer is a memorable story. If you can package up your findings with a clear, logical story, you’re more likely to make an impression and get buy-in for your recommendations. Here are six things to keep in mind as you craft your data story:

6 tips for telling a compelling story with data

1. Set up the story

By now you’ve become very intimate with your research, but, of course, not everyone you’ll present it to will be. It’s important to set up the story to ensure everyone is caught up before you present your findings and recommendations. Storytelling frameworks can help you come up with an outline for your presentation.

One framework you can follow is SCQA, first introduced by Barbara Minto in her book, The Pyramid Principle: Logic and Writing in Thinking. The SCQA acronym stands for: Situation, Complication, Question & Answer, and it’s commonly used by consulting firms.

  • Situation: where the business is today, what the knowns are, and any other relevant context.
  • Complication: the business problem that needs solving - why you set out to do market research.
  • Question: the specific questions that needed answering with your research, and your approach to answering them.
  • Answer: the main insights coming out of your market research that answer your business questions, and how they point to action to solve your complication.
SCQA: Situation, Complication, Question, Answer

2. Layer in business context

If you follow the SCQA framework, this will come naturally. When presenting your research to stakeholders, business context helps you answer the “so what?” question—why your research matters, why stakeholders should care, and how the findings/recommendations can help the business.

3. Find powerful stats

No one’s going to be overly impressed if you’re throwing up percentages in the single digits. The powerful stats are the ones that support your claim with large numbers. This is because once you exceed the 50%-mark, you’re talking about the majority.

If the data that supports your story isn’t a big statistic, try reframing it. Instead of “10% of Americans would feel safe as a passenger in a self-driving car”, try “90% of Americans would not feel safe…” Note that in cases where you’re conducting a structured analysis (e.g.: building a scorecard), reframing stats won’t work.

4. Focus on the “why”

In many cases, your existing business data will describe the “what” but when you collect consumers voices and opinions using market research, you can uncover the “why”. When your story is falling flat, step back and see if your presentation is not just answering the “what” and “how”, but also the “why”.

5. Keep it simple

You’ve done a lot of digging into your data, and while a lot of the findings may be interesting, they might not all be relevant to your story. Only include the results that contribute to or add color to your story and recommendations. It’s OK to leave a lot of your data out of your presentation if it distracts your audience from taking away what matters.

6. Make it human

At the end of the day, the goal of market research is to explain human perceptions and behaviors. The more you can bring real examples into your story, the more tangible the results will be to your audience.

One way to do this is by peppering in quotes from your open-ended responses. You can also mix qualitative information (e.g.: interviews, customer support cases, etc.) with your quantitative results. This will bring your data to life and make your points that much more compelling.

Turning insights into strategic business recommendations

Strategic business recommendations take your findings from “Oh, that’s really interesting” to “Now what?” But, in order to get people on board, you have to make sure your recommendations are realistic and aligned to the overall business strategy.

Strategic recommendations should be:

  • Focused: Start by referring back to your research brief. Recommendations should flow from the business question you started with and be supported by your research insights. Also, keep your recommendations focused on the audience you’re presenting to (e.g.: marketing recommendations to the marketing team).
  • Specific: A vague recommendation will never be implemented. Make sure your recommendations outline specific, clear actions to be taken. It’s possible that the next step is to do more research, so in that case, be ready with a plan (and budget, if applicable).
  • Attainable: If you’re recommendations are unrealistic, you’ll lose your audience to eye-rolls. If they are aggressive or require more resources than you currently have (whether it’s budget or headcount), outline what it would take to accomplish your recommendations.
  • Measurable: Tie your recommendations to quantifiable business outcomes, or better yet, forecast the business impact of implementing your recommendations. Even if they’re based on assumptions, including metrics in your recommendations will help you sway your leadership team.

Inspiring action from your stakeholders

This is the moment you’ve waited for! Your time to shine! How you’re going to leave your mark and impact your organization for the better!

Now, getting buy-in for your recommendations and motivating others to act can be a challenge. Two strategies that we’ve seen work well for our clients are 1) Going on a research roadshow and 2) Creating a follow-up plan.

1. Go on a research roadshow

The research roadshow is pretty much what it sounds like: a series of meetings to present your findings and recommendations to the right stakeholders to motivate them to take action. The order in which you do things here can make a big difference in how your roadshow turns out:

  1. First, preview your results with your immediate team to get support and feedback. Think of this as  a dry-run before presenting to cross-functional stakeholders and leadership. A key benefit to doing this is that you can get feedback on the story your presentation tells and recommendations for how to improve it. You want people on your team to be on the same page and telling the same story as the research gets shared.
  2. Involve cross-functional stakeholders. Instead of rounding them all up into one room s, engaging them in small groups or even 1:1 previews of the findings can open them up to feedback and involvement in developing the recommendations. They’ll have a great sense of what’s achievable in their department. A huge bonus of this step: If stakeholders feel like they’re a part of creating the recommendations, they are more likely to act on them.
  3. Once you have made your rounds and all your  cross-functional partners are on the same page, present an executive summary to leadership. The key here is to know your audience and tailor the message and story to the things they care about. While in some cases this step might not be necessary to make final decisions, it’s a great way to get executive exposure for your work and team’s strategy.

2. Create a follow-up plan

One of the best ways to ensure your recommendations are implemented is to hold your team and stakeholders accountable with follow-up. Make a plan to regroup with the team as needed to check in on action items. Ask yourself and your stakeholders these questions:

  • What actions were taken?
  • What was the business outcome?
  • If no actions were taken, why not?
  • How could our market research be more actionable next time?

Finally, pull your stakeholders into your follow-up research plan. As you continue to do more market research, keeping stakeholders involved will improve the effectiveness and actionability of your research.

And now! Go forth! Use what you learned in this guide to get the market feedback that will catapult your business to the next level. And remember: if you get stuck, our team is only an email away!

Ready to get started with your market research survey?

Design your survey

Create a survey from scratch using our best practices or start from one of our market research survey templates.

Find your audience

Reach a targeted sample of people in over 130 countries using SurveyMonkey Audience. You’ll start seeing responses in minutes.

Thanks for reading! Check out our other Ultimate Guides!

Concept Testing

Get feedback on product ideas, packaging, marketing campaigns, and more. Learn how to run your own concept testing program.

See guide

Content Marketing

Fuel your content strategy with market research. Learn how to build credibility and thought leadership in your industry with surveys.

See guide

Customer Feedback

Act on customer feedback at scale. Learn how to run a customer feedback program from team organization, survey design, and analysis.

See guide