You’ve thoughtfully crafted a survey, sent it to target consumers and received thousands of responses. Now what? Survey analysis.
That unstructured mass of customer data is useless if not analyzed properly. Survey analysis can be approached from a few different angles, depending on the type of survey data and your analysis goals.
Know your goals
Understanding the intended outcome of any research project is the absolute first step in survey construction and survey analysis, so you need to think about the following:
- What do you want to understand?
- What do you want to uncover?
- How do you plan to utilize research insights?
These three points will shape the questions you choose to ask, but they will also guide your course of analysis. When dealing with large numbers of words, it can be easy to get lost in the data. Keeping in mind your research goals allows you to sift through the survey responses to find the most relevant information.
The second part of this process is exploratory survey analysis. Approaching research with a hypothesis keeps analysis focused and ensures that the insights are relevant to your brand. Non-hypothesis based analysis leaves room for the ‘unknown unknowns’ and the information you aren’t expecting to find. Often the most valuable insights are the ones you don’t even know to look for.
The last step is essential in ensuring that the hours and hours of work you put into creating and analyzing this survey aren’t wasted. The final stage in any consumer research project is utilizing the research insights. Prior to analysis, understand how you can allocate budget, time and other resources to applying research insights to business practices.
Understand survey performance
- What was your response rate?
- What questions produced the highest drop-off rate?
- How did respondents from different demographics or channels (social, email, etc) perform?
Analyzing the performance of a survey can inform future research efforts. It’s helpful to know which questions, question formats and distribution channels produced the best results. Moving forward, surveys can be created with these insights in mind.
Survey performance and response rate can also impact the legitimacy of insights. A sound survey requires large numbers of respondents to avoid bias. It’s important to be aware of response rate to determine if survey data can be used.
Creating a survey
The only way to collect valuable feedback and conduct useful survey analysis is by asking the right questions. All survey questions can be broken down into two groups: quantitative and qualitative.
Quantitative questions can be answered with prewritten options. This includes yes/no, multiple choice and rating scales. Quantitative questions are easy to analyze and easy for respondents to answer. The catch is that they can lack detail and only provide surface level insights.
- How do you feel about our brand?
- How likely are you to recommend our brand to a friend?
- Somewhat likely
- Very likely
Qualitative questions are open-ended and allow respondents to write whatever they want in their own words. Qualitative questions take more time and effort to answer and analyze, but provide more detail. This also allows for survey analysis to uncover those ‘unknown unknowns’ and reveal more information about each individual respondent through their word choices.
- What characteristics do you associate with our brand?
- Would you recommend our brand to a friend and why?
Conducting a survey
Throughout the customer journey there are a few key areas to solicit survey respondents. When a customer hits a milestone (first purchase, 1-year anniversary, etc.), unsubscribing or canceling membership, post-customer service interaction, and the classic abandoned cart scenario. These points in the customer journey usually indicate that a customer is active and has an opinion – whether good or bad.
The big picture
Big picture survey analysis includes the demographic breakdown of respondents, including metadata points like age, location, gender and occupation. It’s crucial to collect this information from respondents so that analysis can be hyper-focused on specific consumer audiences.
Quantitative responses can also be included in big picture analysis. When working with close-ended questions, you can easily see overarching themes and patterns. The black and white nature of quantitative analysis provides a great starting point by revealing what requires deeper digging.
Once you discover those big picture themes, the next step is detailed survey analysis. Zoom in on qualitative responses to discover justification for those overarching quantitative patterns. At this point, we understand the what – but we’re still searching for the all important why. Qualitative responses will give us a more detailed explanation as to why respondents feel the way they do. This is where you’ll gain insight as to how to correct any issues or maintain any strengths.
For example, if we ask a respondent “Do you like our product?” – to which they answer no – that’s all we have. We know they don’t like the product, but we don’t know how to fix it. If we dig deeper into qualitative responses, we can see customers have regular issues with battery strength and customer service representatives.
This portion of survey analysis can also include understanding the relationship between the metadata points mentioned earlier and response patterns. Are respondents over the age of 35 more likely to call your product helpful? That’s what you’ll find out during this step of analysis.
How can I analyze survey data?
Our NLP technology helps users extract meaningful consumer insights from any textual data source – surveys, reviews, social media and more. The platform compares two or more written data sets to pinpoint the differences in topics, words, phrases, grammar and emotion. This text analysis method provides an additional layer of context, allowing you to discover what makes different segments of your audience unique.
Using Relative Insight for survey analysis can save you time and money. Our clients report saving up to 10% of analysts’ time through the automated analysis of open-ends. Others have replaced expensive, full service survey analysis providers with Relative Insight – saving costs on outsourced research while developing their internal capabilities.
By enabling the scalable analysis of open-ends, our clients have also been able to change their approach to surveying by including a greater volume of these questions – allowing them to capture audience feedback in a new way and realize the benefits of text data as a source of business intelligence.
Relative Insight analyzes open-ended survey responses and compares respondents using a range of metadata points. This approach reveals trends, unique opinions and tone of voice specific to your target audiences. Read our case studies or get in touch to learn how Relative Insight can help you.