How to Analyze Your Survey Responses in 2021

Customer satisfaction surveys can be one of the quickest and most dependable ways to get useful information about your brand, your products, and individual product features, at every step of the customer journey.

Once you’ve collected your survey data, though, you’ll need to analyze the results.

The type of questions you ask and the number of responses you receive will shape the way you perform survey analysis. Ask close-ended questions and you’ll be ready to analyze your data with everyday tools like Excel. Ask open-ended questions, and you’ll need more advanced data analysis tools that are equipped with AI.

In this post, we’ll show you how easy it is to perform survey data analysis on both quantitative and qualitative customer feedback.

What Is Survey Analysis?

By digging into the details of survey responses, you can better understand customers, employees, and your overall brand performance.

Types of Survey Data

There are two basic data types in surveys: open-ended questions and close-ended questions.

Close-ended questions deliver quantitative data, also known as structured data. You’d perform statistical analysis on this survey data since it’s quantifiable. The answers to close-ended questions are usually multiple-choice, rated on a number scale, or one-word answers, like Yes/No.

The customer satisfaction (CSAT) survey is the most widely used questionnaire for requesting customer feedback, and can quickly help companies measure how happy their customers are with their product, service, event, or brand. The NPS score is by far the most well known CSAT score and buckets customers into three groups based on their responses.

Most surveys, like NPS surveys, often ask a follow-up question to understand why customers leave a high or low score. These open-ended questions elicit more detailed responses and can help companies really understand what they’re doing right or wrong.

Open-ended responses can be categorized as qualitative data or unstructured data. They’re harder to analyze because they are not organized in any predefined manner.

The differences between qualitative and quantitative data, including a definition, types of data, what it answers, and examples of both data types

Later on, we’ll introduce you to survey analysis tools that make it easy to break down open-ended survey results for real qualitative insights. But first, let’s explore how to analyze survey data in Excel.

How to Analyze Quantitative Survey Data in Excel

To get the most out of your survey responses, you need to decide what you’re looking for. What are your goals? What are the insights you want to gather? Start with the end result in mind.

Analyzing quantitative survey data in Excel can be a snap with built-in formulas, tables, and charts. The below examples show the results of over 2,500 survey responses for the messaging app, Slack. Reviewers ranked the app from 1 to 5 stars on general use (“Stars” column), value for money, ease of use, features, and customer support. A snapshot of the overall data in Excel:

Data from Slack reviews organized into an Excel table.

Filter survey data by different criteria

The below shows reviews filtered by survey respondents that delivered a rating of 3 stars or lower. To do this, click the left-hand corner to select the entire table, and choose ‘turn on filter’ by selecting the filter icon. Then, choose a conditional value of 3 or lower, and rename this view to ‘detractors’, survey respondents that give you a below-average rating.

Slack reviews in Excel s filtered by those that ranked 3 stars or lower.

Calculate mean, maximum, and minimum

The columns, rows, and cells in Excel have built-in formulas. In this case, simply highlight the entire column (or group of columns), and choose the correct formula in Excel. Type ‘=average’, ‘=max’, or ‘=min’ in the corresponding field and Excel will calculate the statistics for you.

Mean, maximum, and minimum calculations for Slack reviews in Excel.

Perform cross-tabulation with a pivot table

A pivot table is a new (pivoted) table that summarizes the data of a more comprehensive table. Also referred to as cross-tabulation, it can provide a quick comparison of how different groups of respondents answered your survey questions, for example, you could focus on respondents in a particular age group, with a specific occupation, and more.

To create a pivot table, select the cells you want to use, click ‘Insert’from the menu bar, select ‘PivotTable,’ then choose the location for your pivot table. Drag the fields you want to use into the ‘PivotTable Fields’ pane that pops up.

In this table, we can see a summary of the column "Stars." It counts the number of times each star appears in a review.

An Excel pivot table of Slack review data.

Create charts and graphs to visualize data

Simply choose the chart or graph you’d like to use from the ‘Insert’ menu, and Excel will walk you through choosing your fields. This graph shows the percentage value by star for each category. We can see that the results are quite positive, with more than 50% of the reviews being 5 stars (in all categories). 

Graph showing percentage results of star ratings for each category of Slack ratings.

How to Analyze Qualitative Survey Data with AI

Qualitative data can be harder to analyze because it can’t simply be turned into numerical values to run through Excel formulas. But asking open-ended questions will help you to discover reactions and emotions that can’t easily be understood with Yes/No questions.

With AI tools, like MonkeyLearn, you can transform open-ended survey responses into real, actionable insights in just a few steps, and in next to no time at all.

MonkeyLearn is a SaaS platform with powerful text analysis tools that can automatically organize surveys, read responses for sentiment and emotion, extract the most important words, and more. MonkeyLearn offers no-code solutions to build your own machine learning models and analyze huge amounts of data in real-time.

You can simply upload an Excel file with long-form survey responses and perform any number of text analyses, then output the results back into Excel. Or upload and download directly from survey tools you already use, like SurveyMonkey, Survey Gizmo, Google Forms, and more.

Sentiment analysis is a cornerstone of text analysis that uses natural language processing (NLP) to categorize text data, like survey responses, into positive, negative, or neutral.

Sentiment analysis goes beyond mere word definitions into opinion and emotion, even sarcasm. And it can be used on social media, customer service tickets, regular documents, all manner of text.

Try our this MonkeyLearn’s pre-trained sentiment analyzer:

Test with your own text

Results

TagConfidence
Positive100.0%

Powerful algorithms break the comment down and easily score the sentiment as Positive.

Some open-ended survey responses, however, may contain more than one opinion. To gain a more accurate analysis, these have to be separated into individual opinion units (fragments of text), otherwise, responses will be categorized as ‘neutral’ since opposing sentiments in one text will cancel each other out. 

Try out how this pre-trained opinion unit extractor, below, separates text into smaller fragments: 

Test with your own text

Results

TagValue
OPINIONThe hotel has a great location
OPINIONbut all in all it was a horrible experience!
OPINIONOnly stayed here because it was the pre-accommodation choice for one of our tours
OPINIONWill never stay here again!

Once you have separate opinion units, you can get even more granular with your data analysis

Aspect-based sentiment analysis, for example, will break the text down into aspects (categories), and then assign a sentiment to each piece of text. 

If we’re surveying software customers, we could use categories, like Usability, Functionality, and Support, to sort texts.

So, we would end up with aspect-based sentiment analysis in Excel that looks like this:

Opinion units organized in an excel spreadsheet

Aspect-based sentiment analysis is perfect for analyzing your open-ended questions because customers will have different opinions about different products or aspects of the business.

Getting started with qualitative data analysis is easy. You can even train your own models, specific to your business, and your criteria in just a few minutes.

Just sign up to MonkeyLearn, then watch this tutorial to learn how to create your own survey analyzer.

To connect your survey data, you have one of three options:

Create A Survey Results Report

Bring your data together and uncover even more insights with MonkeyLearn Studio – your all-in-one survey analysis and data visualization tool.

The below is a sample dashboard showing an analysis of reviews for Zoom:

The MonkeyLearn Studio dashboard showing multiple text analysis results together.

You can see sentiment analysis by category (aspect), individual responses by date, and keyword extraction that shows the most used and most important words: in the ‘Qualifier’ section on the bottom left and in the word clouds created on the bottom right.

Data visualization tools can put all of your survey data into one place for an expansive view or finite insights. Get to the data that’s not simply visible on a simple spreadsheet or graph.

With MonkeyLearn Studio you can manipulate your survey data and your findings right in the data visualization dashboard.

MonkeyLearn Studio offers countless insights and visualizations that can produce game-changing results for presentations, board meetings, and personal edification.

Get Started with Survey Data Analysis

Customer surveys are an undeniably important aspect of business development, and Excel can be a helpful tool to analyze your quantitative survey results.

Using Excel in concert with survey applications and AI survey analysis tools can take your data to the next level.

Find out how MonkeyLearn can help you analyze your surveys and improve your analytics by showcasing your data in broad strokes or finite detail.

You can try out most of the powerful text analysis tools for free to see how they work. Or schedule a demo to understand the true potential of machine learning.

Tobias Geisler Mesevage

September 4th, 2020