From Data to Decisions: Identify Insights Faster through Text Analysis

Almost three years ago, I wrote an article on how to analyze open-ended questions faster with a quick excel trick. It is still a very popular post, and I answer a lot of questions every month (particularly on Step 2!) about how to do it affectively. The main reason it's so popular is because analyzing large quantities of open-ended survey questions has always been painful. It's reason enough to not put comment boxes in surveys. However, it’s a trade off, you save yourself the time and effort of analysis but you’re missing out on valuable insights. Often the most valuable insights are hidden within open-ended comments. In fact, we have clients who have hired employees where all they do is read through survey responses and categorize them. For these organizations, the open-ended comments are critical to their feedback program, and in many cases actually impact their front-line employees' compensation. However, Cvent Web Surveys users can put their aspirin back in their drawers because everything about analyzing open-ended survey questions is about to change.

After next week, all Enterprise and Professional Cvent Web Survey users will see a new Section Link option under the Survey Reports section:
 

TEXT ANALYSIS


Our text analysis reports are not just a word cloud. While you can choose to see the results as a word cloud, if you prefer, you can view the output as a ranked list. What really sets our text analysis apart is the ability to create custom categories and drill down into each common phrase and custom category grouping. Consider the following market research survey example:

 

 

A company is doing some market research for their new energy drink, Pizzazz, and wants to know about people’s leisure activities so they can spend their advertising and marketing dollars effectively. Because an exhaustive list would simply be too long and significantly increase the cognitive load placed on the respondent, the survey author chose to leave this question open ended. These are some of the phrases that came out of the analysis:

 

 

 

Example Survey Question with Text Analysis

 

 

You can see that there are natural groupings between some of the phrases, for example, Saturday, Sunday and Weekend can easily be grouped into a single “Weekend” category. Similarly, you may want to group Baseball, Football, Soccer and Tennis into a single “Sports” category.

 

 

Example Survey Question with Text Analysis and Categories

 

 

By grouping these things together, you can easily report on how many people play Sports on the Weekend.


The Cvent Text Analysis engine is going to be most helpful for large data sets. You’ll want to create these analysis reports after a survey is completed, or when you have on-going surveys such as a customer service satisfaction survey that’s sent after every support call. Using a rolling date, you’ll be able to easily spot spikes in categories – for example, maybe more people are complaining about a Long Wait when contacting customer support – or identify new areas that had never been categorized before – perhaps after a new release, the types of questions customers have changes and you realize you need to update your training documents.

We offer a range of options from font type to size to color for word clouds. Taking the example above for customer support, at a glance you’re able to see that over all people are happy with support (Good Support is the largest category), but poor support and long waits are creeping up. This should indicate you should be looking closer at what’s causing the poor support. In this particular case, my guess would be that the long wait during off hours is causing the overall dissatisfaction.

 

 

 

Example Survey Question with Text Analysis Word Cloud


To know for sure, just click on one of the categories and it will return all answers that were tagged in a category. Or if you want to see who said something about a long wait in the off hours, you can go from the category ranking list, select multiple categories and see all comments that were tagged in both categories. This gives you the ability to quickly see who said A and B.

Like all other Cvent Web Surveys reports, you’re able to add your text analysis into a parked report group for easy sharing within the organization. By adding your text analysis into parked report groups, you can give eyes-only access to various members of your organizations without having to give them access to the full Cvent application. Most importantly, for the first time ever, this type of parked report has the option to be interactive. So just like you can drill down into a phrase or category from within the survey reporting application, you’ll be able to drill down to read the responses from a parked text analysis report as well. Pairing text analysis with parked reports is going to make it much faster to identify insights from survey data and make decisions based on these new findings.

Finally, it’s important to note that our text analysis engine also works for foreign language surveys so if you’re conducting surveys in multiple languages, you’ll be able to create text analysis reports by language.

To learn more about our new text analysis feature and other survey reports, register for new release training. Classes start January 23rd and run through February 6th.

Not a current client but want to learn more about other ways Cvent’s online survey and feedback management solution can help streamline your feedback projects?
Register for an upcoming demonstration.

 

 

blog comments powered by Disqus
Testimonial Video on Reporting