Skip to main content
Text Analytics in Reports

How to use text analytics to get the most out of your written feedback

Jared Ellis avatar
Written by Jared Ellis
Updated over a week ago

What can I learn from this page?

How to use text analytics to get the most out of your written feedback

Who is this guide for?

Account Admins, Survey Admins, Survey Creators, Report Viewers


  • Engagement comments have a new classification model for enhanced analysis of comments. This model is 25% more precise, has 56% more topics than the previous model and makes 30% fewer mistakes. This model will be applied to new surveys from August 31st 2023

  • If you have launched a survey prior to this date you will see a reduced list of topics. You can speak with a Product Support Specialist by replying with “Ask a Person” in a Support Conversation to discuss adding the new topics model to surveys launched prior to this date.

📌 Note: Text Analytics is available for Engagement, Manager Effectiveness, Lifecycle (onboarding/exit) and Team Effectiveness surveys only. It is not available for Reviewer notes in Lifecycle surveys or Individual Effectiveness surveys.

Text Analytics is a set of tools backed by smart algorithms designed to help you get maximum insight from written feedback. How does it work?

  • automatically classifies every comment into one or more of twenty-eight pre-defined high-level themes (like ‘leadership’, ‘workload or “”compensation)

  • analyzes responses for sentiment (‘positive’, ‘negative’ or ‘neutral’)

  • provides topic, sentiment, rating and question filters that allow you to slice and dice comments into the most useful subset

  • visualizes all the comments in a bubble chart that gives you a bird’s-eye view of what people are saying and how they’re feeling

You will find Text Analytics in the Comments Report via the link in the reporting navigation.

How to Read the Topic Sentiment Chart

The Topic Sentiment Chart, also known as the Topics and Themes chart, is visible on the Comments page of a survey under the Topics and Themes section. This chart combines topic and sentiment analysis to give you a quick overview of all survey comments. Each bubble represents the percentage of comments assigned to that topic. Topics on the right tend to be more positive, while those on the left are more negative. Neutral or polarized topics are in the center. The biggest bubbles represent the largest groups of comments.

The average number underneath the sentiment line indicates whether the overall comments in your survey are more positive or negative. For example, if the number showing is +0.8, it means the overall average (AVG) sentiment of the comments is slightly more positive than negative.

Hover over any bubble for the sentiment breakdown, especially for neutral topics, and see the total number of comments it represents. Percentages on bubbles show comment proportions.

How to Use Comment Filters


We have four additional ways to filter comments: by question, by topic, by sentiment and by rating. You can combine these filters to further narrow the comments in a set – for example, you might like to see negative comments left on positively-rated questions, or just the comments about “Compensation” in response to a specific question.

You can also apply standard demographic filters.

The Text Analytics report can be exported in CSV or Excel format. The Excel export contains a summary sheet in addition to a sheet of all comments and their classifications. Both CSV and Excel exports respect the currently applied demographic and comment filters.

How to Edit a Sentiment or Topic Tag

You can delete or reclassify / retag the sentiment and topic of any comment using the edit function.

Admins can still only choose from the prescribed 4 sentiment options (Positive, Negative, Neutral, No sentiment) and 28 prescribed topics (Autonomy, Benefits, Career, Collaboration, Communication, Company Culture, Company Strategy and Goals, Compensation, Diversity and Inclusion, Environmental and Social Responsibility, Facilities, Feedback, Flexible Working, Hiring, Innovation, Leadership, Learning and Development, Manager, Motivation, Physical Safety, Product and Services, Recognition, Resources and Information, Systems and Equipment, Turnover, Wellbeing, Workload, No Comment) with only 1 sentiment tag. There is no restriction on the number of topics that a comment can be tagged with.


Save Comments to Return to Later

To save any comments, click Save next to the comment in your reporting view:


Each user has their own list of saved comments which are unique to them. Other users are not able to see which comments you have saved within this view.


Use Keyword Search for Comments

  1. Go to the Comments report

  2. To the right of the report you will see the search field

  3. This appears on any survey with a 'comments report' (excludes 360 & performance surveys) on open and closed surveys


Whatever word you enter into the search bar will return comments with exact word matches in any order sequence within the one comment. For example when searching: "Hot desk" you might see: "my desk gets hot being near the window" or "I really like the idea of hot desk". Key points to note:

  • Special characters eg: $!@ are ignored in searches

  • This is an exact match feature eg: when searching for "on" the results will show all comments with "on" NOT "only" or "one."

  • Some languages that use picture characters (such as Chinese or Japanese)can be searched, but require a space between the picture characters to find a match. The platform searches whole words only, not partial words.

Comment Statistics

When viewing the Commenter Statistics, you can see:

  • The percentage of participants who commented on the survey

  • The median number of comments per commenter

  • The commenters' combined average Engagement score, which breaks down the Engagement score for people who commented on the survey.

In the example below, the Commenters Engagement score is 69%. This score is compared to the overall Engagement score of all survey participants (which also includes those who didn't comment). Here, there is a 0% difference compared to the overall score because 100% of participants commented.

The Commenters Engagement score also shows the breakdown of favorable, neutral, and unfavorable responses for that factor:

Learn more about how scores are calculated for reports here.

Note: To display "Commenter Statistics," at least 5 comments are needed. It doesn't mean every question needs 5 comments; just one question needs 5 comments for the statistics to appear.


Who can see text analytics in the comments report?

The access to the Text Analytics report is the same as for the Comments report. If you allow comments access, then Text Analytics will be enabled as well.

Which topics does the classifier recognize?

Our algorithm is trained on over half a million pieces of employee feedback to recognize this set of common topics.

  • Autonomy: Individual or team's ability to make independent decisions

  • Benefits: Non-monetary remuneration including perks, insurance, and parental leave

  • Career: Career opportunities, development plans, and growth support

  • Collaboration: Collaboration within a team or across 2 or more teams

  • Communication: Communication from leadership, between teams, or between employees

  • Company Culture: The sense of belonging in a workplace

  • Company Strategy and Goals: The company's strategic direction, vision, and overall goals

  • Compensation: Monetary remuneration including bonuses, equity, 401k and superannuation, or the need to increase compensation

  • Diversity and Inclusion: Demographic diversity and representation in key roles and a sense of belonging for employees of diverse backgrounds

  • Environmental and Social Responsibility: The company's impact on environmental sustainability and social responsibility, including risk mitigation and public perception

  • Facilities: The physical working environment, break rooms and office spaces

  • Performance Feedback: Feedback on an employee's execution of their role or the feedback process

  • Flexible Working: Remote and flexible working arrangements

  • Hiring: The process or need for hiring

  • Innovation: The company's willingness to adopt new ideas or create new processes and products

  • Leadership: Senior and executive leadership who drive company strategy

  • Learning and Development: Employee learning and development programs and material

  • Manager: The employee's own or another employee's line manager

  • Motivation: The drive to complete tasks at work or go above and beyond

  • Physical Safety: Physical safety concerns and standards, and risk mitigation

  • Product and Services: The company's main offering to customers

  • Recognition: Reward and praise for good work or how recognition is given/isn’t given

  • Resources and Information: Information, resources, and support needed to perform their role

  • Systems and Equipment: Internal systems, processes, tooling, and technology or equipment

  • Turnover: Employees voluntarily or involuntarily leaving the company

  • Wellbeing: Issues impacting the employee's ability to achieve work/life balance including burnout, mental health, and stress

  • Workload: The amount of work or number of responsibilities assigned

  • No Comment: Comments that do not provide further context to a rating response (e.g. "N/A", "all good")

Note: If you have launched a survey prior to the end of August 2023 you will see a reduced list of topics as below:

  • Career: Career progress and development

  • Salary: Compensation and benefits

  • Learning & Development: Learning and development

  • Collaboration: Inter and intra team collaboration

  • Autonomy: Autonomy as individuals or team in making decisions

  • Company Performance: General company performance

  • Products & Services: Related to products and services offered by the company

  • Innovation: Innovation / general improvements in products/services

  • Communication: Management and general organizational communication

  • Systems & Resources: Internal systems. processes and resources

  • Environment: Aspects related to the workplace environment and facilities

  • Feedback: Performance reviews and general feedback

  • Recognition: Role and work recognition

  • Leadership: Related to senior leaders and management in general

  • Manager: Manager-specific comments

  • Work Life: Work life balance

  • Feeling Good: Short comments expressing positive sentiment

  • No comment: Comments that are synonyms for ‘no comment’.

  • Unclassified: Comments that are unable to be classified.

Can I create my own custom topics?

Not at this time.

Are there benchmarks to help me understand whether my results are 'normal'?

We don't have benchmarks in the platform yet, but this article documents some of our findings including which topics are popular in which industries and regions, which topics people talk most negatively about, and which topics people talk most positively about, in general.

Text Analytics isn't very helpful for me – can it be disabled?

Yes! By default, your Comments report will show the topic sentiment bubble chart and topic and sentiment tags for each comment, but you can ask Support to disable this functionality at an account or survey level. Eventually we'll put this power in your hands, but for now you'll need to ask one of our team to do it for you.

Who can adjust the comment sentiment and topics?

Only those with admin permissions can adjust comment sentiment and topics.

How many topics can be assigned per comment?

Only two topics maximum can be assigned per comment.

I have strong feelings about the Comments report and Text Analytics. Can I share them with you?

By all means! Just reply with "Ask a Person" in a Support Conversation to speak with a Product Support Specialist. Please share your ideas and feedback with the specialist and they'll forward them onto our product team. It's a great time to share your thoughts, because we'll be investing a lot of time in this part of the platform over the next few months.

💬 Need further help? Just reply with "Ask a Person" in a Support Conversation to speak with a Product Support Specialist.

Did this answer your question?