Skip to main content
Documentation
Documentation
Peakon
Last Updated: 2023-09-08
Concept: Keyword Topics Analysis

Concept: Keyword Topics Analysis

Employee comments provide a rich source of insight and add context to quantitative responses. Topics analysis uses Natural Language Processing (NLP), a field of machine learning, to analyze employee comments and identify important themes and sentiment.

Text Analysis

The algorithm can uncover both industry-wide employee experience topics and specific team and time-dependent topics.
You don't need to use predefined terms or manually add keywords of your own. Peakon summarizes a topic by combining the most representative sentences from multiple comments to provide you with a meaningful overview.
Peakon calculates the comments topics once the survey round closes at 1:00 AM UTC the next day.
New topics generate after:
  • Questions from 2 active core driver questions. Example: A driver from the Engagement question set and a driver from the Company question set.
  • 200 or more comments across all kinds of question sets over the last 3 months for 1 language.
People leaders further down the hierarchy will have fewer topics than leaders further up the hierarchy, who have access to more survey results and comments.
Older comments (more than 3 months old) can contribute to new, related topics. However, they can't generate new themes.
Peakon will seek to populate the top 50 topics, and then place the topics under the respective driver. You could potentially have some drivers without topics, despite there being more than 300 comments.

Sentiment Analysis

Sentiment analysis consists of an algorithm that processes text comments and assigns a numerical value of probability of positive or negative. Example: Scores between -1 and +1, with zero being neutral.
The greater the number of positive words contained in a sentence (Example: Great, friendly, knowledgeable), the more positive sentiment it will have. However, ideas and opinions are complex. Peakon bypasses the potential inaccuracies in sentiment analysis by using the scores given by employees.
This score is more reliable as it comes directly from an employee in relation to a specific driver question that sets a context. As Peakon asks for feedback after every question, our comments are typically short, simple, and focused, and rarely express conflicting sentiment.

Relevant Themes

The Peakon algorithm finds topics from short text answers containing the same words but used in different contexts.
Example 1. Answering a growth question (
I feel that I'm growing professionally
) an employee might comment:
I'm definitely growing in a direction that's more interesting to me, thanks to the development plan that my manager created
.
Example 2. Answering an open-ended question (
If you had a magic wand what’s the one thing you would change about Kinetar?
) An employee could comment:
I really wish we would provide more line manager training, this would be really helpful
.
In example 1, the word
manager
 relates to an employee interaction with the manager. In example 2, relates to a manager asking for more training.
Using examples 1 and 2, we would find that the word
manager
 is important because it displays twice. However, with no context to act on, we get no interesting actionable information about why the word is important.
Using contextual analysis, the Peakon algorithm makes topics more actionable and interesting by:
  • Attaching topics to drivers, which add context to topics. Using Example 1, the algorithm might find
    development plan
     as a topic if it's a combination of words used in relation to growth.
  • Finding multiword topics: In example 1, the answer included the word combination
    development plan
    . When this word combination displays in many answers, Peakon highlights the topic as
    development plan
     and not
    development
    . From Example 2, the topic could be
    line manager training
    ,
    manager training
    , or
    training
     depending on how often the words occur.
  • Removing words from the question phrasing. People have a tendency to repeat the words of the question in the answer. In Example 1, the answer includes the word
    growing
    , which was also in the question. It isn't useful to find this word as a topic, since we already know that we're asking about growth. Therefore, Peakon automatically removes the words from the question phrasing.
  • Reducing the priority on commonly occurring words. The algorithm reduces the priority on words with no specific meaning that display throughout all survey answers.
    Manager
     only displays as a topic if it has a specific meaning in a driver context, providing topics specific to the driver. Example:
    Office
     or
    Room
     for the environment driver or
    Pay rise
     for the driver reward. If we didn't down-prioritize commonly occurring words, we would find
    manager
     as a topic for all drivers.
  • Delivering a topic extract. The algorithm generates a paragraph from the sentences that are most indicative of all the comments that make up a topic. This paragraph gives you an insight into a broad conversation.
  • Ensuring that your topics change over time. Peakon also uses any older comments that are also relevant to the topic. However, the topic will depend on comments made in the last 3 months.