Sentiment analysis

Find out how happy your customers are

Live chat messages are constantly pouring in through your website. Some of these messages are looking for information, some for support, some to offer an opinion – and it can be hard work just to respond to them all.

But being responsive doesn’t give you time for analysis. In dealing with each chat session as it unfolds, you’re missing out on the bigger picture: the overall satisfaction of your customers. Enter sentiment analysis.

Introducing sentiment analysis

Sentiment analysis is a great way to gauge how well your customers think you’re doing. It works by analysing your live chat sessions for positive or negative feeling – detecting the real mood and meaning beneath words.

With WhosOn, this sentiment analysis is automatic and out of the box. A smart sentiment analysis algorithm is built into the product, ready to help you tap into your customers’ mindset.

Both real-time and reflective

There are two stages of our live chat sentiment analysis. The first is a real-time reflection of how well the chat is going, which adjusts as the chat unfolds. The second is a post-chat analysis, which looks at the completed session and calculates an overall satisfaction score.

This gives you live insights into customer mood, as well as more considered calculations of performance and trends.

A journey through live chat

So, what does that sentiment analysis look like in action? Here’s what happens inside every live chat conversation:

WhosOn detects customer mood

Line by line, WhosOn assesses chat sessions to pick up on mood. It’s trained to recognise emotive keywords, syntactic effect and tone, and can ascertain sentiment in real-time.

A visual mood indicator is displayed

WhosOn uses this mood detection to inform your live chat session. You’ll see a visual mood indicator at the top of each chat window, displayed in simple smiley face format.

Fluid, ongoing measurement

Mood indicators adjust as the chat conversation unfolds. This helps you see at a glance how the customer’s attitude is changing, based on the agent’s responses.

Post-chat cross-referencing

After the chat session has closed, WhosOn then cross-references all data to work out overall customer satisfaction. This combines language, mood analysis, and post-chat survey results.

An overall sentiment score is calculated

Finally, WhosOn produces a sentiment score for each live chat session. Digging deeper than a simplistic “positive” or “negative” rating, WhosOn calculates a measured score of 1-100.

Get smarter over time

Sentiment analysis doesn’t stop after the live chat session has finished. To help you get better and better customer insights, WhosOn uses ongoing sentiment training.

Our algorithm improves with every chat you handle. As more chats and ratings are completed, more data is fed to WhosOn for sentiment analysis. This data helps WhosOn get smarter, and produce ever-more accurate scores. In turn, this helps you get smarter with identifying trends and enhancing service quality.

Transform your customer service

Sentiment analysis transforms the caliber of your customer service. When rolled into your live chat software, it helps you to:

  • Increase first contact resolution
  • Identify or escalate issues quickly
  • Spot upsell or advocacy opportunities
  • Track customer satisfaction trends
  • Assess agent performance
  • Find key emotive triggers
  • See live levels of customer mood
  • Give personalised, adaptive service
  • Strategically segment customers
  • Reduce churn rates

Take WhosOn for a spin

Free 30 day trial on your site - no card required


Please enter a valid domain