_

Client: Global technology firm

TOPIC MODELLING AT A GLOBAL TECHNOLOGY FIRM

A hot topic in analytics is making sense of the vast amount of written text that is produced as we go about our daily work. Here we discuss the application of a method called topic modelling, which can summarise large amounts of written text in fewer key themes.

Brief

In this firm, the CEO regularly records video messages to staff as a way of communicating important information about business priorities. These videos are posted on the corporate intranet for viewing by the staff of over 100,000 employees.

Within a few days, hundreds, sometimes thousands, of comments have been left by employees. This information needs summarising so that it can be fed back to the CEO to take any required action.

There are too many comments to read one-by-one; an analytical approach is required. Using web-scraping technology and text analytics, these comments are analysed quickly by the HR analytics team. A report is provided back to the CEO in a matter of days.

Solution

The report contains not only the results of a sentiment analysis – whether the tone of the comments was positive or negative on balance, right down to the counts of different emojis – but also a summary of the core themes based on a method called topic analysis.

This information is used by the CEO to identify pockets of the organisation where focus is required to persuade and convince about relevant issues. Such analyses were difficult as few as 10 years ago.

Primarily because the approach to communicating with employees via video technology was less common, but the analytics methods were also less well known outside of advanced technology firms. Now such capabilities are within reach of most HR analytics teams.

COULD THIS APPROACH WORK FOR YOUR BUSINESS

Previous
Previous

Podium improves hiring at a professional services firm