STOP Measuring Reputation Like It’s 1999

Qantas is tying part of CEO Vanessa Hudson's bonus directly to the airline's reputation score. This development signals an important shift in how boards view reputation risk. Australian companies have long understood that reputation crises carry material financial consequences. The past year has been no exception, with a parade of corporate disasters.

Yet most companies still rely on outdated manual surveys and quarterly brand tracking studies to understand their reputation. These traditional approaches are expensive, slow, and capture only a fraction of the conversation happening about their brands in real-time. By the time the quarterly report lands on the CEO's desk, the damage is already done.

What is the alternative? Machine learning and AI of course.

Recent advances in AI have transformed our ability to extract actionable intelligence from a company’s communication ecosystem. When analysed at scale across millions of posts, comments and reactions, clear patterns emerge that traditional focus groups could never capture. The technology can now detect sarcasm, understand context, and differentiate between a temporary service glitch generating frustration and a fundamental product flaw triggering customer exodus.

The importance of this is that platforms like Reddit, which for years corporate communications teams have dismissed as a chaotic echo chamber of anonymous complaints and memes, now hold a wealth of information about a company's reputation.

These same tools are proving equally powerful when turned inward. Companies are discovering they can gauge workforce apathy and engagement levels by analysing workforce communication patterns. The technology reveals whether an organisation has developed dangerous complacency levels that significantly increase risk. These risk signals are of vital importance for cybersecurity breaches and safety incidents that literally cost people their livelihoods and lives.

The real breakthrough isn't just in understanding what people are saying, it's in predicting what comes next. Through my work at Traffyk, we’ve been able to detect issues 2-12 months in advance for ASX 20 companies. Models connected to a diverse range of datasets across all the company's stakeholder channels can identify predictive fingerprints of early warning signals for each type of crisis.

It is important to note: This is not about ‘witch hunting’ or ‘Big Brother’ type surveillance, it is about understanding the wants and needs of the employees, customers and other stakeholders to address the root cause of issues that are often deeply ingrained in many enterprise companies.

When the tools and technology exist today, already deployed across some of Australia's most trusted brands, every day that passes without them is a day closer to the next preventable disaster.

There are no more excuses.

‍

29 September 2025