By Lucy Lavery, Tom Foley & Tim Zecchin

AVE is dead. Buried. Never coming back.

But here we are again – chasing shiny metrics and calling it strategy. This time, it’s sentiment. The question is: are we doing analysis which is transformative, or getting hooked on something just as flawed?

Spoiler alert: sentiment can be powerful. But only if you stop treating it like a prettier version of AVE. On its own, it’s just another number. Add intelligence, context, and intention – now you’ve got real insight.

At UNICEPTA by PRophet, we’ve built sentiment models that inform global health policy, shape brand narratives, and measure trust at scale. But we’ve also seen how fast sentiment metrics can be mis-used and mis-interpreted. As we head to Vienna and immerse ourselves in everything communication measurement at AMEC’s Global Summit, here’s our take – unfiltered, unsugarcoated, and unapologetically practical.

1. Define the Rule of the Game

Positive. Neutral. Negative.

That’s the default language of sentiment. But without context, it’s noise. Your brand could be drowning in “positive” coverage and still lose trust. Or weather a wave of “negativity” and come out stronger.

So start by asking:

If you don’t answer those questions first, sentiment isn’t a signal. It’s a smokescreen.

2. If Your Model Is Generic, Your Insight Will Be Too 

Fast is good. Simple is fine. But choosing generic over custom-built is careless.

Sure, a three-point scale is useful for snapshots. It helps in a crisis. It benchmarks competitors. But it won’t tell you if that “positive” headline was actually lukewarm, or if the “negative” spike was a sign of relevance – not risk.

That’s why we’ve built sentiment models that go deeper:

Because when sentiment gets smart, it stops being reactive. It starts being predictive.

3. AI Is Only as Smart as the Questions You Ask It 

Let’s be real: AI is fast, scalable, and powerful. But it’s not a mind reader.

Drop a legal report into a generic sentiment engine, and you’ll see “negative” all over the place: litigation, defamation, lawsuit. Except… those are wins in that world.

Same with humanitarian comms. “Famine” and “crisis” aren’t red flags – they’re the mission.

AI alone won’t get that. That’s why our models blend automation with human oversight from day one. We don’t just fine-tune the tech – we teach it to think like your brand.

4. Don’t Just Show Data – Shape the Story  

Presentation matters. It changes how your audience interprets the data – and what they do with it.

One chart shows Microsoft on top. Another which shows absolute volumes rather than sentiment breakdown as a percentage? Amazon. Add a timeline? Now Google looks like the future.

Data visualization isn’t just pretty slides. It’s strategy. We help brands choose the right lens, but light and digestible does not rule. Telling the full story could mean showing more, not less.

And when simple visuals won’t cut it, we build custom impact scores that go deeper.

5. Want Real Insight? Go Beyond Positive vs Negative  

You can’t understand emotion in black and white.

When the World Health Organization worked with us to decode public reactions during the pandemic, “positive” and “negative” weren’t enough. So we built a model based on the stages of grief – anger, denial, anxiety, acceptance – to help them shape communications that moved audiences toward informed behaviors.

At COP26, we tracked public mood from DOOM to HOPE – because that’s what climate comms is about: mindset, not just mentions.

And for a global financial institution, we measured trust through three attributes: competence, integrity, efficiency. It didn’t just show how people felt – it showed where the brand needed to grow.

That’s what happens when sentiment analysis stops being a checkbox – and becomes a strategy.

6. What’s the Playbook?  

Here’s what we believe:

If you’re still chasing sentiment as a standalone stat, you’re already behind. If you’re using it to ask smarter questions, spot the real signals, and guide action – congrats. You’re doing it right.

At UNICEPTA by PRophet, we don’t do cookie-cutter anything. We build tailored frameworks that help the world’s biggest brands see through the noise and act with clarity. And we do it fast – powered by AI, guided by expertise, and founded on a solid understanding of your brand and purpose

Because the future of media intelligence isn’t about measuring everything. It’s about measuring what matters. More importantly still, it’s about what matters to you.