Michael Williams

Let’s start with you. Who are you, and what lens do you bring to understanding people online?

I’m a Market Intelligence Director at Jellyfish, where I’ve been for nearly ten years. I originally started in social media before gradually moving into research and insights, and today I’m the global specialist in social listening. I’ve been conducting social listening analysis for over seven years within our insights team.

My role focuses on understanding audiences to inform our wider creative, media, and planning teams. Having worked with a wide range of clients across different categories and markets, I bring a comparative lens to understanding people online. I’m a firm believer in needs-based segmentation and consistently try to identify the underlying motivations, tensions, and unmet needs that drive online behaviour, rather than taking surface-level conversation at face value.

‍

What’s a working theory you have right now about how people behave online?

A working theory I have is that meaning in online conversations lives more in context than in direct mentions. People rarely speak cleanly about topics - they hint, reference, and emote around them. That’s why precision, proximity, and exclusions matter more than raw volume in social intelligence.

‍

What’s an insight you surfaced that you still think about? What one stuck with you?

Many of the insights that stay with me come from deep-dive projects. One that stands out was a global CPG project exploring lifestyle motivations around laundry in India and Thailand. We built a complex social listening framework around core personas; including community mapping, unmet needs analysis and emerging trends analysis. 

In India, we uncovered that hostel dwellers regularly share washing drums, leading them to combine laundry loads with others. In Thailand, we identified a growing niche community of “house husbands” - men who manage household tasks, including laundry. These small but culturally specific insights revealed unique needs and behaviours that ultimately fed into strategic brand and product recommendations. It reinforced the value of uncovering subcultures rather than relying on broad assumptions.

‍

What’s the weirdest rabbit hole your work has ever sent you down? And what did it teach you?

One of the strangest rabbit holes my work has sent me down was realizing how often context mattered more than the actual keyword I was tracking. By following surrounding language rather than the core term, I uncovered entirely different conversations that explained the “why” behind mentions. It taught me that brands and topics are often secondary characters in much larger narratives.‍

‍

What skills or mindsets do you think the next generation of analysts will need?

Looking ahead, AI and generative workflows will be central to the next generation of analysts. Those who rise to the top will be the ones who use these tools to drive efficiency while enabling deeper insight mining. Strong prompt-writing skills, combined with a solid foundation in social listening and platform knowledge, will be essential.

As more manual tasks become automated, verification, validation, and interpretation of insights will become increasingly important. Strategic thinking and narrative-building will ultimately be what differentiates great analysts from good ones.

‍

What’s a niche community, account, or corner of the internet you’re watching right now? And why?

At Jellyfish, we’re heavy users of Reddit and closely follow multiple category-level subreddits for our clients. These communities consistently surface rich insights around user needs, interests, pain points, and evolving preferences. To us, Reddit has become particularly valuable for understanding how consumer attitudes shift over time and for informing strategic brand and product positioning.

‍

Last non-work thing you read that shaped your thinking?

Mark Ritson’s Marketing MBA shaped how I think about the difference between strategy and execution. It reinforced that many marketing problems aren’t execution issues, but strategy problems - often compounded by poor measurement. That perspective has stayed with me, particularly when interpreting social and audience data.

‍

‍