Olga Logunova

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

I’ve been working in media studies and social intelligence for over 12 years, and my expertise lies at the intersection of academia and business consultancy, with experience at King's College London, the Healthtech unicorn Flo, and multiple consultancy projects for start-ups and entrepreneurs focused on business growth through social media.

My academic background has equipped me with a range of frameworks and theories for understanding people's behaviour online. For instance, I use celebrity studies frameworks for research goals related to influencer marketing and social media campaign assessments; the theory of platform affordances is highly valuable for exploring cross-platform interaction and content movements; studies on inauthentic behaviour are useful for tasks related to noise and bot filtering, and for understanding media attacks.

At the same time, I have hands-on experience in data analysis and enjoy applying mixed-methods (quantitative and qualitative) to generate insights. I also like to support social media data with insights from web-analytics, surveys, UX studies, etc.

‍

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

From my academic background, no single theory really works on its own; the key is to combine theories and develop frameworks relevant to each specific task.

Social capital theory explains online interactions very well. Capital can take different forms, but at the core are the relationships people build over their lives - classmates, partners, colleagues. Sometimes it feels like this happens much faster online and reaches much larger audiences, but the underlying mechanisms are the same.

Beyond networking, there are also mechanisms of influence over an audience. And if you add Crystal Abidin’s celebrity theory and the idea of post-based fame, where popularity and views are not necessarily tied to an existing audience but just rooted in great value content and being viral purely because of strong content, you recognise TikTok. It's now actively produced on the platform, and you see different types of influencers emerging every week.

And when you combine this with affordance theory, which explains how platforms shape the content we create on them, you get a picture of multi-platform communication among users, brands, and celebrities.

‍

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

Each day, I work with hundreds of unstructured signals from social media data, which transform into tens of insights for different objectives, e.g., generating new product ideas, expanding into new markets, trend watching, supporting creative strategy and user engagement. That’s why it’s really hard to choose one.

Let me tell you about one case that combines business relevance and methodological angle.

Using a social listening tool, I detected a spike in consumer conversation in a non-relevant subreddit thread. The discussion captured was relevant and helped in generating new product ideas. However, the Reddit topic-starter was a screenshot from a random user's Instagram Story that the social listening tool did not capture. Later, TikTok users began creating content inspired by the same Instagram Story, significantly boosting brand visibility.

So, one piece of user content, moving across three platforms, with zero brand control, resulted in massive brand visibility. UGC doesn’t stay where it was created. It travels where the audience decides. Brands need to carefully explore the media landscape and capitalise on these unexpected organic waves of consumer interest.

‍

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

One of my most recent studies was a botnet study, focused on understanding inauthentic behaviour in political conversations, fake news, misinformation, and manipulation of public opinion.

The first challenge was defining the concept itself, because to identify inauthentic behaviour, you first need to understand what authentic behaviour looks like. So we worked with several data layers.

At the message layer, we looked at factors such as post length, likes, views, posting rhythm and sequences, aggression index in the text, and the narratives used.

At the author layer, we analysed subscribers and subscriptions, profile completeness (geo, description, date of birth), account creation date, and on YouTube, also publishing patterns and playlists.

And finally, the network layer: who comments on whom, the actual connections and cluster structures. This became the key element.

When you start looking at thousands of these connections and networks, you realise that you can keep splitting and rebuilding the network into different clusters almost endlessly, depending on which variables you apply.

Every time you build and rebuild a network, you’re trying to find the “ideal” structure but that structure only exists in relation to your questions. As soon as you change the question, you build a new map. And you can keep searching for insights almost infinitely - the complexity of this approach really allows for that.

And the questions are endless too:

Which communication strategies are most effective against misinformation?

Which groups have more influence, and what are their characteristics?

All of this helps to reveal what information wars look like today.

‍

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

The media landscape is becoming increasingly fragmented across platforms and markets. But to understand each piece, you need to see the full context. That’s why systems thinking and the ability to build complex conceptual models are so important. They help you get to cause-and-effect relationships, not just linear descriptions. Mathematics, logic, and data storytelling are also essential.

Being able to answer the question “why” is critical.

I also think that over-reliance on AI summarisation tools can lead to a loss of deep understanding. Analysts need to be able to think conceptually and critically, structure signals into frameworks, and turn data into relevant business recommendations - not just summaries.

‍

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

Right now, it’s definitely Reddit.

It’s a platform that continues to attract a very strong organic audience, with many niche communities. It’s highly organic and is often used as the final step in decision-making. Recent data also shows that it’s the number one source for LLM models, making its content valuable not only for users but also for GEO (Generative Engine Optimisation).

Reddit's audience is clearly growing, and it has already overtaken TikTok in the UK to become the fourth most-visited social media platform.

One more domain - LLM data. Chat GPT, Gemini, Claude, etc., have created a new type of content, available for users and impacting their behaviours. We need to learn how to properly analyse this new data layer.  Social intelligence is evolving into a broader system, not only about social media, but also about search, LLM search, and reviews. It is not just social. Digital intelligence? What is our next way of presenting ourselves as experts?

And I’ve even tested it myself; I already have two followers there after just a couple of months, and it ranks number three in Google search results for my own name. Modest results, but huge impact — that’s what Reddit is today for brands.

‍

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

For me, it’s art, and I’ll explain this using opera as a multi-layered case.
My love from music school evolved into going to productions where you disconnect for 3–5 hours and fully immerse yourself in a space created by music, voices, and performance. It’s an incredibly multi-layered form of culture that you don’t just decode, but interpret depending on your mood. The same symbols can be highlighted differently, and this endless play of meanings is fascinating.

This constant reconstruction, deconstruction, and work with signals feels very close to how I work with social data. As if you’re given raw data, and you build a framework in which it works best for a specific business task. There is always room for new interpretation: new markets, different audience perceptions. And the audience itself is a whole separate layer: observing live opera crowds gives insights into interactions, social structures, and the communities that form around making a theatre visit unforgettable.

‍

‍

‍